Stowage/go
2
0
Fork 0
mirror of https://github.com/golang/go.git synced 2025-12-08 06:10:04 +00:00
Code Issues Projects Releases Packages Wiki Activity
go/src/cmd/compile/internal/ssa/opGen.go

8030 lines
159 KiB
Go
Raw Normal View History

[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
// autogenerated: do not edit!
// generated from gen/*Ops.go
cmd/compile/internal/ssa: cleanup godocs Add a blank line before the "package ssa" lines so the "autogenerated don't edit" comments don't end up in godoc output. Change-Id: I82bf90d52d426ce1a8e21483fc8f47b3689259c7 Reviewed-on: https://go-review.googlesource.com/20086 Reviewed-by: Keith Randall <khr@golang.org>
2016-03-01 10:58:06 -08:00
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
package ssa
cmd/internal/obj: add As type for assembly opcodes Passes toolstash/buildall. Fixes #14692. Change-Id: I4352678d8251309f2b8b7793674c550fac948006 Reviewed-on: https://go-review.googlesource.com/20350 Reviewed-by: Dave Cheney <dave@cheney.net> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
2016-03-07 18:00:08 -08:00
import (
"cmd/internal/obj"
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
"cmd/internal/obj/arm"
[dev.ssa] cmd/compile: Add initial SSA configuration for PPC64 This adds the initial SSA implementation for PPC64. Builds golang and all.bash runs correctly. Simple hello.go builds but does not run. Change-Id: I7cec211b934cd7a2dd75a6cdfaf9f71867063466 Reviewed-on: https://go-review.googlesource.com/24453 Reviewed-by: David Chase <drchase@google.com> Reviewed-by: Cherry Zhang <cherryyz@google.com> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-06-24 14:37:17 -05:00
"cmd/internal/obj/ppc64"
cmd/internal/obj: add As type for assembly opcodes Passes toolstash/buildall. Fixes #14692. Change-Id: I4352678d8251309f2b8b7793674c550fac948006 Reviewed-on: https://go-review.googlesource.com/20350 Reviewed-by: Dave Cheney <dave@cheney.net> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
2016-03-07 18:00:08 -08:00
"cmd/internal/obj/x86"
)
[dev.ssa] cmd/compile/internal/gc: reduce genValue redundancy Add an asm field to opcodeTable containing the Prog's as field. Then instructions that fill the Prog the same way can be collapsed into a single switch case. I'm still thinking of a better way to reduce redundancy, but I think this might be a good temporary solution to prevent duplication from getting out of control. What do you think? Change-Id: I0c4a0992741f908bd357ee2707edb82e76e4ce61 Reviewed-on: https://go-review.googlesource.com/11130 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Keith Randall <khr@golang.org>
2015-06-16 11:11:16 -07:00
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
const (
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
BlockInvalid BlockKind = iota
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
BlockAMD64EQ
BlockAMD64NE
BlockAMD64LT
BlockAMD64LE
BlockAMD64GT
BlockAMD64GE
BlockAMD64ULT
BlockAMD64ULE
BlockAMD64UGT
BlockAMD64UGE
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
BlockAMD64EQF
BlockAMD64NEF
BlockAMD64ORD
BlockAMD64NAN
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
BlockARMEQ
BlockARMNE
BlockARMLT
BlockARMLE
BlockARMGT
BlockARMGE
BlockARMULT
BlockARMULE
BlockARMUGT
BlockARMUGE
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
BlockPlain
BlockIf
BlockCall
cmd/compile: fix defer/deferreturn Make sure we do any just-before-return cleanup on all paths out of a function, including when recovering. Each exit path should include deferreturn (if there are any defers) and then the exit code (e.g. copying heap-escaping return values back to the stack). Introduce a Defer SSA block type which has two outgoing edges - one the fallthrough edge (the defer was queued successfully) and one which immediately returns (the defer had a successful recover() call and normal execution should resume at the return point). Fixes #14725 Change-Id: Iad035c9fd25ef8b7a74dafbd7461cf04833d981f Reviewed-on: https://go-review.googlesource.com/20486 Reviewed-by: David Chase <drchase@google.com>
2016-03-09 19:27:57 -08:00
BlockDefer
[dev.ssa] cmd/compile: optimize nil checks Use faulting loads instead of test/jeq to do nil checks. Fold nil checks into a following load/store if possible. Makes binaries about 2% smaller. Change-Id: I54af0f0a93c853f37e34e0ce7e3f01dd2ac87f64 Reviewed-on: https://go-review.googlesource.com/16287 Reviewed-by: David Chase <drchase@google.com>
2015-10-23 19:12:49 -07:00
BlockCheck
[dev.ssa] cmd/compile/internal/ssa: distinguish exit and return blocks It is confusing to have exceptional edges jump back into real code. Distinguish return blocks, which execute acutal code, and the exit block, which is a merge point for the regular and exceptional return flow. Prevent critical edge insertion from adding blocks on edges into the exit block. These added blocks serve no purpose and add a bunch of dead jumps to the assembly output. Furthermore, live variable analysis is confused by these jumps. Change-Id: Ifd69e6c00e90338ed147e7cb351b5100dc0364df Reviewed-on: https://go-review.googlesource.com/14254 Reviewed-by: David Chase <drchase@google.com>
2015-09-03 09:09:59 -07:00
BlockRet
[dev.ssa] cmd/compile/internal/ssa: implement ORETJMP Change-Id: I352c7b9aab053959bc74c15861339e1dbe545ddc Reviewed-on: https://go-review.googlesource.com/14404 Reviewed-by: David Chase <drchase@google.com>
2015-09-08 21:28:44 -07:00
BlockRetJmp
[dev.ssa] cmd/compile/internal/ssa: simplify how exit blocks are used Move to implicit (mostly) instead of explicit exit blocks. RET and RETJMP have no outgoing edges - they implicitly exit. CALL only has one outgoing edge, as its exception edge is implicit as well. Exit blocks are only used for unconditionally panicking code, like the failed branches of nil and bounds checks. There may now be more than one exit block. No merges happen at exit blocks. The only downside is it is harder to find all the places code can exit the method. See the reverse dominator code for an example. Change-Id: I42e2fd809a4bf81301ab993e29ad9f203ce48eb0 Reviewed-on: https://go-review.googlesource.com/14462 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-09-09 18:03:41 -07:00
BlockExit
BlockFirst
[dev.ssa] cmd/compile: Add initial SSA configuration for PPC64 This adds the initial SSA implementation for PPC64. Builds golang and all.bash runs correctly. Simple hello.go builds but does not run. Change-Id: I7cec211b934cd7a2dd75a6cdfaf9f71867063466 Reviewed-on: https://go-review.googlesource.com/24453 Reviewed-by: David Chase <drchase@google.com> Reviewed-by: Cherry Zhang <cherryyz@google.com> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-06-24 14:37:17 -05:00
BlockPPC64EQ
BlockPPC64NE
BlockPPC64LT
BlockPPC64LE
BlockPPC64GT
BlockPPC64GE
BlockPPC64ULT
BlockPPC64ULE
BlockPPC64UGT
BlockPPC64UGE
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
)
var blockString = [...]string{
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
BlockInvalid: "BlockInvalid",
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
BlockAMD64EQ: "EQ",
BlockAMD64NE: "NE",
BlockAMD64LT: "LT",
BlockAMD64LE: "LE",
BlockAMD64GT: "GT",
BlockAMD64GE: "GE",
BlockAMD64ULT: "ULT",
BlockAMD64ULE: "ULE",
BlockAMD64UGT: "UGT",
BlockAMD64UGE: "UGE",
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
BlockAMD64EQF: "EQF",
BlockAMD64NEF: "NEF",
BlockAMD64ORD: "ORD",
BlockAMD64NAN: "NAN",
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
BlockARMEQ: "EQ",
BlockARMNE: "NE",
BlockARMLT: "LT",
BlockARMLE: "LE",
BlockARMGT: "GT",
BlockARMGE: "GE",
BlockARMULT: "ULT",
BlockARMULE: "ULE",
BlockARMUGT: "UGT",
BlockARMUGE: "UGE",
[dev.ssa] cmd/compile/internal/ssa: implement ORETJMP Change-Id: I352c7b9aab053959bc74c15861339e1dbe545ddc Reviewed-on: https://go-review.googlesource.com/14404 Reviewed-by: David Chase <drchase@google.com>
2015-09-08 21:28:44 -07:00
BlockPlain: "Plain",
BlockIf: "If",
BlockCall: "Call",
cmd/compile: fix defer/deferreturn Make sure we do any just-before-return cleanup on all paths out of a function, including when recovering. Each exit path should include deferreturn (if there are any defers) and then the exit code (e.g. copying heap-escaping return values back to the stack). Introduce a Defer SSA block type which has two outgoing edges - one the fallthrough edge (the defer was queued successfully) and one which immediately returns (the defer had a successful recover() call and normal execution should resume at the return point). Fixes #14725 Change-Id: Iad035c9fd25ef8b7a74dafbd7461cf04833d981f Reviewed-on: https://go-review.googlesource.com/20486 Reviewed-by: David Chase <drchase@google.com>
2016-03-09 19:27:57 -08:00
BlockDefer: "Defer",
[dev.ssa] cmd/compile: optimize nil checks Use faulting loads instead of test/jeq to do nil checks. Fold nil checks into a following load/store if possible. Makes binaries about 2% smaller. Change-Id: I54af0f0a93c853f37e34e0ce7e3f01dd2ac87f64 Reviewed-on: https://go-review.googlesource.com/16287 Reviewed-by: David Chase <drchase@google.com>
2015-10-23 19:12:49 -07:00
BlockCheck: "Check",
[dev.ssa] cmd/compile/internal/ssa: implement ORETJMP Change-Id: I352c7b9aab053959bc74c15861339e1dbe545ddc Reviewed-on: https://go-review.googlesource.com/14404 Reviewed-by: David Chase <drchase@google.com>
2015-09-08 21:28:44 -07:00
BlockRet: "Ret",
BlockRetJmp: "RetJmp",
[dev.ssa] cmd/compile/internal/ssa: simplify how exit blocks are used Move to implicit (mostly) instead of explicit exit blocks. RET and RETJMP have no outgoing edges - they implicitly exit. CALL only has one outgoing edge, as its exception edge is implicit as well. Exit blocks are only used for unconditionally panicking code, like the failed branches of nil and bounds checks. There may now be more than one exit block. No merges happen at exit blocks. The only downside is it is harder to find all the places code can exit the method. See the reverse dominator code for an example. Change-Id: I42e2fd809a4bf81301ab993e29ad9f203ce48eb0 Reviewed-on: https://go-review.googlesource.com/14462 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-09-09 18:03:41 -07:00
BlockExit: "Exit",
BlockFirst: "First",
[dev.ssa] cmd/compile: Add initial SSA configuration for PPC64 This adds the initial SSA implementation for PPC64. Builds golang and all.bash runs correctly. Simple hello.go builds but does not run. Change-Id: I7cec211b934cd7a2dd75a6cdfaf9f71867063466 Reviewed-on: https://go-review.googlesource.com/24453 Reviewed-by: David Chase <drchase@google.com> Reviewed-by: Cherry Zhang <cherryyz@google.com> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-06-24 14:37:17 -05:00
BlockPPC64EQ: "EQ",
BlockPPC64NE: "NE",
BlockPPC64LT: "LT",
BlockPPC64LE: "LE",
BlockPPC64GT: "GT",
BlockPPC64GE: "GE",
BlockPPC64ULT: "ULT",
BlockPPC64ULE: "ULE",
BlockPPC64UGT: "UGT",
BlockPPC64UGE: "UGE",
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
}
func (k BlockKind) String() string { return blockString[k] }
const (
OpInvalid Op = iota
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
OpAMD64ADDSS
OpAMD64ADDSD
OpAMD64SUBSS
OpAMD64SUBSD
OpAMD64MULSS
OpAMD64MULSD
OpAMD64DIVSS
OpAMD64DIVSD
OpAMD64MOVSSload
OpAMD64MOVSDload
OpAMD64MOVSSconst
OpAMD64MOVSDconst
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
OpAMD64MOVSSloadidx1
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
OpAMD64MOVSSloadidx4
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
OpAMD64MOVSDloadidx1
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
OpAMD64MOVSDloadidx8
OpAMD64MOVSSstore
OpAMD64MOVSDstore
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
OpAMD64MOVSSstoreidx1
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
OpAMD64MOVSSstoreidx4
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
OpAMD64MOVSDstoreidx1
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
OpAMD64MOVSDstoreidx8
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
OpAMD64ADDQ
OpAMD64ADDL
OpAMD64ADDQconst
OpAMD64ADDLconst
OpAMD64SUBQ
OpAMD64SUBL
OpAMD64SUBQconst
OpAMD64SUBLconst
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpAMD64MULQ
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
OpAMD64MULL
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpAMD64MULQconst
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
OpAMD64MULLconst
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
OpAMD64HMULQ
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
OpAMD64HMULL
OpAMD64HMULW
OpAMD64HMULB
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
OpAMD64HMULQU
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
OpAMD64HMULLU
OpAMD64HMULWU
OpAMD64HMULBU
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
OpAMD64AVGQU
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
OpAMD64DIVQ
OpAMD64DIVL
OpAMD64DIVW
OpAMD64DIVQU
OpAMD64DIVLU
OpAMD64DIVWU
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
OpAMD64MODQ
OpAMD64MODL
OpAMD64MODW
OpAMD64MODQU
OpAMD64MODLU
OpAMD64MODWU
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
OpAMD64ANDQ
OpAMD64ANDL
OpAMD64ANDQconst
OpAMD64ANDLconst
OpAMD64ORQ
OpAMD64ORL
OpAMD64ORQconst
OpAMD64ORLconst
OpAMD64XORQ
OpAMD64XORL
[dev.ssa] cmd/compile: implement ONOT Co-hacking with josharian at Gophercon. Change-Id: Ia59dfab676c6ed598c2c25483439cd1395a4ea87 Reviewed-on: https://go-review.googlesource.com/12029 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
2015-07-10 11:25:48 -06:00
OpAMD64XORQconst
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
OpAMD64XORLconst
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpAMD64CMPQ
[dev.ssa] cmd/compile/internal/ssa/gen: implement all bit sizes for Eq and Neq Change-Id: I2c5c75153493b5dd3f1a743e5edf04403e83b31b Reviewed-on: https://go-review.googlesource.com/12474 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-21 18:06:15 +02:00
OpAMD64CMPL
OpAMD64CMPW
OpAMD64CMPB
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
OpAMD64CMPQconst
OpAMD64CMPLconst
OpAMD64CMPWconst
OpAMD64CMPBconst
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
OpAMD64UCOMISS
OpAMD64UCOMISD
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpAMD64TESTQ
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
OpAMD64TESTL
OpAMD64TESTW
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpAMD64TESTB
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
OpAMD64TESTQconst
OpAMD64TESTLconst
OpAMD64TESTWconst
OpAMD64TESTBconst
OpAMD64SHLQ
OpAMD64SHLL
OpAMD64SHLQconst
OpAMD64SHLLconst
OpAMD64SHRQ
OpAMD64SHRL
OpAMD64SHRW
OpAMD64SHRB
OpAMD64SHRQconst
OpAMD64SHRLconst
OpAMD64SHRWconst
OpAMD64SHRBconst
OpAMD64SARQ
OpAMD64SARL
OpAMD64SARW
OpAMD64SARB
OpAMD64SARQconst
OpAMD64SARLconst
OpAMD64SARWconst
OpAMD64SARBconst
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
OpAMD64ROLQconst
OpAMD64ROLLconst
OpAMD64ROLWconst
OpAMD64ROLBconst
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
OpAMD64NEGQ
OpAMD64NEGL
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
OpAMD64NOTQ
OpAMD64NOTL
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
OpAMD64BSFQ
OpAMD64BSFL
OpAMD64BSFW
OpAMD64BSRQ
OpAMD64BSRL
OpAMD64BSRW
OpAMD64CMOVQEQconst
OpAMD64CMOVLEQconst
OpAMD64CMOVWEQconst
OpAMD64CMOVQNEconst
OpAMD64CMOVLNEconst
OpAMD64CMOVWNEconst
OpAMD64BSWAPQ
OpAMD64BSWAPL
[dev.ssa] cmd/compile/internal/ssa: implement OSQRT Change-Id: Iec61ca1bdc064c29ceca6d47f600d5643d0a64dd Reviewed-on: https://go-review.googlesource.com/14533 Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
2015-09-12 13:26:57 -07:00
OpAMD64SQRTSD
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
OpAMD64SBBQcarrymask
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
OpAMD64SBBLcarrymask
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpAMD64SETEQ
OpAMD64SETNE
OpAMD64SETL
[dev.ssa] cmd/compile/ssa: add comparison ops Increase SSA coverage of functions in the standard library from 20.79% to 27.81%. The most significant unimplemented items are now: 10.16% 2597 SSA unimplemented: zero for type error not implemented 8.44% 2157 SSA unimplemented: addr: bad op DOTPTR 7.98% 2039 SSA unimplemented: unhandled OLITERAL 7 6.29% 1607 SSA unimplemented: unhandled expr OROR 4.73% 1209 SSA unimplemented: unhandled expr LEN 4.55% 1163 SSA unimplemented: unhandled expr LROT 3.42% 874 SSA unimplemented: unhandled OLITERAL 6 2.46% 629 SSA unimplemented: unhandled expr DOT 2.41% 615 SSA unimplemented: zero for type []byte not implemented 2.02% 516 SSA unimplemented: unhandled stmt CALLMETH 1.90% 486 SSA unimplemented: unhandled expr ANDAND 1.79% 458 SSA unimplemented: unhandled expr CALLINTER 1.69% 433 SSA unimplemented: unhandled stmt SWITCH 1.67% 428 SSA unimplemented: unhandled expr CALLMETH 1.67% 426 SSA unimplemented: unhandled expr CLOSUREVAR Change-Id: I40959b22993c4f70784b4eca472cae752347879c Reviewed-on: https://go-review.googlesource.com/11452 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-24 17:48:22 -07:00
OpAMD64SETLE
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpAMD64SETG
OpAMD64SETGE
OpAMD64SETB
[dev.ssa] cmd/compile: finish implementing comparisons Change-Id: I4e496c7c7239111133631f76ca25e14be64800c6 Reviewed-on: https://go-review.googlesource.com/12656 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-24 12:47:00 -07:00
OpAMD64SETBE
OpAMD64SETA
OpAMD64SETAE
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
OpAMD64SETEQF
OpAMD64SETNEF
OpAMD64SETORD
OpAMD64SETNAN
OpAMD64SETGF
OpAMD64SETGEF
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
OpAMD64MOVBQSX
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
OpAMD64MOVBQZX
OpAMD64MOVWQSX
OpAMD64MOVWQZX
OpAMD64MOVLQSX
OpAMD64MOVLQZX
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
OpAMD64MOVLconst
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpAMD64MOVQconst
[dev.ssa] cmd/compile: truncate when converting floats to ints Modified to use the truncating conversion. Fixes reflect. Change-Id: I47bf3200abc2d2c662939a2a2351e2ff84168f4a Reviewed-on: https://go-review.googlesource.com/14167 Reviewed-by: David Chase <drchase@google.com>
2015-09-01 19:05:44 -05:00
OpAMD64CVTTSD2SL
OpAMD64CVTTSD2SQ
OpAMD64CVTTSS2SL
OpAMD64CVTTSS2SQ
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
OpAMD64CVTSL2SS
OpAMD64CVTSL2SD
OpAMD64CVTSQ2SS
OpAMD64CVTSQ2SD
OpAMD64CVTSD2SS
OpAMD64CVTSS2SD
[dev.ssa] cmd/compile: add complex arithmetic Still to do: details, more testing corner cases. (e.g. negative zero) Includes small cleanups for previous CL. Note: complex division is currently done in the runtime, so the division code here is apparently not yet necessary and also not tested. Seems likely better to open code division and expose the widening/narrowing to optimization. Complex64 multiplication and division is done in wide format to avoid cancellation errors; for division, this also happens to be compatible with pre-SSA practice (which uses a single complex128 division function). It would-be-nice to widen for complex128 multiplication intermediates as well, but that is trickier to implement without a handy wider-precision format. Change-Id: I595a4300f68868fb7641852a54674c6b2b78855e Reviewed-on: https://go-review.googlesource.com/14028 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-28 14:24:10 -04:00
OpAMD64PXOR
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpAMD64LEAQ
[dev.ssa] cmd/compile/internal/ssa: Handle variables correctly Use *Node of type ONAME instead of string as the key for variable maps. This will prevent aliasing between two identically named but differently scoped variables. Introduce an Aux value that encodes the offset of a variable from a base pointer (either global base pointer or stack pointer). Allow LEAQ and derivatives (MOVQ, etc.) to also have such an Aux field. Allocate space for AUTO variables in stackalloc. Change-Id: Ibdccdaea4bbc63a1f4882959ac374f2b467e3acd Reviewed-on: https://go-review.googlesource.com/11238 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-06-19 21:02:28 -07:00
OpAMD64LEAQ1
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpAMD64LEAQ2
OpAMD64LEAQ4
OpAMD64LEAQ8
OpAMD64MOVBload
OpAMD64MOVBQSXload
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
OpAMD64MOVWload
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
OpAMD64MOVWQSXload
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
OpAMD64MOVLload
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
OpAMD64MOVLQSXload
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpAMD64MOVQload
OpAMD64MOVBstore
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
OpAMD64MOVWstore
OpAMD64MOVLstore
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpAMD64MOVQstore
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
OpAMD64MOVOload
OpAMD64MOVOstore
OpAMD64MOVBloadidx1
cmd/compile: fix plan9-amd64 build The previous rules to combine indexed loads produced addresses like: From: obj.Addr{ Type: TYPE_MEM, Reg: REG_CX, Name: NAME_AUTO, Offset: 121, ... } which are erroneous because NAME_AUTO implies a base register of REG_SP, and cmd/internal/obj/x86 makes many assumptions to this effect. Note that previously we were also producing an extra "ADDQ SP, CX" instruction, so indexing off of SP was already handled. The approach taken by this CL to address the problem is to instead produce addresses like: From: obj.Addr{ Type: TYPE_MEM, Reg: REG_SP, Name: NAME_AUTO, Offset: 121, Index: REG_CX, Scale: 1, } and to omit the "ADDQ SP, CX" instruction. Downside to this approach is it requires adding a lot of new MOV[WLQ]loadidx1 instructions that nearly duplicate functionality of the existing MOV[WLQ]loadidx[248] instructions, but with a different Scale. Fixes #15001. Change-Id: Iad9a1a41e5e2552f8d22e3ba975e4ea0862dffd2 Reviewed-on: https://go-review.googlesource.com/21245 Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
2016-03-28 19:10:13 -07:00
OpAMD64MOVWloadidx1
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
OpAMD64MOVWloadidx2
cmd/compile: fix plan9-amd64 build The previous rules to combine indexed loads produced addresses like: From: obj.Addr{ Type: TYPE_MEM, Reg: REG_CX, Name: NAME_AUTO, Offset: 121, ... } which are erroneous because NAME_AUTO implies a base register of REG_SP, and cmd/internal/obj/x86 makes many assumptions to this effect. Note that previously we were also producing an extra "ADDQ SP, CX" instruction, so indexing off of SP was already handled. The approach taken by this CL to address the problem is to instead produce addresses like: From: obj.Addr{ Type: TYPE_MEM, Reg: REG_SP, Name: NAME_AUTO, Offset: 121, Index: REG_CX, Scale: 1, } and to omit the "ADDQ SP, CX" instruction. Downside to this approach is it requires adding a lot of new MOV[WLQ]loadidx1 instructions that nearly duplicate functionality of the existing MOV[WLQ]loadidx[248] instructions, but with a different Scale. Fixes #15001. Change-Id: Iad9a1a41e5e2552f8d22e3ba975e4ea0862dffd2 Reviewed-on: https://go-review.googlesource.com/21245 Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
2016-03-28 19:10:13 -07:00
OpAMD64MOVLloadidx1
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
OpAMD64MOVLloadidx4
cmd/compile: fix plan9-amd64 build The previous rules to combine indexed loads produced addresses like: From: obj.Addr{ Type: TYPE_MEM, Reg: REG_CX, Name: NAME_AUTO, Offset: 121, ... } which are erroneous because NAME_AUTO implies a base register of REG_SP, and cmd/internal/obj/x86 makes many assumptions to this effect. Note that previously we were also producing an extra "ADDQ SP, CX" instruction, so indexing off of SP was already handled. The approach taken by this CL to address the problem is to instead produce addresses like: From: obj.Addr{ Type: TYPE_MEM, Reg: REG_SP, Name: NAME_AUTO, Offset: 121, Index: REG_CX, Scale: 1, } and to omit the "ADDQ SP, CX" instruction. Downside to this approach is it requires adding a lot of new MOV[WLQ]loadidx1 instructions that nearly duplicate functionality of the existing MOV[WLQ]loadidx[248] instructions, but with a different Scale. Fixes #15001. Change-Id: Iad9a1a41e5e2552f8d22e3ba975e4ea0862dffd2 Reviewed-on: https://go-review.googlesource.com/21245 Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
2016-03-28 19:10:13 -07:00
OpAMD64MOVQloadidx1
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
OpAMD64MOVQloadidx8
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
OpAMD64MOVBstoreidx1
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
OpAMD64MOVWstoreidx1
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
OpAMD64MOVWstoreidx2
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
OpAMD64MOVLstoreidx1
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
OpAMD64MOVLstoreidx4
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
OpAMD64MOVQstoreidx1
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpAMD64MOVQstoreidx8
[dev.ssa] cmd/compile: introduce storeconst ops Introduce opcodes that store a constant value. AuxInt now needs to hold both the value to be stored and the constant offset at which to store it. Introduce a StoreConst type to help encode/decode these parts to/from an AuxInt. Change-Id: I1631883abe035cff4b16368683e1eb3d2ccb674d Reviewed-on: https://go-review.googlesource.com/16170 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-10-21 13:13:56 -07:00
OpAMD64MOVBstoreconst
OpAMD64MOVWstoreconst
OpAMD64MOVLstoreconst
OpAMD64MOVQstoreconst
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
OpAMD64MOVBstoreconstidx1
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
OpAMD64MOVWstoreconstidx1
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
OpAMD64MOVWstoreconstidx2
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
OpAMD64MOVLstoreconstidx1
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
OpAMD64MOVLstoreconstidx4
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
OpAMD64MOVQstoreconstidx1
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
OpAMD64MOVQstoreconstidx8
[dev.ssa] cmd/compile: clean up zeroing. Use duffzero when appropriate. Change-Id: I4deb03340e87f43179d5e22bf81843c17b5581fc Reviewed-on: https://go-review.googlesource.com/14756 Reviewed-by: David Chase <drchase@google.com>
2015-09-18 18:23:34 -07:00
OpAMD64DUFFZERO
[dev.ssa] Merge remote-tracking branch 'origin/master' into mergebranch The only major fixup is that duffzero changed from 8-byte writes to 16-byte writes. Change-Id: I1762b74ce67a8e4b81c11568027cdb3572f7f87c
2015-10-19 13:56:55 -07:00
OpAMD64MOVOconst
[dev.ssa] cmd/compile/internal/ssa: Initial implementation of memory zeroing This will make it possible for us to start implementing interfaces and other stack allocated types which are more than one machine word. Change-Id: I52b187a791cf1919cb70ed6dabdc9f57b317ea83 Reviewed-on: https://go-review.googlesource.com/11631 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-27 15:45:20 +01:00
OpAMD64REPSTOSQ
[dev.ssa] cmd/compile/internal/ssa: add call opcodes Add calls, particularly closure calls. Reorg SSAable variable test for converting to SSA. Change-Id: Ia75c04295e6b0b040122f97e2381836a393b7f42 Reviewed-on: https://go-review.googlesource.com/10912 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-06-10 15:03:06 -07:00
OpAMD64CALLstatic
OpAMD64CALLclosure
[dev.ssa] cmd/compile/internal/gc: implement go and defer TODO: for now, just function calls. Do method and interface calls. Change-Id: Ib262dfa31cb753996cde899beaad4dc2e66705ac Reviewed-on: https://go-review.googlesource.com/14035 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-28 22:51:01 -07:00
OpAMD64CALLdefer
OpAMD64CALLgo
[dev.ssa] cmd/compile/internal/ssa: complete call ops OCALLINTER, as well as ODEFER/OPROC with OCALLMETH/OCALLINTER. Move all the call logic to its own routine, a lot of the code is shared. Change-Id: Ieac59596165e434cc6d1d7b5e46b78957e9c5ed3 Reviewed-on: https://go-review.googlesource.com/14464 Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2015-09-09 23:56:59 -07:00
OpAMD64CALLinter
[dev.ssa] cmd/compile: better copying Replace REP MOVSB with all the copying techniques used by the old compiler. Copy in chunks, DUFFCOPY, etc. Introduces MOVO opcodes and an Int128 type to move around 16 bytes at a time. Change-Id: I1e73e68ca1d8b3dd58bb4af2f4c9e5d9bf13a502 Reviewed-on: https://go-review.googlesource.com/16174 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2015-10-21 17:18:07 -07:00
OpAMD64DUFFCOPY
OpAMD64REPMOVSQ
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpAMD64InvertFlags
[dev.ssa] cmd/compile: implement OGETG Change-Id: I7ecf62cf399c710b4a617803c43e83fce09b8a7d Reviewed-on: https://go-review.googlesource.com/13585 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 11:22:16 -07:00
OpAMD64LoweredGetG
[dev.ssa] cmd/compile: addressed vars and closures Cleaned up first-block-in-function code. Added cases for |PHEAP for PPARAM and PAUTO. Made PPARAMOUT act more like PAUTO for purposes of address generation and vardef placement. Added cases for OCLOSUREVAR and Ops for getting closure pointer. Closure ops are scheduled at top of entry block to capture DX. Wrote test that seems to show proper behavior for addressed parameters, locals, and returns. Change-Id: Iee93ebf9e3d9f74cfb4d1c1da8038eb278d8a857 Reviewed-on: https://go-review.googlesource.com/14650 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com>
2015-09-11 16:40:05 -04:00
OpAMD64LoweredGetClosurePtr
[dev.ssa] cmd/compile: optimize nil checks Use faulting loads instead of test/jeq to do nil checks. Fold nil checks into a following load/store if possible. Makes binaries about 2% smaller. Change-Id: I54af0f0a93c853f37e34e0ce7e3f01dd2ac87f64 Reviewed-on: https://go-review.googlesource.com/16287 Reviewed-by: David Chase <drchase@google.com>
2015-10-23 19:12:49 -07:00
OpAMD64LoweredNilCheck
[dev.ssa] cmd/compile: be safer about uintptr/unsafe.Pointer conversions Make sure that when a pointer value is live across a function call, we save it as a pointer. (And similarly a uintptr live across a function call should not be saved as a pointer.) Add a nasty test case. This is probably what is preventing the merge from master to dev.ssa. Signs point to something like this bug happening in mallocgc. Change-Id: Ib23fa1251b8d1c50d82c6a448cb4a4fc28219029 Reviewed-on: https://go-review.googlesource.com/16830 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: David Chase <drchase@google.com>
2015-11-10 15:35:36 -08:00
OpAMD64MOVQconvert
[dev.ssa] cmd/compile: clean up comparisons Add new constant-flags opcodes. These can be generated from comparisons that we know the result of, like x&31 < 32. Constant-fold the constant-flags opcodes into all flag users. Reorder some CMPxconst args so they read in the comparison direction. Reorg deadcode removal a bit - it needs to remove the OpCopy ops it generates when strength-reducing Phi ops. So it needs to splice out all the dead blocks and do a copy elimination before it computes live values. Change-Id: Ie922602033592ad8212efe4345394973d3b94d9f Reviewed-on: https://go-review.googlesource.com/18267 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-05 14:56:26 -08:00
OpAMD64FlagEQ
OpAMD64FlagLT_ULT
OpAMD64FlagLT_UGT
OpAMD64FlagGT_UGT
OpAMD64FlagGT_ULT
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
OpARMADD
OpARMADDconst
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
OpARMSUB
OpARMSUBconst
OpARMRSB
OpARMRSBconst
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
OpARMMUL
OpARMHMUL
OpARMHMULU
[dev.ssa] cmd/compile: handle Div, Convert, GetClosurePtr etc. on ARM This CL adds support of Div, Mod, Convert, GetClosurePtr and 64-bit indexing support to SSA backend for ARM. Add tests for 64-bit indexing to cmd/compile/internal/gc/testdata/string_ssa.go. Tests cmd/compile/internal/gc/testdata/*_ssa.go passed, except compound_ssa.go and fp_ssa.go. Progress on SSA for ARM. Still not complete. Essentially the only unsupported part is floating point. Updates #15365. Change-Id: I269e88b67f641c25e7a813d910c96d356d236bff Reviewed-on: https://go-review.googlesource.com/23542 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 09:49:28 -04:00
OpARMDIV
OpARMDIVU
OpARMMOD
OpARMMODU
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
OpARMADDS
OpARMADC
OpARMSUBS
OpARMSBC
OpARMMULLU
OpARMMULA
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
OpARMADDF
OpARMADDD
OpARMSUBF
OpARMSUBD
OpARMMULF
OpARMMULD
OpARMDIVF
OpARMDIVD
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
OpARMAND
OpARMANDconst
OpARMOR
OpARMORconst
OpARMXOR
OpARMXORconst
OpARMBIC
OpARMBICconst
[dev.ssa] cmd/compile: handle boolean values for SSA on ARM Fix hardcoded flag register mask in ssa/flagalloc.go by auto-generating the mask. Also fix a mistake (in previous CL) about conditional branches. Progress on SSA backend for ARM. Still not complete. Now "container/ring" package compiles and tests passed. Updates #15365. Change-Id: Id7c8805c30dbb8107baedb485ed0f71f59ed6ea8 Reviewed-on: https://go-review.googlesource.com/23093 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 11:25:07 -04:00
OpARMMVN
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
OpARMSQRTD
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
OpARMSLL
OpARMSLLconst
OpARMSRL
OpARMSRLconst
OpARMSRA
OpARMSRAconst
[dev.ssa] cmd/compile: handle 64-bit shifts on ARM Also fix a mistake in previous CL about x8 and x16 shifts: the shift needs ZeroExt. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: Ibc352760023d38bc6b9c5251e929fe26e016637a Reviewed-on: https://go-review.googlesource.com/23486 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 23:17:42 -04:00
OpARMSRRconst
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
OpARMCMP
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
OpARMCMPconst
OpARMCMN
OpARMCMNconst
OpARMTST
OpARMTSTconst
OpARMTEQ
OpARMTEQconst
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
OpARMCMPF
OpARMCMPD
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
OpARMMOVWconst
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
OpARMMOVFconst
OpARMMOVDconst
[dev.ssa] cmd/compile: use MOVWaddr for address on ARM Introduce an op MOVWaddr for addresses on ARM, instead of overuse ADDconst. Mark MOVWaddr as rematerializable. This fixes a liveness problem: if it were not rematerializable, the address of a variable may be spilled and later use of the address may just load the spilled value without mentioning the variable, and the liveness code may think it is dead prematurely. Update #15365. Change-Id: Ib0b0fa826bdb75c9e6bb362b95c6cf132cc6b1c0 Reviewed-on: https://go-review.googlesource.com/23942 Reviewed-by: David Chase <drchase@google.com>
2016-06-06 22:36:45 -04:00
OpARMMOVWaddr
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
OpARMMOVBload
OpARMMOVBUload
OpARMMOVHload
OpARMMOVHUload
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
OpARMMOVWload
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
OpARMMOVFload
OpARMMOVDload
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
OpARMMOVBstore
OpARMMOVHstore
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
OpARMMOVWstore
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
OpARMMOVFstore
OpARMMOVDstore
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
OpARMMOVBreg
OpARMMOVBUreg
OpARMMOVHreg
OpARMMOVHUreg
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
OpARMMOVWF
OpARMMOVWD
OpARMMOVWUF
OpARMMOVWUD
OpARMMOVFW
OpARMMOVDW
OpARMMOVFWU
OpARMMOVDWU
OpARMMOVFD
OpARMMOVDF
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
OpARMCALLstatic
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
OpARMCALLclosure
OpARMCALLdefer
OpARMCALLgo
OpARMCALLinter
OpARMLoweredNilCheck
OpARMEqual
OpARMNotEqual
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
OpARMLessThan
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
OpARMLessEqual
OpARMGreaterThan
OpARMGreaterEqual
OpARMLessThanU
OpARMLessEqualU
OpARMGreaterThanU
OpARMGreaterEqualU
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
OpARMCarry
OpARMLoweredSelect0
OpARMLoweredSelect1
[dev.ssa] cmd/compile: handle 64-bit shifts on ARM Also fix a mistake in previous CL about x8 and x16 shifts: the shift needs ZeroExt. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: Ibc352760023d38bc6b9c5251e929fe26e016637a Reviewed-on: https://go-review.googlesource.com/23486 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 23:17:42 -04:00
OpARMLoweredZeromask
[dev.ssa] cmd/compile: implement Zero, Move, Copy for SSA on ARM Generate load/stores for small zeroing/move, DUFFZERO/DUFFCOPY for medium zeroing/move, and loops for large zeroing/move. cmd/compile/internal/gc/testdata/{copy_ssa.go,zero_ssa.go} tests passed. Progress on SSA backend for ARM. Still not complete. A few packages in the standard library compile and tests passed, including container/list, hash/crc32, unicode/utf8, etc. Updates #15365. Change-Id: Ieb4b68b44ee7de66bf7b68f5f33a605349fcc6fa Reviewed-on: https://go-review.googlesource.com/23097 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:31:14 -04:00
OpARMDUFFZERO
OpARMDUFFCOPY
OpARMLoweredZero
OpARMLoweredMove
[dev.ssa] cmd/compile: handle Div, Convert, GetClosurePtr etc. on ARM This CL adds support of Div, Mod, Convert, GetClosurePtr and 64-bit indexing support to SSA backend for ARM. Add tests for 64-bit indexing to cmd/compile/internal/gc/testdata/string_ssa.go. Tests cmd/compile/internal/gc/testdata/*_ssa.go passed, except compound_ssa.go and fp_ssa.go. Progress on SSA for ARM. Still not complete. Essentially the only unsupported part is floating point. Updates #15365. Change-Id: I269e88b67f641c25e7a813d910c96d356d236bff Reviewed-on: https://go-review.googlesource.com/23542 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 09:49:28 -04:00
OpARMLoweredGetClosurePtr
OpARMMOVWconvert
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
OpAdd8
OpAdd16
OpAdd32
OpAdd64
OpAddPtr
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
OpAdd32F
OpAdd64F
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
OpSub8
OpSub16
OpSub32
OpSub64
[dev.ssa] cmd/compile: implement OSLICESTR Add a new function and generic operation to handle bounds checking for slices. Unlike the index bounds checking the index can be equal to the upper bound. Do gc-friendly slicing that generates proper code for 0-length result slices. This is a takeover of Alexandru's original change, (https://go-review.googlesource.com/#/c/12764/) submittable now that the decompose phase is in. Change-Id: I17d164cf42ed7839f84ca949c6ad3289269c9160 Reviewed-on: https://go-review.googlesource.com/13903 Reviewed-by: David Chase <drchase@google.com>
2015-08-24 23:52:03 -07:00
OpSubPtr
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
OpSub32F
OpSub64F
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
OpMul8
OpMul16
OpMul32
OpMul64
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
OpMul32F
OpMul64F
OpDiv32F
OpDiv64F
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
OpHmul8
OpHmul8u
OpHmul16
OpHmul16u
OpHmul32
OpHmul32u
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
OpHmul64
OpHmul64u
OpAvg64u
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
OpDiv8
OpDiv8u
OpDiv16
OpDiv16u
OpDiv32
OpDiv32u
OpDiv64
OpDiv64u
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
OpMod8
OpMod8u
OpMod16
OpMod16u
OpMod32
OpMod32u
OpMod64
OpMod64u
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
OpAnd8
OpAnd16
OpAnd32
OpAnd64
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
OpOr8
OpOr16
OpOr32
OpOr64
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
OpXor8
OpXor16
OpXor32
OpXor64
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
OpLsh8x8
OpLsh8x16
OpLsh8x32
OpLsh8x64
OpLsh16x8
OpLsh16x16
OpLsh16x32
OpLsh16x64
OpLsh32x8
OpLsh32x16
OpLsh32x32
OpLsh32x64
OpLsh64x8
OpLsh64x16
OpLsh64x32
OpLsh64x64
OpRsh8x8
OpRsh8x16
OpRsh8x32
OpRsh8x64
OpRsh16x8
OpRsh16x16
OpRsh16x32
OpRsh16x64
OpRsh32x8
OpRsh32x16
OpRsh32x32
OpRsh32x64
OpRsh64x8
OpRsh64x16
OpRsh64x32
OpRsh64x64
OpRsh8Ux8
OpRsh8Ux16
OpRsh8Ux32
OpRsh8Ux64
OpRsh16Ux8
OpRsh16Ux16
OpRsh16Ux32
OpRsh16Ux64
OpRsh32Ux8
OpRsh32Ux16
OpRsh32Ux32
OpRsh32Ux64
OpRsh64Ux8
OpRsh64Ux16
OpRsh64Ux32
OpRsh64Ux64
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
OpLrot8
OpLrot16
OpLrot32
OpLrot64
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
OpEq8
OpEq16
OpEq32
OpEq64
[dev.ssa] cmd/compile: implement non-numeric comparisons The only slice/interface comparisons that reach the backend are comparisons to nil. Funcs, maps, and channels are references types, so pointer equality is enough. Change-Id: I60a71da46a36202e9bd62ed370ab7d7f2e2800e7 Reviewed-on: https://go-review.googlesource.com/12715 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-27 13:17:45 -07:00
OpEqPtr
[dev.ssa] cmd/compile/internal/ssa: fix iface and slice comparisons A simpler way to do iface/slice comparisons. Fixes some cases of failed lowerings. Change-Id: Ia252bc8648293a2d460f63c41f1591785543a1e9 Reviewed-on: https://go-review.googlesource.com/14493 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-09-10 13:53:27 -07:00
OpEqInter
OpEqSlice
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
OpEq32F
OpEq64F
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
OpNeq8
OpNeq16
OpNeq32
OpNeq64
[dev.ssa] cmd/compile: implement non-numeric comparisons The only slice/interface comparisons that reach the backend are comparisons to nil. Funcs, maps, and channels are references types, so pointer equality is enough. Change-Id: I60a71da46a36202e9bd62ed370ab7d7f2e2800e7 Reviewed-on: https://go-review.googlesource.com/12715 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-27 13:17:45 -07:00
OpNeqPtr
[dev.ssa] cmd/compile/internal/ssa: fix iface and slice comparisons A simpler way to do iface/slice comparisons. Fixes some cases of failed lowerings. Change-Id: Ia252bc8648293a2d460f63c41f1591785543a1e9 Reviewed-on: https://go-review.googlesource.com/14493 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-09-10 13:53:27 -07:00
OpNeqInter
OpNeqSlice
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
OpNeq32F
OpNeq64F
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
OpLess8
OpLess8U
OpLess16
OpLess16U
OpLess32
OpLess32U
OpLess64
OpLess64U
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
OpLess32F
OpLess64F
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
OpLeq8
OpLeq8U
OpLeq16
OpLeq16U
OpLeq32
OpLeq32U
OpLeq64
OpLeq64U
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
OpLeq32F
OpLeq64F
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
OpGreater8
OpGreater8U
OpGreater16
OpGreater16U
OpGreater32
OpGreater32U
OpGreater64
OpGreater64U
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
OpGreater32F
OpGreater64F
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
OpGeq8
OpGeq8U
OpGeq16
OpGeq16U
OpGeq32
OpGeq32U
OpGeq64
OpGeq64U
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
OpGeq32F
OpGeq64F
cmd/compile: introduce bool operations. Introduce OrB, EqB, NeqB, AndB to handle bool operations. Change-Id: I53e4d5125a8090d5eeb4576db619103f19fff58d Reviewed-on: https://go-review.googlesource.com/22412 Reviewed-by: Keith Randall <khr@golang.org>
2016-04-24 21:21:07 +02:00
OpAndB
OpOrB
OpEqB
OpNeqB
[dev.ssa] cmd/compile: implement ONOT Co-hacking with josharian at Gophercon. Change-Id: Ia59dfab676c6ed598c2c25483439cd1395a4ea87 Reviewed-on: https://go-review.googlesource.com/12029 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
2015-07-10 11:25:48 -06:00
OpNot
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
OpNeg8
OpNeg16
OpNeg32
OpNeg64
[dev.ssa] cmd/compile: add complex arithmetic Still to do: details, more testing corner cases. (e.g. negative zero) Includes small cleanups for previous CL. Note: complex division is currently done in the runtime, so the division code here is apparently not yet necessary and also not tested. Seems likely better to open code division and expose the widening/narrowing to optimization. Complex64 multiplication and division is done in wide format to avoid cancellation errors; for division, this also happens to be compatible with pre-SSA practice (which uses a single complex128 division function). It would-be-nice to widen for complex128 multiplication intermediates as well, but that is trickier to implement without a handy wider-precision format. Change-Id: I595a4300f68868fb7641852a54674c6b2b78855e Reviewed-on: https://go-review.googlesource.com/14028 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-28 14:24:10 -04:00
OpNeg32F
OpNeg64F
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
OpCom8
OpCom16
OpCom32
OpCom64
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
OpCtz16
OpCtz32
OpCtz64
OpClz16
OpClz32
OpClz64
OpBswap32
OpBswap64
[dev.ssa] cmd/compile/internal/ssa: implement OSQRT Change-Id: Iec61ca1bdc064c29ceca6d47f600d5643d0a64dd Reviewed-on: https://go-review.googlesource.com/14533 Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
2015-09-12 13:26:57 -07:00
OpSqrt
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpPhi
OpCopy
[dev.ssa] cmd/compile: enhance SSA filtering, add OpConvert Modified GOSSA{HASH.PKG} environment variable filters to make it easier to make/run with all SSA for testing. Disable attempts at SSA for architectures that are not amd64 (avoid spurious errors/unimplementeds.) Removed easy out for unimplemented features. Add convert op for proper liveness in presence of uintptr to/from unsafe.Pointer conversions. Tweaked stack sizes to get a pass on windows; 1024 instead 768, was observed to pass at least once. Change-Id: Ida3800afcda67d529e3b1cf48ca4a3f0fa48b2c5 Reviewed-on: https://go-review.googlesource.com/16201 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com>
2015-10-19 11:36:07 -04:00
OpConvert
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
OpConstBool
OpConstString
OpConstNil
OpConst8
OpConst16
OpConst32
OpConst64
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
OpConst32F
OpConst64F
[dev.ssa] cmd/compile: add decompose pass Decompose breaks compound objects up into pieces that can be operated on by the target architecture. The decompose pass only does phi ops, the rest is done by the rewrite rules in generic.rules. Compound objects include strings,slices,interfaces,structs,arrays. Arrays aren't decomposed because of indexing (we could support constant indexes, but dynamic indexes can't be handled using SSA). Structs will come in a subsequent CL. TODO: after this pass we have lost the association between, e.g., a string's pointer and its size. It would be nice if we could keep that information around for debugging info somehow. Change-Id: I6379ab962a7beef62297d0f68c421f22aa0a0901 Reviewed-on: https://go-review.googlesource.com/13683 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-18 10:26:28 -07:00
OpConstInterface
OpConstSlice
[dev.ssa] cmd/compile: start arguments as spilled Declare a function's arguments as having already been spilled so their use just requires a restore. Allow spill locations to be portions of larger objects the stack. Required to load portions of compound input arguments. Rename the memory input to InputMem. Use Arg for the pre-spilled argument values. Change-Id: I8fe2a03ffbba1022d98bfae2052b376b96d32dda Reviewed-on: https://go-review.googlesource.com/16536 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: David Chase <drchase@google.com>
2015-11-02 08:10:26 -08:00
OpInitMem
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpArg
[dev.ssa] cmd/compile/internal/ssa: Handle variables correctly Use *Node of type ONAME instead of string as the key for variable maps. This will prevent aliasing between two identically named but differently scoped variables. Introduce an Aux value that encodes the offset of a variable from a base pointer (either global base pointer or stack pointer). Allow LEAQ and derivatives (MOVQ, etc.) to also have such an Aux field. Allocate space for AUTO variables in stackalloc. Change-Id: Ibdccdaea4bbc63a1f4882959ac374f2b467e3acd Reviewed-on: https://go-review.googlesource.com/11238 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-06-19 21:02:28 -07:00
OpAddr
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpSP
[dev.ssa] cmd/compile/internal/ssa: Handle variables correctly Use *Node of type ONAME instead of string as the key for variable maps. This will prevent aliasing between two identically named but differently scoped variables. Introduce an Aux value that encodes the offset of a variable from a base pointer (either global base pointer or stack pointer). Allow LEAQ and derivatives (MOVQ, etc.) to also have such an Aux field. Allocate space for AUTO variables in stackalloc. Change-Id: Ibdccdaea4bbc63a1f4882959ac374f2b467e3acd Reviewed-on: https://go-review.googlesource.com/11238 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-06-19 21:02:28 -07:00
OpSB
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpFunc
OpLoad
OpStore
OpMove
[dev.ssa] cmd/compile/internal/ssa: Initial implementation of memory zeroing This will make it possible for us to start implementing interfaces and other stack allocated types which are more than one machine word. Change-Id: I52b187a791cf1919cb70ed6dabdc9f57b317ea83 Reviewed-on: https://go-review.googlesource.com/11631 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-27 15:45:20 +01:00
OpZero
[dev.ssa] cmd/compile/internal/ssa: add call opcodes Add calls, particularly closure calls. Reorg SSAable variable test for converting to SSA. Change-Id: Ia75c04295e6b0b040122f97e2381836a393b7f42 Reviewed-on: https://go-review.googlesource.com/10912 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-06-10 15:03:06 -07:00
OpClosureCall
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpStaticCall
[dev.ssa] cmd/compile/internal/gc: implement go and defer TODO: for now, just function calls. Do method and interface calls. Change-Id: Ib262dfa31cb753996cde899beaad4dc2e66705ac Reviewed-on: https://go-review.googlesource.com/14035 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-28 22:51:01 -07:00
OpDeferCall
OpGoCall
[dev.ssa] cmd/compile/internal/ssa: complete call ops OCALLINTER, as well as ODEFER/OPROC with OCALLMETH/OCALLINTER. Move all the call logic to its own routine, a lot of the code is shared. Change-Id: Ieac59596165e434cc6d1d7b5e46b78957e9c5ed3 Reviewed-on: https://go-review.googlesource.com/14464 Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2015-09-09 23:56:59 -07:00
OpInterCall
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
OpSignExt8to16
OpSignExt8to32
OpSignExt8to64
OpSignExt16to32
OpSignExt16to64
OpSignExt32to64
OpZeroExt8to16
OpZeroExt8to32
OpZeroExt8to64
OpZeroExt16to32
OpZeroExt16to64
OpZeroExt32to64
OpTrunc16to8
OpTrunc32to8
OpTrunc32to16
OpTrunc64to8
OpTrunc64to16
OpTrunc64to32
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
OpCvt32to32F
OpCvt32to64F
OpCvt64to32F
OpCvt64to64F
OpCvt32Fto32
OpCvt32Fto64
OpCvt64Fto32
OpCvt64Fto64
OpCvt32Fto64F
OpCvt64Fto32F
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpIsNonNil
OpIsInBounds
[dev.ssa] cmd/compile: implement OSLICESTR Add a new function and generic operation to handle bounds checking for slices. Unlike the index bounds checking the index can be equal to the upper bound. Do gc-friendly slicing that generates proper code for 0-length result slices. This is a takeover of Alexandru's original change, (https://go-review.googlesource.com/#/c/12764/) submittable now that the decompose phase is in. Change-Id: I17d164cf42ed7839f84ca949c6ad3289269c9160 Reviewed-on: https://go-review.googlesource.com/13903 Reviewed-by: David Chase <drchase@google.com>
2015-08-24 23:52:03 -07:00
OpIsSliceInBounds
[dev.ssa] cmd/compile: optimize nil checks Use faulting loads instead of test/jeq to do nil checks. Fold nil checks into a following load/store if possible. Makes binaries about 2% smaller. Change-Id: I54af0f0a93c853f37e34e0ce7e3f01dd2ac87f64 Reviewed-on: https://go-review.googlesource.com/16287 Reviewed-by: David Chase <drchase@google.com>
2015-10-23 19:12:49 -07:00
OpNilCheck
[dev.ssa] cmd/compile: implement OGETG Change-Id: I7ecf62cf399c710b4a617803c43e83fce09b8a7d Reviewed-on: https://go-review.googlesource.com/13585 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 11:22:16 -07:00
OpGetG
[dev.ssa] cmd/compile: addressed vars and closures Cleaned up first-block-in-function code. Added cases for |PHEAP for PPARAM and PAUTO. Made PPARAMOUT act more like PAUTO for purposes of address generation and vardef placement. Added cases for OCLOSUREVAR and Ops for getting closure pointer. Closure ops are scheduled at top of entry block to capture DX. Wrote test that seems to show proper behavior for addressed parameters, locals, and returns. Change-Id: Iee93ebf9e3d9f74cfb4d1c1da8038eb278d8a857 Reviewed-on: https://go-review.googlesource.com/14650 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com>
2015-09-11 16:40:05 -04:00
OpGetClosurePtr
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpArrayIndex
OpPtrIndex
OpOffPtr
OpSliceMake
OpSlicePtr
OpSliceLen
OpSliceCap
[dev.ssa] cmd/compile: add compose/decompose for complex, phi, constants Still to do: arithmetic Change-Id: I31fd23b34980c9ed4b4e304b8597134b2ba6ca5c Reviewed-on: https://go-review.googlesource.com/14024 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-28 14:24:10 -04:00
OpComplexMake
OpComplexReal
OpComplexImag
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpStringMake
OpStringPtr
OpStringLen
[dev.ssa] cmd/compile: add decompose pass Decompose breaks compound objects up into pieces that can be operated on by the target architecture. The decompose pass only does phi ops, the rest is done by the rewrite rules in generic.rules. Compound objects include strings,slices,interfaces,structs,arrays. Arrays aren't decomposed because of indexing (we could support constant indexes, but dynamic indexes can't be handled using SSA). Structs will come in a subsequent CL. TODO: after this pass we have lost the association between, e.g., a string's pointer and its size. It would be nice if we could keep that information around for debugging info somehow. Change-Id: I6379ab962a7beef62297d0f68c421f22aa0a0901 Reviewed-on: https://go-review.googlesource.com/13683 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-18 10:26:28 -07:00
OpIMake
[dev.ssa] cmd/compile/internal/ssa: implement ITAB Implement ITAB, selecting the itable field of an interface. Soften the lowering check to allow lowerings that leave generic but dead ops behind. (The ITAB lowering does this.) Change-Id: Icc84961dd4060d143602f001311aa1d8be0d7fc0 Reviewed-on: https://go-review.googlesource.com/13144 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-04 15:47:22 -07:00
OpITab
[dev.ssa] cmd/compile: add decompose pass Decompose breaks compound objects up into pieces that can be operated on by the target architecture. The decompose pass only does phi ops, the rest is done by the rewrite rules in generic.rules. Compound objects include strings,slices,interfaces,structs,arrays. Arrays aren't decomposed because of indexing (we could support constant indexes, but dynamic indexes can't be handled using SSA). Structs will come in a subsequent CL. TODO: after this pass we have lost the association between, e.g., a string's pointer and its size. It would be nice if we could keep that information around for debugging info somehow. Change-Id: I6379ab962a7beef62297d0f68c421f22aa0a0901 Reviewed-on: https://go-review.googlesource.com/13683 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-18 10:26:28 -07:00
OpIData
[dev.ssa] cmd/compile: Allow structs to be SSAd Break small structs up into their components so they can be registerized. Change StructSelect to use field indexes instead of field offsets, as field offsets aren't unique in the presence of zero-sized fields. Change-Id: I2f1dc89f7fa58e1cf58aa1a32b238959d53f62e4 Reviewed-on: https://go-review.googlesource.com/18570 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-11 21:05:33 -08:00
OpStructMake0
OpStructMake1
OpStructMake2
OpStructMake3
OpStructMake4
OpStructSelect
[dev.ssa] cmd/compile: respect stack slot width when storing/loading registers Prior to this, we were smashing our own stack, which caused the crypto/sha256 tests to fail. Change-Id: I7dd94cf466d175b3be0cd65f9c4fe8b1223081fe Reviewed-on: https://go-review.googlesource.com/12660 Reviewed-by: Daniel Morsing <daniel.morsing@gmail.com> Reviewed-by: Keith Randall <khr@golang.org>
2015-07-24 14:51:51 -07:00
OpStoreReg
OpLoadReg
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
OpFwdRef
[dev.ssa] cmd/compile: optimize phi ops Redo how we keep track of forward references when building SSA. When the forward reference is resolved, update the Value node in place. Improve the phi elimination pass so it can simplify phis of phis. Give SSA package access to decoded line numbers. Fix line numbers for constant booleans. Change-Id: I3dc9896148d260be2f3dd14cbe5db639ec9fa6b7 Reviewed-on: https://go-review.googlesource.com/18674 Reviewed-by: David Chase <drchase@google.com> Run-TryBot: Keith Randall <khr@golang.org>
2016-01-14 16:02:23 -08:00
OpUnknown
[dev.ssa] cmd/runtime: generate gc bitmaps for SSA-compiled code This change is all about leveraging the gc bitmap generation that is already done by the current compiler. We rearrange how stack allocation is done so that we generate a variable declaration for each spill. We also reorganize how args/locals are recorded during SSA. Then we can use the existing allocauto/defframe to allocate the stack frame and liveness to make the gc bitmaps. With this change, stack copying works correctly and we no longer need hacks in runtime/stack*.go to make tests work. GC is close to working, it just needs write barriers. Change-Id: I990fb4e3fbe98850c6be35c3185a1c85d9e1a6ba Reviewed-on: https://go-review.googlesource.com/13894 Reviewed-by: David Chase <drchase@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-24 02:16:19 -07:00
OpVarDef
OpVarKill
[dev.ssa] Merge remote-tracking branch 'origin/master' into mergebranch Semi-regular merge from tip to dev.ssa. Conflicts: src/runtime/sys_windows_amd64.s Change-Id: I5f733130049c810e6ceacd46dad85faebca52b29
2016-01-19 09:59:21 -08:00
OpVarLive
cmd/compile: keep pointer input arguments live throughout function Introduce a KeepAlive op which makes sure that its argument is kept live until the KeepAlive. Use KeepAlive to mark pointer input arguments as live after each function call and at each return. We do this change only for pointer arguments. Those are the critical ones to handle because they might have finalizers. Doing compound arguments (slices, structs, ...) is more complicated because we would need to track field liveness individually (we do that for auto variables now, but inputs requires extra trickery). Turn off the automatic marking of args as live. That way, when args are explicitly nulled, plive will know that the original argument is dead. The KeepAlive op will be the eventual implementation of runtime.KeepAlive. Fixes #15277 Change-Id: I5f223e65d99c9f8342c03fbb1512c4d363e903e5 Reviewed-on: https://go-review.googlesource.com/22365 Reviewed-by: David Chase <drchase@google.com> Reviewed-by: Russ Cox <rsc@golang.org>
2016-04-21 19:28:28 -07:00
OpKeepAlive
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
OpInt64Make
OpInt64Hi
OpInt64Lo
OpAdd32carry
OpAdd32withcarry
OpSub32carry
OpSub32withcarry
OpMul32uhilo
OpSignmask
[dev.ssa] cmd/compile: handle 64-bit shifts on ARM Also fix a mistake in previous CL about x8 and x16 shifts: the shift needs ZeroExt. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: Ibc352760023d38bc6b9c5251e929fe26e016637a Reviewed-on: https://go-review.googlesource.com/23486 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 23:17:42 -04:00
OpZeromask
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
OpCvt32Uto32F
OpCvt32Uto64F
OpCvt32Fto32U
OpCvt64Fto32U
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
OpSelect0
OpSelect1
[dev.ssa] cmd/compile: Add initial SSA configuration for PPC64 This adds the initial SSA implementation for PPC64. Builds golang and all.bash runs correctly. Simple hello.go builds but does not run. Change-Id: I7cec211b934cd7a2dd75a6cdfaf9f71867063466 Reviewed-on: https://go-review.googlesource.com/24453 Reviewed-by: David Chase <drchase@google.com> Reviewed-by: Cherry Zhang <cherryyz@google.com> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-06-24 14:37:17 -05:00
OpPPC64ADD
OpPPC64ADDconst
OpPPC64FADD
OpPPC64FADDS
OpPPC64SUB
OpPPC64FSUB
OpPPC64FSUBS
OpPPC64MULLD
OpPPC64MULLW
OpPPC64FMUL
OpPPC64FMULS
OpPPC64FDIV
OpPPC64FDIVS
OpPPC64AND
OpPPC64ANDconst
OpPPC64OR
OpPPC64ORconst
OpPPC64XOR
OpPPC64XORconst
OpPPC64NEG
OpPPC64MOVBreg
OpPPC64MOVBZreg
OpPPC64MOVHreg
OpPPC64MOVHZreg
OpPPC64MOVWreg
OpPPC64MOVWZreg
OpPPC64MOVBload
OpPPC64MOVBZload
OpPPC64MOVHload
OpPPC64MOVHZload
OpPPC64MOVWload
OpPPC64MOVWZload
OpPPC64MOVDload
OpPPC64FMOVDload
OpPPC64FMOVSload
OpPPC64MOVBstore
OpPPC64MOVHstore
OpPPC64MOVWstore
OpPPC64MOVDstore
OpPPC64FMOVDstore
OpPPC64FMOVSstore
OpPPC64MOVBstoreconst
OpPPC64MOVHstoreconst
OpPPC64MOVWstoreconst
OpPPC64MOVDstoreconst
OpPPC64MOVDconst
OpPPC64MOVWconst
OpPPC64MOVHconst
OpPPC64MOVBconst
OpPPC64FMOVDconst
OpPPC64FMOVSconst
OpPPC64FCMPU
OpPPC64CMP
OpPPC64CMPU
OpPPC64CMPW
OpPPC64CMPWU
OpPPC64CMPconst
OpPPC64CALLstatic
OpPPC64Equal
OpPPC64NotEqual
OpPPC64LessThan
OpPPC64LessEqual
OpPPC64GreaterThan
OpPPC64GreaterEqual
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
)
var opcodeTable = [...]opInfo{
{name: "OpInvalid"},
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "ADDSS",
argLen: 2,
commutative: true,
resultInArg0: true,
asm: x86.AADDSS,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
{1, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "ADDSD",
argLen: 2,
commutative: true,
resultInArg0: true,
asm: x86.AADDSD,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
{1, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SUBSS",
argLen: 2,
resultInArg0: true,
asm: x86.ASUBSS,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 2147418112}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
{1, 2147418112}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 2147483648, // X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
2147418112, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SUBSD",
argLen: 2,
resultInArg0: true,
asm: x86.ASUBSD,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 2147418112}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
{1, 2147418112}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 2147483648, // X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
2147418112, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "MULSS",
argLen: 2,
commutative: true,
resultInArg0: true,
asm: x86.AMULSS,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
{1, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "MULSD",
argLen: 2,
commutative: true,
resultInArg0: true,
asm: x86.AMULSD,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
{1, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "DIVSS",
argLen: 2,
resultInArg0: true,
asm: x86.ADIVSS,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 2147418112}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
{1, 2147418112}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 2147483648, // X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
2147418112, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "DIVSD",
argLen: 2,
resultInArg0: true,
asm: x86.ADIVSD,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 2147418112}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
{1, 2147418112}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 2147483648, // X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
2147418112, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVSSload",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVSS,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVSDload",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVSD,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVSSconst",
cmd/compile: const folding for float32/64 Split the auxFloat type into 32/64 bit versions and perform checking for exactly representable float32 values. Perform const folding on float32/64. Comment out some const negation rules that the frontend already performs. Change-Id: Ib3f8d59fa8b30e50fe0267786cfb3c50a06169d2 Reviewed-on: https://go-review.googlesource.com/20568 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
2016-03-11 19:36:54 -06:00
auxType: auxFloat32,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
rematerializeable: true,
asm: x86.AMOVSS,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVSDconst",
cmd/compile: const folding for float32/64 Split the auxFloat type into 32/64 bit versions and perform checking for exactly representable float32 values. Perform const folding on float32/64. Comment out some const negation rules that the frontend already performs. Change-Id: Ib3f8d59fa8b30e50fe0267786cfb3c50a06169d2 Reviewed-on: https://go-review.googlesource.com/20568 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
2016-03-11 19:36:54 -06:00
auxType: auxFloat64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
rematerializeable: true,
asm: x86.AMOVSD,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
{
name: "MOVSSloadidx1",
auxType: auxSymOff,
argLen: 3,
asm: x86.AMOVSS,
reg: regInfo{
inputs: []inputInfo{
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
},
outputs: []regMask{
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
},
},
},
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVSSloadidx4",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVSS,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
{
name: "MOVSDloadidx1",
auxType: auxSymOff,
argLen: 3,
asm: x86.AMOVSD,
reg: regInfo{
inputs: []inputInfo{
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
},
outputs: []regMask{
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
},
},
},
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVSDloadidx8",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVSD,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVSSstore",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVSS,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVSDstore",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVSD,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
{
name: "MOVSSstoreidx1",
auxType: auxSymOff,
argLen: 4,
asm: x86.AMOVSS,
reg: regInfo{
inputs: []inputInfo{
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{2, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
},
},
},
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVSSstoreidx4",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 4,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVSS,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{2, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
{
name: "MOVSDstoreidx1",
auxType: auxSymOff,
argLen: 4,
asm: x86.AMOVSD,
reg: regInfo{
inputs: []inputInfo{
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{2, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
},
},
},
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVSDstoreidx8",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 4,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVSD,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{2, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
name: "ADDQ",
argLen: 2,
commutative: true,
asm: x86.AADDQ,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{1, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
name: "ADDL",
argLen: 2,
commutative: true,
asm: x86.AADDL,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{1, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
name: "ADDQconst",
auxType: auxInt64,
argLen: 1,
asm: x86.AADDQ,
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
},
},
},
{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
name: "ADDLconst",
auxType: auxInt32,
argLen: 1,
asm: x86.AADDL,
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
},
},
},
[dev.ssa] cmd/compile: implement ONOT Co-hacking with josharian at Gophercon. Change-Id: Ia59dfab676c6ed598c2c25483439cd1395a4ea87 Reviewed-on: https://go-review.googlesource.com/12029 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
2015-07-10 11:25:48 -06:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SUBQ",
argLen: 2,
resultInArg0: true,
asm: x86.ASUBQ,
[dev.ssa] cmd/compile: implement ONOT Co-hacking with josharian at Gophercon. Change-Id: Ia59dfab676c6ed598c2c25483439cd1395a4ea87 Reviewed-on: https://go-review.googlesource.com/12029 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
2015-07-10 11:25:48 -06:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: implement ONOT Co-hacking with josharian at Gophercon. Change-Id: Ia59dfab676c6ed598c2c25483439cd1395a4ea87 Reviewed-on: https://go-review.googlesource.com/12029 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
2015-07-10 11:25:48 -06:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile: implement ONOT Co-hacking with josharian at Gophercon. Change-Id: Ia59dfab676c6ed598c2c25483439cd1395a4ea87 Reviewed-on: https://go-review.googlesource.com/12029 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
2015-07-10 11:25:48 -06:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: implement ONOT Co-hacking with josharian at Gophercon. Change-Id: Ia59dfab676c6ed598c2c25483439cd1395a4ea87 Reviewed-on: https://go-review.googlesource.com/12029 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
2015-07-10 11:25:48 -06:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SUBL",
argLen: 2,
resultInArg0: true,
asm: x86.ASUBL,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa/gen: implement all bit sizes for Eq and Neq Change-Id: I2c5c75153493b5dd3f1a743e5edf04403e83b31b Reviewed-on: https://go-review.googlesource.com/12474 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-21 18:06:15 +02:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SUBQconst",
auxType: auxInt64,
argLen: 1,
resultInArg0: true,
asm: x86.ASUBQ,
[dev.ssa] cmd/compile/internal/ssa/gen: implement all bit sizes for Eq and Neq Change-Id: I2c5c75153493b5dd3f1a743e5edf04403e83b31b Reviewed-on: https://go-review.googlesource.com/12474 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-21 18:06:15 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement all bit sizes for Eq and Neq Change-Id: I2c5c75153493b5dd3f1a743e5edf04403e83b31b Reviewed-on: https://go-review.googlesource.com/12474 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-21 18:06:15 +02:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa/gen: implement all bit sizes for Eq and Neq Change-Id: I2c5c75153493b5dd3f1a743e5edf04403e83b31b Reviewed-on: https://go-review.googlesource.com/12474 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-21 18:06:15 +02:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement all bit sizes for Eq and Neq Change-Id: I2c5c75153493b5dd3f1a743e5edf04403e83b31b Reviewed-on: https://go-review.googlesource.com/12474 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-21 18:06:15 +02:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SUBLconst",
auxType: auxInt32,
argLen: 1,
resultInArg0: true,
asm: x86.ASUBL,
[dev.ssa] cmd/compile/internal/ssa/gen: implement all bit sizes for Eq and Neq Change-Id: I2c5c75153493b5dd3f1a743e5edf04403e83b31b Reviewed-on: https://go-review.googlesource.com/12474 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-21 18:06:15 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement all bit sizes for Eq and Neq Change-Id: I2c5c75153493b5dd3f1a743e5edf04403e83b31b Reviewed-on: https://go-review.googlesource.com/12474 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-21 18:06:15 +02:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa/gen: implement all bit sizes for Eq and Neq Change-Id: I2c5c75153493b5dd3f1a743e5edf04403e83b31b Reviewed-on: https://go-review.googlesource.com/12474 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-21 18:06:15 +02:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement all bit sizes for Eq and Neq Change-Id: I2c5c75153493b5dd3f1a743e5edf04403e83b31b Reviewed-on: https://go-review.googlesource.com/12474 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-21 18:06:15 +02:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "MULQ",
argLen: 2,
commutative: true,
resultInArg0: true,
asm: x86.AIMULQ,
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "MULL",
argLen: 2,
commutative: true,
resultInArg0: true,
asm: x86.AIMULL,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "MULQconst",
auxType: auxInt64,
argLen: 1,
resultInArg0: true,
asm: x86.AIMULQ,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
[dev.ssa] cmd/compile/ssa: add comparison ops Increase SSA coverage of functions in the standard library from 20.79% to 27.81%. The most significant unimplemented items are now: 10.16% 2597 SSA unimplemented: zero for type error not implemented 8.44% 2157 SSA unimplemented: addr: bad op DOTPTR 7.98% 2039 SSA unimplemented: unhandled OLITERAL 7 6.29% 1607 SSA unimplemented: unhandled expr OROR 4.73% 1209 SSA unimplemented: unhandled expr LEN 4.55% 1163 SSA unimplemented: unhandled expr LROT 3.42% 874 SSA unimplemented: unhandled OLITERAL 6 2.46% 629 SSA unimplemented: unhandled expr DOT 2.41% 615 SSA unimplemented: zero for type []byte not implemented 2.02% 516 SSA unimplemented: unhandled stmt CALLMETH 1.90% 486 SSA unimplemented: unhandled expr ANDAND 1.79% 458 SSA unimplemented: unhandled expr CALLINTER 1.69% 433 SSA unimplemented: unhandled stmt SWITCH 1.67% 428 SSA unimplemented: unhandled expr CALLMETH 1.67% 426 SSA unimplemented: unhandled expr CLOSUREVAR Change-Id: I40959b22993c4f70784b4eca472cae752347879c Reviewed-on: https://go-review.googlesource.com/11452 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-24 17:48:22 -07:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "MULLconst",
auxType: auxInt32,
argLen: 1,
resultInArg0: true,
asm: x86.AIMULL,
[dev.ssa] cmd/compile/ssa: add comparison ops Increase SSA coverage of functions in the standard library from 20.79% to 27.81%. The most significant unimplemented items are now: 10.16% 2597 SSA unimplemented: zero for type error not implemented 8.44% 2157 SSA unimplemented: addr: bad op DOTPTR 7.98% 2039 SSA unimplemented: unhandled OLITERAL 7 6.29% 1607 SSA unimplemented: unhandled expr OROR 4.73% 1209 SSA unimplemented: unhandled expr LEN 4.55% 1163 SSA unimplemented: unhandled expr LROT 3.42% 874 SSA unimplemented: unhandled OLITERAL 6 2.46% 629 SSA unimplemented: unhandled expr DOT 2.41% 615 SSA unimplemented: zero for type []byte not implemented 2.02% 516 SSA unimplemented: unhandled stmt CALLMETH 1.90% 486 SSA unimplemented: unhandled expr ANDAND 1.79% 458 SSA unimplemented: unhandled expr CALLINTER 1.69% 433 SSA unimplemented: unhandled stmt SWITCH 1.67% 428 SSA unimplemented: unhandled expr CALLMETH 1.67% 426 SSA unimplemented: unhandled expr CLOSUREVAR Change-Id: I40959b22993c4f70784b4eca472cae752347879c Reviewed-on: https://go-review.googlesource.com/11452 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-24 17:48:22 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/ssa: add comparison ops Increase SSA coverage of functions in the standard library from 20.79% to 27.81%. The most significant unimplemented items are now: 10.16% 2597 SSA unimplemented: zero for type error not implemented 8.44% 2157 SSA unimplemented: addr: bad op DOTPTR 7.98% 2039 SSA unimplemented: unhandled OLITERAL 7 6.29% 1607 SSA unimplemented: unhandled expr OROR 4.73% 1209 SSA unimplemented: unhandled expr LEN 4.55% 1163 SSA unimplemented: unhandled expr LROT 3.42% 874 SSA unimplemented: unhandled OLITERAL 6 2.46% 629 SSA unimplemented: unhandled expr DOT 2.41% 615 SSA unimplemented: zero for type []byte not implemented 2.02% 516 SSA unimplemented: unhandled stmt CALLMETH 1.90% 486 SSA unimplemented: unhandled expr ANDAND 1.79% 458 SSA unimplemented: unhandled expr CALLINTER 1.69% 433 SSA unimplemented: unhandled stmt SWITCH 1.67% 428 SSA unimplemented: unhandled expr CALLMETH 1.67% 426 SSA unimplemented: unhandled expr CLOSUREVAR Change-Id: I40959b22993c4f70784b4eca472cae752347879c Reviewed-on: https://go-review.googlesource.com/11452 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-24 17:48:22 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/ssa: add comparison ops Increase SSA coverage of functions in the standard library from 20.79% to 27.81%. The most significant unimplemented items are now: 10.16% 2597 SSA unimplemented: zero for type error not implemented 8.44% 2157 SSA unimplemented: addr: bad op DOTPTR 7.98% 2039 SSA unimplemented: unhandled OLITERAL 7 6.29% 1607 SSA unimplemented: unhandled expr OROR 4.73% 1209 SSA unimplemented: unhandled expr LEN 4.55% 1163 SSA unimplemented: unhandled expr LROT 3.42% 874 SSA unimplemented: unhandled OLITERAL 6 2.46% 629 SSA unimplemented: unhandled expr DOT 2.41% 615 SSA unimplemented: zero for type []byte not implemented 2.02% 516 SSA unimplemented: unhandled stmt CALLMETH 1.90% 486 SSA unimplemented: unhandled expr ANDAND 1.79% 458 SSA unimplemented: unhandled expr CALLINTER 1.69% 433 SSA unimplemented: unhandled stmt SWITCH 1.67% 428 SSA unimplemented: unhandled expr CALLMETH 1.67% 426 SSA unimplemented: unhandled expr CLOSUREVAR Change-Id: I40959b22993c4f70784b4eca472cae752347879c Reviewed-on: https://go-review.googlesource.com/11452 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-24 17:48:22 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/ssa: add comparison ops Increase SSA coverage of functions in the standard library from 20.79% to 27.81%. The most significant unimplemented items are now: 10.16% 2597 SSA unimplemented: zero for type error not implemented 8.44% 2157 SSA unimplemented: addr: bad op DOTPTR 7.98% 2039 SSA unimplemented: unhandled OLITERAL 7 6.29% 1607 SSA unimplemented: unhandled expr OROR 4.73% 1209 SSA unimplemented: unhandled expr LEN 4.55% 1163 SSA unimplemented: unhandled expr LROT 3.42% 874 SSA unimplemented: unhandled OLITERAL 6 2.46% 629 SSA unimplemented: unhandled expr DOT 2.41% 615 SSA unimplemented: zero for type []byte not implemented 2.02% 516 SSA unimplemented: unhandled stmt CALLMETH 1.90% 486 SSA unimplemented: unhandled expr ANDAND 1.79% 458 SSA unimplemented: unhandled expr CALLINTER 1.69% 433 SSA unimplemented: unhandled stmt SWITCH 1.67% 428 SSA unimplemented: unhandled expr CALLMETH 1.67% 426 SSA unimplemented: unhandled expr CLOSUREVAR Change-Id: I40959b22993c4f70784b4eca472cae752347879c Reviewed-on: https://go-review.googlesource.com/11452 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-24 17:48:22 -07:00
},
},
},
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "HMULQ",
argLen: 2,
asm: x86.AIMULQ,
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "HMULL",
argLen: 2,
asm: x86.AIMULL,
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "HMULW",
argLen: 2,
asm: x86.AIMULW,
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "HMULB",
argLen: 2,
asm: x86.AIMULB,
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
},
},
},
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "HMULQU",
argLen: 2,
asm: x86.AMULQ,
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "HMULLU",
argLen: 2,
asm: x86.AMULL,
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "HMULWU",
argLen: 2,
asm: x86.AMULW,
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "HMULBU",
argLen: 2,
asm: x86.AMULB,
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
},
},
},
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "AVGQU",
argLen: 2,
commutative: true,
resultInArg0: true,
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "DIVQ",
argLen: 2,
asm: x86.AIDIVQ,
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65531}, // AX CX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934596, // DX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
1, // AX
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "DIVL",
argLen: 2,
asm: x86.AIDIVL,
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65531}, // AX CX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934596, // DX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
1, // AX
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "DIVW",
argLen: 2,
asm: x86.AIDIVW,
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65531}, // AX CX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934596, // DX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
1, // AX
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "DIVQU",
argLen: 2,
asm: x86.ADIVQ,
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65531}, // AX CX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934596, // DX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
1, // AX
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "DIVLU",
argLen: 2,
asm: x86.ADIVL,
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65531}, // AX CX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934596, // DX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
1, // AX
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "DIVWU",
argLen: 2,
asm: x86.ADIVW,
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65531}, // AX CX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934596, // DX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
1, // AX
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "MODQ",
argLen: 2,
asm: x86.AIDIVQ,
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65531}, // AX CX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "MODL",
argLen: 2,
asm: x86.AIDIVL,
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65531}, // AX CX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "MODW",
argLen: 2,
asm: x86.AIDIVW,
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65531}, // AX CX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "MODQU",
argLen: 2,
asm: x86.ADIVQ,
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65531}, // AX CX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "MODLU",
argLen: 2,
asm: x86.ADIVL,
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65531}, // AX CX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "MODWU",
argLen: 2,
asm: x86.ADIVW,
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 1}, // AX
{1, 65531}, // AX CX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "ANDQ",
argLen: 2,
commutative: true,
resultInArg0: true,
asm: x86.AANDQ,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "ANDL",
argLen: 2,
commutative: true,
resultInArg0: true,
asm: x86.AANDL,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: finish implementing comparisons Change-Id: I4e496c7c7239111133631f76ca25e14be64800c6 Reviewed-on: https://go-review.googlesource.com/12656 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-24 12:47:00 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "ANDQconst",
auxType: auxInt64,
argLen: 1,
resultInArg0: true,
asm: x86.AANDQ,
[dev.ssa] cmd/compile: finish implementing comparisons Change-Id: I4e496c7c7239111133631f76ca25e14be64800c6 Reviewed-on: https://go-review.googlesource.com/12656 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-24 12:47:00 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: finish implementing comparisons Change-Id: I4e496c7c7239111133631f76ca25e14be64800c6 Reviewed-on: https://go-review.googlesource.com/12656 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-24 12:47:00 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile: finish implementing comparisons Change-Id: I4e496c7c7239111133631f76ca25e14be64800c6 Reviewed-on: https://go-review.googlesource.com/12656 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-24 12:47:00 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "ANDLconst",
auxType: auxInt32,
argLen: 1,
resultInArg0: true,
asm: x86.AANDL,
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: Complete 64-bit shifts Implement correct Go shifts. Allow multi-line rewrite rules. Fix offset & alignment in stack alloc. Change-Id: I0ae9e522c83df9205bbe4ab94bc0e43d16dace58 Reviewed-on: https://go-review.googlesource.com/10891 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-10 10:39:57 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "ORQ",
argLen: 2,
commutative: true,
resultInArg0: true,
asm: x86.AORQ,
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "ORL",
argLen: 2,
commutative: true,
resultInArg0: true,
asm: x86.AORL,
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "ORQconst",
auxType: auxInt64,
argLen: 1,
resultInArg0: true,
asm: x86.AORQ,
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "ORLconst",
auxType: auxInt32,
argLen: 1,
resultInArg0: true,
asm: x86.AORL,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "XORQ",
argLen: 2,
commutative: true,
resultInArg0: true,
asm: x86.AXORQ,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "XORL",
argLen: 2,
commutative: true,
resultInArg0: true,
asm: x86.AXORL,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "XORQconst",
auxType: auxInt64,
argLen: 1,
resultInArg0: true,
asm: x86.AXORQ,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "XORLconst",
auxType: auxInt32,
argLen: 1,
resultInArg0: true,
asm: x86.AXORL,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "CMPQ",
argLen: 2,
asm: x86.ACMPQ,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "CMPL",
argLen: 2,
asm: x86.ACMPL,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "CMPW",
argLen: 2,
asm: x86.ACMPW,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "CMPB",
argLen: 2,
asm: x86.ACMPB,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "CMPQconst",
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.ACMPQ,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "CMPLconst",
auxType: auxInt32,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.ACMPL,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "CMPWconst",
auxType: auxInt16,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.ACMPW,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "CMPBconst",
auxType: auxInt8,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.ACMPB,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "UCOMISS",
argLen: 2,
asm: x86.AUCOMISS,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
{1, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "UCOMISD",
argLen: 2,
asm: x86.AUCOMISD,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
{1, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "TESTQ",
argLen: 2,
asm: x86.ATESTQ,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "TESTL",
argLen: 2,
asm: x86.ATESTL,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "TESTW",
argLen: 2,
asm: x86.ATESTW,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "TESTB",
argLen: 2,
asm: x86.ATESTB,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "TESTQconst",
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.ATESTQ,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "TESTLconst",
auxType: auxInt32,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.ATESTL,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "TESTWconst",
auxType: auxInt16,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.ATESTW,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "TESTBconst",
auxType: auxInt8,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.ATESTB,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SHLQ",
argLen: 2,
resultInArg0: true,
asm: x86.ASHLQ,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 2}, // CX
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
outputs: []regMask{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SHLL",
argLen: 2,
resultInArg0: true,
asm: x86.ASHLL,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 2}, // CX
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
outputs: []regMask{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SHLQconst",
auxType: auxInt64,
argLen: 1,
resultInArg0: true,
asm: x86.ASHLQ,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SHLLconst",
auxType: auxInt32,
argLen: 1,
resultInArg0: true,
asm: x86.ASHLL,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SHRQ",
argLen: 2,
resultInArg0: true,
asm: x86.ASHRQ,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 2}, // CX
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
outputs: []regMask{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SHRL",
argLen: 2,
resultInArg0: true,
asm: x86.ASHRL,
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 2}, // CX
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
outputs: []regMask{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SHRW",
argLen: 2,
resultInArg0: true,
asm: x86.ASHRW,
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 2}, // CX
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
outputs: []regMask{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SHRB",
argLen: 2,
resultInArg0: true,
asm: x86.ASHRB,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 2}, // CX
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SHRQconst",
auxType: auxInt64,
argLen: 1,
resultInArg0: true,
asm: x86.ASHRQ,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SHRLconst",
auxType: auxInt32,
argLen: 1,
resultInArg0: true,
asm: x86.ASHRL,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SHRWconst",
auxType: auxInt16,
argLen: 1,
resultInArg0: true,
asm: x86.ASHRW,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SHRBconst",
auxType: auxInt8,
argLen: 1,
resultInArg0: true,
asm: x86.ASHRB,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SARQ",
argLen: 2,
resultInArg0: true,
asm: x86.ASARQ,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 2}, // CX
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: Handle variables correctly Use *Node of type ONAME instead of string as the key for variable maps. This will prevent aliasing between two identically named but differently scoped variables. Introduce an Aux value that encodes the offset of a variable from a base pointer (either global base pointer or stack pointer). Allow LEAQ and derivatives (MOVQ, etc.) to also have such an Aux field. Allocate space for AUTO variables in stackalloc. Change-Id: Ibdccdaea4bbc63a1f4882959ac374f2b467e3acd Reviewed-on: https://go-review.googlesource.com/11238 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-06-19 21:02:28 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SARL",
argLen: 2,
resultInArg0: true,
asm: x86.ASARL,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 2}, // CX
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SARW",
argLen: 2,
resultInArg0: true,
asm: x86.ASARW,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 2}, // CX
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SARB",
argLen: 2,
resultInArg0: true,
asm: x86.ASARB,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 2}, // CX
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SARQconst",
auxType: auxInt64,
argLen: 1,
resultInArg0: true,
asm: x86.ASARQ,
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SARLconst",
auxType: auxInt32,
argLen: 1,
resultInArg0: true,
asm: x86.ASARL,
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SARWconst",
auxType: auxInt16,
argLen: 1,
resultInArg0: true,
asm: x86.ASARW,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "SARBconst",
auxType: auxInt8,
argLen: 1,
resultInArg0: true,
asm: x86.ASARB,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "ROLQconst",
auxType: auxInt64,
argLen: 1,
resultInArg0: true,
asm: x86.AROLQ,
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "ROLLconst",
auxType: auxInt32,
argLen: 1,
resultInArg0: true,
asm: x86.AROLL,
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "ROLWconst",
auxType: auxInt16,
argLen: 1,
resultInArg0: true,
asm: x86.AROLW,
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "ROLBconst",
auxType: auxInt8,
argLen: 1,
resultInArg0: true,
asm: x86.AROLB,
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "NEGQ",
argLen: 1,
resultInArg0: true,
asm: x86.ANEGQ,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "NEGL",
argLen: 1,
resultInArg0: true,
asm: x86.ANEGL,
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "NOTQ",
argLen: 1,
resultInArg0: true,
asm: x86.ANOTQ,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
},
},
},
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "NOTL",
argLen: 1,
resultInArg0: true,
asm: x86.ANOTL,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
},
},
},
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
{
name: "BSFQ",
argLen: 1,
asm: x86.ABSFQ,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
},
clobbers: 8589934592, // FLAGS
outputs: []regMask{
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
{
name: "BSFL",
argLen: 1,
asm: x86.ABSFL,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
},
clobbers: 8589934592, // FLAGS
outputs: []regMask{
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
{
name: "BSFW",
argLen: 1,
asm: x86.ABSFW,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
},
clobbers: 8589934592, // FLAGS
outputs: []regMask{
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
{
name: "BSRQ",
argLen: 1,
asm: x86.ABSRQ,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
},
clobbers: 8589934592, // FLAGS
outputs: []regMask{
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
{
name: "BSRL",
argLen: 1,
asm: x86.ABSRL,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
},
clobbers: 8589934592, // FLAGS
outputs: []regMask{
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
{
name: "BSRW",
argLen: 1,
asm: x86.ABSRW,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
},
clobbers: 8589934592, // FLAGS
outputs: []regMask{
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
{
name: "CMOVQEQconst",
auxType: auxInt64,
argLen: 2,
resultInArg0: true,
asm: x86.ACMOVQEQ,
reg: regInfo{
inputs: []inputInfo{
{1, 8589934592}, // FLAGS
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65518}, // CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
},
clobbers: 8589934593, // AX FLAGS
outputs: []regMask{
65518, // CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
{
name: "CMOVLEQconst",
auxType: auxInt32,
argLen: 2,
resultInArg0: true,
asm: x86.ACMOVLEQ,
reg: regInfo{
inputs: []inputInfo{
{1, 8589934592}, // FLAGS
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65518}, // CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
},
clobbers: 8589934593, // AX FLAGS
outputs: []regMask{
65518, // CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
{
name: "CMOVWEQconst",
auxType: auxInt16,
argLen: 2,
resultInArg0: true,
asm: x86.ACMOVLEQ,
reg: regInfo{
inputs: []inputInfo{
{1, 8589934592}, // FLAGS
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65518}, // CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
},
clobbers: 8589934593, // AX FLAGS
outputs: []regMask{
65518, // CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
{
name: "CMOVQNEconst",
auxType: auxInt64,
argLen: 2,
resultInArg0: true,
asm: x86.ACMOVQNE,
reg: regInfo{
inputs: []inputInfo{
{1, 8589934592}, // FLAGS
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65518}, // CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
},
clobbers: 8589934593, // AX FLAGS
outputs: []regMask{
65518, // CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
{
name: "CMOVLNEconst",
auxType: auxInt32,
argLen: 2,
resultInArg0: true,
asm: x86.ACMOVLNE,
reg: regInfo{
inputs: []inputInfo{
{1, 8589934592}, // FLAGS
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65518}, // CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
},
clobbers: 8589934593, // AX FLAGS
outputs: []regMask{
65518, // CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
{
name: "CMOVWNEconst",
auxType: auxInt16,
argLen: 2,
resultInArg0: true,
asm: x86.ACMOVLNE,
reg: regInfo{
inputs: []inputInfo{
{1, 8589934592}, // FLAGS
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65518}, // CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
},
clobbers: 8589934593, // AX FLAGS
outputs: []regMask{
65518, // CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
{
name: "BSWAPQ",
argLen: 1,
resultInArg0: true,
asm: x86.ABSWAPQ,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
},
clobbers: 8589934592, // FLAGS
outputs: []regMask{
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
{
name: "BSWAPL",
argLen: 1,
resultInArg0: true,
asm: x86.ABSWAPL,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: regalloc enforces 2-address instructions Instead of being a hint, resultInArg0 is now enforced by regalloc. This allows us to delete all the code from amd64/ssa.go which deals with converting from a semantically three-address instruction into some copies plus a two-address instruction. Change-Id: Id4f39a80be4b678718bfd42a229f9094ab6ecd7c Reviewed-on: https://go-review.googlesource.com/21816 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-04-10 08:26:43 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
},
clobbers: 8589934592, // FLAGS
outputs: []regMask{
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
[dev.ssa] cmd/compile/internal/ssa: implement OSQRT Change-Id: Iec61ca1bdc064c29ceca6d47f600d5643d0a64dd Reviewed-on: https://go-review.googlesource.com/14533 Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
2015-09-12 13:26:57 -07:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SQRTSD",
argLen: 1,
asm: x86.ASQRTSD,
[dev.ssa] cmd/compile/internal/ssa: implement OSQRT Change-Id: Iec61ca1bdc064c29ceca6d47f600d5643d0a64dd Reviewed-on: https://go-review.googlesource.com/14533 Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
2015-09-12 13:26:57 -07:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile/internal/ssa: implement OSQRT Change-Id: Iec61ca1bdc064c29ceca6d47f600d5643d0a64dd Reviewed-on: https://go-review.googlesource.com/14533 Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
2015-09-12 13:26:57 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile/internal/ssa: implement OSQRT Change-Id: Iec61ca1bdc064c29ceca6d47f600d5643d0a64dd Reviewed-on: https://go-review.googlesource.com/14533 Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
2015-09-12 13:26:57 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SBBQcarrymask",
argLen: 1,
asm: x86.ASBBQ,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SBBLcarrymask",
argLen: 1,
asm: x86.ASBBL,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: Initial implementation of memory zeroing This will make it possible for us to start implementing interfaces and other stack allocated types which are more than one machine word. Change-Id: I52b187a791cf1919cb70ed6dabdc9f57b317ea83 Reviewed-on: https://go-review.googlesource.com/11631 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-27 15:45:20 +01:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETEQ",
argLen: 1,
asm: x86.ASETEQ,
[dev.ssa] cmd/compile/internal/ssa: Initial implementation of memory zeroing This will make it possible for us to start implementing interfaces and other stack allocated types which are more than one machine word. Change-Id: I52b187a791cf1919cb70ed6dabdc9f57b317ea83 Reviewed-on: https://go-review.googlesource.com/11631 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-27 15:45:20 +01:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: Initial implementation of memory zeroing This will make it possible for us to start implementing interfaces and other stack allocated types which are more than one machine word. Change-Id: I52b187a791cf1919cb70ed6dabdc9f57b317ea83 Reviewed-on: https://go-review.googlesource.com/11631 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-27 15:45:20 +01:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETNE",
argLen: 1,
asm: x86.ASETNE,
[dev.ssa] cmd/compile/internal/ssa: Initial implementation of memory zeroing This will make it possible for us to start implementing interfaces and other stack allocated types which are more than one machine word. Change-Id: I52b187a791cf1919cb70ed6dabdc9f57b317ea83 Reviewed-on: https://go-review.googlesource.com/11631 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-27 15:45:20 +01:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: Initial implementation of memory zeroing This will make it possible for us to start implementing interfaces and other stack allocated types which are more than one machine word. Change-Id: I52b187a791cf1919cb70ed6dabdc9f57b317ea83 Reviewed-on: https://go-review.googlesource.com/11631 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-27 15:45:20 +01:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETL",
argLen: 1,
asm: x86.ASETLT,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETLE",
argLen: 1,
asm: x86.ASETLE,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
[dev.ssa] cmd/compile/internal/ssa: add call opcodes Add calls, particularly closure calls. Reorg SSAable variable test for converting to SSA. Change-Id: Ia75c04295e6b0b040122f97e2381836a393b7f42 Reviewed-on: https://go-review.googlesource.com/10912 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-06-10 15:03:06 -07:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETG",
argLen: 1,
asm: x86.ASETGT,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
[dev.ssa] cmd/compile/internal/ssa: add call opcodes Add calls, particularly closure calls. Reorg SSAable variable test for converting to SSA. Change-Id: Ia75c04295e6b0b040122f97e2381836a393b7f42 Reviewed-on: https://go-review.googlesource.com/10912 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-06-10 15:03:06 -07:00
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETGE",
argLen: 1,
asm: x86.ASETGE,
[dev.ssa] cmd/compile/internal/ssa: add call opcodes Add calls, particularly closure calls. Reorg SSAable variable test for converting to SSA. Change-Id: Ia75c04295e6b0b040122f97e2381836a393b7f42 Reviewed-on: https://go-review.googlesource.com/10912 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-06-10 15:03:06 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: add call opcodes Add calls, particularly closure calls. Reorg SSAable variable test for converting to SSA. Change-Id: Ia75c04295e6b0b040122f97e2381836a393b7f42 Reviewed-on: https://go-review.googlesource.com/10912 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-06-10 15:03:06 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETB",
argLen: 1,
asm: x86.ASETCS,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETBE",
argLen: 1,
asm: x86.ASETLS,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETA",
argLen: 1,
asm: x86.ASETHI,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETAE",
argLen: 1,
asm: x86.ASETCC,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETEQF",
argLen: 1,
asm: x86.ASETEQ,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65518, // CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETNEF",
argLen: 1,
asm: x86.ASETNE,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934593, // AX FLAGS
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65518, // CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETORD",
argLen: 1,
asm: x86.ASETPC,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETNAN",
argLen: 1,
asm: x86.ASETPS,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETGF",
argLen: 1,
asm: x86.ASETHI,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "SETGEF",
argLen: 1,
asm: x86.ASETCC,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 8589934592}, // FLAGS
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "MOVBQSX",
argLen: 1,
asm: x86.AMOVBQSX,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "MOVBQZX",
argLen: 1,
asm: x86.AMOVBQZX,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "MOVWQSX",
argLen: 1,
asm: x86.AMOVWQSX,
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "MOVWQZX",
argLen: 1,
asm: x86.AMOVWQZX,
[dev.ssa] cmd/compile/internal/ssa: adds for 8,16,32 bit ints Change-Id: I33025a4a41fd91f6ee317d33a6eebf27fa00ab51 Reviewed-on: https://go-review.googlesource.com/11115 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-14 11:38:46 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "MOVLQSX",
argLen: 1,
asm: x86.AMOVLQSX,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "MOVLQZX",
argLen: 1,
asm: x86.AMOVLQZX,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVLconst",
auxType: auxInt32,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
rematerializeable: true,
asm: x86.AMOVL,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
},
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVQconst",
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
rematerializeable: true,
asm: x86.AMOVQ,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
},
},
},
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "CVTTSD2SL",
argLen: 1,
asm: x86.ACVTTSD2SL,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "CVTTSD2SQ",
argLen: 1,
asm: x86.ACVTTSD2SQ,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "CVTTSS2SL",
argLen: 1,
asm: x86.ACVTTSS2SL,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "CVTTSS2SQ",
argLen: 1,
asm: x86.ACVTTSS2SQ,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "CVTSL2SS",
argLen: 1,
asm: x86.ACVTSL2SS,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "CVTSL2SD",
argLen: 1,
asm: x86.ACVTSL2SD,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "CVTSQ2SS",
argLen: 1,
asm: x86.ACVTSQ2SS,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "CVTSQ2SD",
argLen: 1,
asm: x86.ACVTSQ2SD,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "CVTSD2SS",
argLen: 1,
asm: x86.ACVTSD2SS,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "CVTSS2SD",
argLen: 1,
asm: x86.ACVTSS2SD,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
},
},
},
[dev.ssa] cmd/compile: add complex arithmetic Still to do: details, more testing corner cases. (e.g. negative zero) Includes small cleanups for previous CL. Note: complex division is currently done in the runtime, so the division code here is apparently not yet necessary and also not tested. Seems likely better to open code division and expose the widening/narrowing to optimization. Complex64 multiplication and division is done in wide format to avoid cancellation errors; for division, this also happens to be compatible with pre-SSA practice (which uses a single complex128 division function). It would-be-nice to widen for complex128 multiplication intermediates as well, but that is trickier to implement without a handy wider-precision format. Change-Id: I595a4300f68868fb7641852a54674c6b2b78855e Reviewed-on: https://go-review.googlesource.com/14028 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-28 14:24:10 -04:00
{
cmd/compile: regalloc of two address instructions x86 has a lot of instructions that require the output to be in the same register as one of the inputs. When allocating the output register, allocate the same register as the input if it is available. Improves the performance of golang.org/x/crypto/sha3 by 10% (from 6% slower than 1.6 to 4% faster). Fixes #14745 Change-Id: I4d81785240c9368e4dc75107b45c959d200df8e6 Reviewed-on: https://go-review.googlesource.com/20488 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2016-03-10 13:05:56 -08:00
name: "PXOR",
argLen: 2,
commutative: true,
resultInArg0: true,
asm: x86.APXOR,
[dev.ssa] cmd/compile: add complex arithmetic Still to do: details, more testing corner cases. (e.g. negative zero) Includes small cleanups for previous CL. Note: complex division is currently done in the runtime, so the division code here is apparently not yet necessary and also not tested. Seems likely better to open code division and expose the widening/narrowing to optimization. Complex64 multiplication and division is done in wide format to avoid cancellation errors; for division, this also happens to be compatible with pre-SSA practice (which uses a single complex128 division function). It would-be-nice to widen for complex128 multiplication intermediates as well, but that is trickier to implement without a handy wider-precision format. Change-Id: I595a4300f68868fb7641852a54674c6b2b78855e Reviewed-on: https://go-review.googlesource.com/14028 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-28 14:24:10 -04:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
{1, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add complex arithmetic Still to do: details, more testing corner cases. (e.g. negative zero) Includes small cleanups for previous CL. Note: complex division is currently done in the runtime, so the division code here is apparently not yet necessary and also not tested. Seems likely better to open code division and expose the widening/narrowing to optimization. Complex64 multiplication and division is done in wide format to avoid cancellation errors; for division, this also happens to be compatible with pre-SSA practice (which uses a single complex128 division function). It would-be-nice to widen for complex128 multiplication intermediates as well, but that is trickier to implement without a handy wider-precision format. Change-Id: I595a4300f68868fb7641852a54674c6b2b78855e Reviewed-on: https://go-review.googlesource.com/14028 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-28 14:24:10 -04:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: add complex arithmetic Still to do: details, more testing corner cases. (e.g. negative zero) Includes small cleanups for previous CL. Note: complex division is currently done in the runtime, so the division code here is apparently not yet necessary and also not tested. Seems likely better to open code division and expose the widening/narrowing to optimization. Complex64 multiplication and division is done in wide format to avoid cancellation errors; for division, this also happens to be compatible with pre-SSA practice (which uses a single complex128 division function). It would-be-nice to widen for complex128 multiplication intermediates as well, but that is trickier to implement without a handy wider-precision format. Change-Id: I595a4300f68868fb7641852a54674c6b2b78855e Reviewed-on: https://go-review.googlesource.com/14028 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-28 14:24:10 -04:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "LEAQ",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
rematerializeable: true,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "LEAQ1",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "LEAQ2",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "LEAQ4",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "LEAQ8",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVBload",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: promote byte/word operation Writing to low 8/16 bits of register creates false dependency Generate 32-bit operations when possible. Change-Id: I8eb6c1c43a66424eec6baa91a660bceb6b80d1d3 Reviewed-on: https://go-review.googlesource.com/19506 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Ilya Tocar <ilya.tocar@intel.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-15 17:01:26 +03:00
asm: x86.AMOVBLZX,
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVBQSXload",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVBQSX,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVWload",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: promote byte/word operation Writing to low 8/16 bits of register creates false dependency Generate 32-bit operations when possible. Change-Id: I8eb6c1c43a66424eec6baa91a660bceb6b80d1d3 Reviewed-on: https://go-review.googlesource.com/19506 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Ilya Tocar <ilya.tocar@intel.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-15 17:01:26 +03:00
asm: x86.AMOVWLZX,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
},
},
},
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVWQSXload",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVWQSX,
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVLload",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVL,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
},
},
},
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVLQSXload",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVLQSX,
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVQload",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVQ,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVBstore",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVB,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
},
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVWstore",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVW,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVLstore",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVL,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
},
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVQstore",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVQ,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
},
},
},
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
{
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
name: "MOVOload",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
asm: x86.AMOVUPS,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
},
},
},
{
name: "MOVOstore",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
asm: x86.AMOVUPS,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 4294901760}, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
},
},
},
{
name: "MOVBloadidx1",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: promote byte/word operation Writing to low 8/16 bits of register creates false dependency Generate 32-bit operations when possible. Change-Id: I8eb6c1c43a66424eec6baa91a660bceb6b80d1d3 Reviewed-on: https://go-review.googlesource.com/19506 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Ilya Tocar <ilya.tocar@intel.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-15 17:01:26 +03:00
asm: x86.AMOVBLZX,
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
},
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
},
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
},
},
cmd/compile: fix plan9-amd64 build The previous rules to combine indexed loads produced addresses like: From: obj.Addr{ Type: TYPE_MEM, Reg: REG_CX, Name: NAME_AUTO, Offset: 121, ... } which are erroneous because NAME_AUTO implies a base register of REG_SP, and cmd/internal/obj/x86 makes many assumptions to this effect. Note that previously we were also producing an extra "ADDQ SP, CX" instruction, so indexing off of SP was already handled. The approach taken by this CL to address the problem is to instead produce addresses like: From: obj.Addr{ Type: TYPE_MEM, Reg: REG_SP, Name: NAME_AUTO, Offset: 121, Index: REG_CX, Scale: 1, } and to omit the "ADDQ SP, CX" instruction. Downside to this approach is it requires adding a lot of new MOV[WLQ]loadidx1 instructions that nearly duplicate functionality of the existing MOV[WLQ]loadidx[248] instructions, but with a different Scale. Fixes #15001. Change-Id: Iad9a1a41e5e2552f8d22e3ba975e4ea0862dffd2 Reviewed-on: https://go-review.googlesource.com/21245 Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
2016-03-28 19:10:13 -07:00
{
name: "MOVWloadidx1",
auxType: auxSymOff,
argLen: 3,
asm: x86.AMOVWLZX,
reg: regInfo{
inputs: []inputInfo{
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
},
outputs: []regMask{
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
{
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
name: "MOVWloadidx2",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: promote byte/word operation Writing to low 8/16 bits of register creates false dependency Generate 32-bit operations when possible. Change-Id: I8eb6c1c43a66424eec6baa91a660bceb6b80d1d3 Reviewed-on: https://go-review.googlesource.com/19506 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Ilya Tocar <ilya.tocar@intel.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-15 17:01:26 +03:00
asm: x86.AMOVWLZX,
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
},
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
},
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
},
},
cmd/compile: fix plan9-amd64 build The previous rules to combine indexed loads produced addresses like: From: obj.Addr{ Type: TYPE_MEM, Reg: REG_CX, Name: NAME_AUTO, Offset: 121, ... } which are erroneous because NAME_AUTO implies a base register of REG_SP, and cmd/internal/obj/x86 makes many assumptions to this effect. Note that previously we were also producing an extra "ADDQ SP, CX" instruction, so indexing off of SP was already handled. The approach taken by this CL to address the problem is to instead produce addresses like: From: obj.Addr{ Type: TYPE_MEM, Reg: REG_SP, Name: NAME_AUTO, Offset: 121, Index: REG_CX, Scale: 1, } and to omit the "ADDQ SP, CX" instruction. Downside to this approach is it requires adding a lot of new MOV[WLQ]loadidx1 instructions that nearly duplicate functionality of the existing MOV[WLQ]loadidx[248] instructions, but with a different Scale. Fixes #15001. Change-Id: Iad9a1a41e5e2552f8d22e3ba975e4ea0862dffd2 Reviewed-on: https://go-review.googlesource.com/21245 Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
2016-03-28 19:10:13 -07:00
{
name: "MOVLloadidx1",
auxType: auxSymOff,
argLen: 3,
asm: x86.AMOVL,
reg: regInfo{
inputs: []inputInfo{
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
},
outputs: []regMask{
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
{
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
name: "MOVLloadidx4",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVL,
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: fix plan9-amd64 build The previous rules to combine indexed loads produced addresses like: From: obj.Addr{ Type: TYPE_MEM, Reg: REG_CX, Name: NAME_AUTO, Offset: 121, ... } which are erroneous because NAME_AUTO implies a base register of REG_SP, and cmd/internal/obj/x86 makes many assumptions to this effect. Note that previously we were also producing an extra "ADDQ SP, CX" instruction, so indexing off of SP was already handled. The approach taken by this CL to address the problem is to instead produce addresses like: From: obj.Addr{ Type: TYPE_MEM, Reg: REG_SP, Name: NAME_AUTO, Offset: 121, Index: REG_CX, Scale: 1, } and to omit the "ADDQ SP, CX" instruction. Downside to this approach is it requires adding a lot of new MOV[WLQ]loadidx1 instructions that nearly duplicate functionality of the existing MOV[WLQ]loadidx[248] instructions, but with a different Scale. Fixes #15001. Change-Id: Iad9a1a41e5e2552f8d22e3ba975e4ea0862dffd2 Reviewed-on: https://go-review.googlesource.com/21245 Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
2016-03-28 19:10:13 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
},
outputs: []regMask{
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
},
},
},
{
name: "MOVQloadidx1",
auxType: auxSymOff,
argLen: 3,
asm: x86.AMOVQ,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
},
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
},
[dev.ssa] cmd/compile: lots of small rewrite optimizations Small optimizations I noticed while looking at Giovanni's test cases. More shifts by constants. Indexed stores for smaller types. Fold LEA into loads/stores. More extending loads. CMP $0 of AND -> TEST Fix order of TEST ops. Giovanni's test cases at https://gist.github.com/rasky/62fba94e3a20d1b05b2a Change-Id: I7077bc0b5319bf05767eeb39f401f4bb4b39f635 Reviewed-on: https://go-review.googlesource.com/19086 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-30 11:25:38 -08:00
},
},
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
{
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
name: "MOVQloadidx8",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVQ,
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
},
},
},
{
name: "MOVBstoreidx1",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 4,
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
asm: x86.AMOVB,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{2, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
},
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
{
name: "MOVWstoreidx1",
auxType: auxSymOff,
argLen: 4,
asm: x86.AMOVW,
reg: regInfo{
inputs: []inputInfo{
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{2, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
},
},
},
[dev.ssa] cmd/compile: better copying Replace REP MOVSB with all the copying techniques used by the old compiler. Copy in chunks, DUFFCOPY, etc. Introduces MOVO opcodes and an Int128 type to move around 16 bytes at a time. Change-Id: I1e73e68ca1d8b3dd58bb4af2f4c9e5d9bf13a502 Reviewed-on: https://go-review.googlesource.com/16174 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2015-10-21 17:18:07 -07:00
{
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
name: "MOVWstoreidx2",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 4,
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
asm: x86.AMOVW,
[dev.ssa] cmd/compile: better copying Replace REP MOVSB with all the copying techniques used by the old compiler. Copy in chunks, DUFFCOPY, etc. Introduces MOVO opcodes and an Int128 type to move around 16 bytes at a time. Change-Id: I1e73e68ca1d8b3dd58bb4af2f4c9e5d9bf13a502 Reviewed-on: https://go-review.googlesource.com/16174 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2015-10-21 17:18:07 -07:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{2, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: better copying Replace REP MOVSB with all the copying techniques used by the old compiler. Copy in chunks, DUFFCOPY, etc. Introduces MOVO opcodes and an Int128 type to move around 16 bytes at a time. Change-Id: I1e73e68ca1d8b3dd58bb4af2f4c9e5d9bf13a502 Reviewed-on: https://go-review.googlesource.com/16174 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2015-10-21 17:18:07 -07:00
},
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
},
},
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
{
name: "MOVLstoreidx1",
auxType: auxSymOff,
argLen: 4,
asm: x86.AMOVL,
reg: regInfo{
inputs: []inputInfo{
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{2, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
},
},
},
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
{
name: "MOVLstoreidx4",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 4,
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
asm: x86.AMOVL,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{2, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: better copying Replace REP MOVSB with all the copying techniques used by the old compiler. Copy in chunks, DUFFCOPY, etc. Introduces MOVO opcodes and an Int128 type to move around 16 bytes at a time. Change-Id: I1e73e68ca1d8b3dd58bb4af2f4c9e5d9bf13a502 Reviewed-on: https://go-review.googlesource.com/16174 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2015-10-21 17:18:07 -07:00
},
},
},
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
{
name: "MOVQstoreidx1",
auxType: auxSymOff,
argLen: 4,
asm: x86.AMOVQ,
reg: regInfo{
inputs: []inputInfo{
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{2, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
},
},
},
[dev.ssa] cmd/compile: better copying Replace REP MOVSB with all the copying techniques used by the old compiler. Copy in chunks, DUFFCOPY, etc. Introduces MOVO opcodes and an Int128 type to move around 16 bytes at a time. Change-Id: I1e73e68ca1d8b3dd58bb4af2f4c9e5d9bf13a502 Reviewed-on: https://go-review.googlesource.com/16174 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2015-10-21 17:18:07 -07:00
{
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
name: "MOVQstoreidx8",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 4,
[dev.ssa] cmd/compile: more combining of ops into instructions Mostly indexed loads. A few more LEA cases. Change-Id: Idc1d447ed0dd6e906cd48e70307a95e77f61cf5f Reviewed-on: https://go-review.googlesource.com/19172 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2016-02-02 11:13:50 -08:00
asm: x86.AMOVQ,
[dev.ssa] cmd/compile: better copying Replace REP MOVSB with all the copying techniques used by the old compiler. Copy in chunks, DUFFCOPY, etc. Introduces MOVO opcodes and an Int128 type to move around 16 bytes at a time. Change-Id: I1e73e68ca1d8b3dd58bb4af2f4c9e5d9bf13a502 Reviewed-on: https://go-review.googlesource.com/16174 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2015-10-21 17:18:07 -07:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{2, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: better copying Replace REP MOVSB with all the copying techniques used by the old compiler. Copy in chunks, DUFFCOPY, etc. Introduces MOVO opcodes and an Int128 type to move around 16 bytes at a time. Change-Id: I1e73e68ca1d8b3dd58bb4af2f4c9e5d9bf13a502 Reviewed-on: https://go-review.googlesource.com/16174 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2015-10-21 17:18:07 -07:00
},
},
},
[dev.ssa] cmd/compile: introduce storeconst ops Introduce opcodes that store a constant value. AuxInt now needs to hold both the value to be stored and the constant offset at which to store it. Introduce a StoreConst type to help encode/decode these parts to/from an AuxInt. Change-Id: I1631883abe035cff4b16368683e1eb3d2ccb674d Reviewed-on: https://go-review.googlesource.com/16170 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-10-21 13:13:56 -07:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVBstoreconst",
auxType: auxSymValAndOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVB,
[dev.ssa] cmd/compile: introduce storeconst ops Introduce opcodes that store a constant value. AuxInt now needs to hold both the value to be stored and the constant offset at which to store it. Introduce a StoreConst type to help encode/decode these parts to/from an AuxInt. Change-Id: I1631883abe035cff4b16368683e1eb3d2ccb674d Reviewed-on: https://go-review.googlesource.com/16170 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-10-21 13:13:56 -07:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: introduce storeconst ops Introduce opcodes that store a constant value. AuxInt now needs to hold both the value to be stored and the constant offset at which to store it. Introduce a StoreConst type to help encode/decode these parts to/from an AuxInt. Change-Id: I1631883abe035cff4b16368683e1eb3d2ccb674d Reviewed-on: https://go-review.googlesource.com/16170 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-10-21 13:13:56 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVWstoreconst",
auxType: auxSymValAndOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVW,
[dev.ssa] cmd/compile: introduce storeconst ops Introduce opcodes that store a constant value. AuxInt now needs to hold both the value to be stored and the constant offset at which to store it. Introduce a StoreConst type to help encode/decode these parts to/from an AuxInt. Change-Id: I1631883abe035cff4b16368683e1eb3d2ccb674d Reviewed-on: https://go-review.googlesource.com/16170 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-10-21 13:13:56 -07:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: introduce storeconst ops Introduce opcodes that store a constant value. AuxInt now needs to hold both the value to be stored and the constant offset at which to store it. Introduce a StoreConst type to help encode/decode these parts to/from an AuxInt. Change-Id: I1631883abe035cff4b16368683e1eb3d2ccb674d Reviewed-on: https://go-review.googlesource.com/16170 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-10-21 13:13:56 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVLstoreconst",
auxType: auxSymValAndOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVL,
[dev.ssa] cmd/compile: introduce storeconst ops Introduce opcodes that store a constant value. AuxInt now needs to hold both the value to be stored and the constant offset at which to store it. Introduce a StoreConst type to help encode/decode these parts to/from an AuxInt. Change-Id: I1631883abe035cff4b16368683e1eb3d2ccb674d Reviewed-on: https://go-review.googlesource.com/16170 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-10-21 13:13:56 -07:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: introduce storeconst ops Introduce opcodes that store a constant value. AuxInt now needs to hold both the value to be stored and the constant offset at which to store it. Introduce a StoreConst type to help encode/decode these parts to/from an AuxInt. Change-Id: I1631883abe035cff4b16368683e1eb3d2ccb674d Reviewed-on: https://go-review.googlesource.com/16170 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-10-21 13:13:56 -07:00
},
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVQstoreconst",
auxType: auxSymValAndOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
asm: x86.AMOVQ,
[dev.ssa] cmd/compile: introduce storeconst ops Introduce opcodes that store a constant value. AuxInt now needs to hold both the value to be stored and the constant offset at which to store it. Introduce a StoreConst type to help encode/decode these parts to/from an AuxInt. Change-Id: I1631883abe035cff4b16368683e1eb3d2ccb674d Reviewed-on: https://go-review.googlesource.com/16170 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-10-21 13:13:56 -07:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: introduce storeconst ops Introduce opcodes that store a constant value. AuxInt now needs to hold both the value to be stored and the constant offset at which to store it. Introduce a StoreConst type to help encode/decode these parts to/from an AuxInt. Change-Id: I1631883abe035cff4b16368683e1eb3d2ccb674d Reviewed-on: https://go-review.googlesource.com/16170 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-10-21 13:13:56 -07:00
},
},
},
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
{
name: "MOVBstoreconstidx1",
auxType: auxSymValAndOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
asm: x86.AMOVB,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
},
},
},
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
{
name: "MOVWstoreconstidx1",
auxType: auxSymValAndOff,
argLen: 3,
asm: x86.AMOVW,
reg: regInfo{
inputs: []inputInfo{
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
},
},
},
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
{
name: "MOVWstoreconstidx2",
auxType: auxSymValAndOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
asm: x86.AMOVW,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
},
},
},
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
{
name: "MOVLstoreconstidx1",
auxType: auxSymValAndOff,
argLen: 3,
asm: x86.AMOVL,
reg: regInfo{
inputs: []inputInfo{
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
},
},
},
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
{
name: "MOVLstoreconstidx4",
auxType: auxSymValAndOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
asm: x86.AMOVL,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
},
},
},
cmd/compile: Add more idx1 load/store instructions Helpful for indexed loads and stores when the stride is not equal to the size being loaded/stored. Update #7927 Change-Id: I8714dd4c7b18a96a611bf5647ee21f753d723945 Reviewed-on: https://go-review.googlesource.com/21346 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-31 09:34:35 -07:00
{
name: "MOVQstoreconstidx1",
auxType: auxSymValAndOff,
argLen: 3,
asm: x86.AMOVQ,
reg: regInfo{
inputs: []inputInfo{
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
},
},
},
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
{
name: "MOVQstoreconstidx8",
auxType: auxSymValAndOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
asm: x86.AMOVQ,
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
{0, 4295032831}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 SB
[dev.ssa] cmd/compile: add store constant indexed operations Change-Id: Ifb8eba1929c79ee7a8cae2191613c55a3b8f74e5 Reviewed-on: https://go-review.googlesource.com/19236 Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-04 15:53:33 -08:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "DUFFZERO",
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 128}, // DI
{1, 65536}, // X0
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934720, // DI FLAGS
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
},
},
[dev.ssa] Merge remote-tracking branch 'origin/master' into mergebranch The only major fixup is that duffzero changed from 8-byte writes to 16-byte writes. Change-Id: I1762b74ce67a8e4b81c11568027cdb3572f7f87c
2015-10-19 13:56:55 -07:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "MOVOconst",
cmd/compile: more sanity checks on rewrite rules Make sure ops have the right number of args, set aux and auxint only if allowed, etc. Normalize error reporting format. Change-Id: Ie545fcc5990c8c7d62d40d9a0a55885f941eb645 Reviewed-on: https://go-review.googlesource.com/22320 Reviewed-by: David Chase <drchase@google.com>
2016-04-20 11:17:41 -07:00
auxType: auxInt128,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
rematerializeable: true,
[dev.ssa] Merge remote-tracking branch 'origin/master' into mergebranch The only major fixup is that duffzero changed from 8-byte writes to 16-byte writes. Change-Id: I1762b74ce67a8e4b81c11568027cdb3572f7f87c
2015-10-19 13:56:55 -07:00
reg: regInfo{
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4294901760, // X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
[dev.ssa] Merge remote-tracking branch 'origin/master' into mergebranch The only major fixup is that duffzero changed from 8-byte writes to 16-byte writes. Change-Id: I1762b74ce67a8e4b81c11568027cdb3572f7f87c
2015-10-19 13:56:55 -07:00
},
},
},
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "REPSTOSQ",
argLen: 4,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 128}, // DI
{1, 2}, // CX
{2, 1}, // AX
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 130, // CX DI
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "CALLstatic",
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
reg: regInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 12884901871, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 FLAGS
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
},
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "CALLclosure",
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{1, 4}, // DX
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 12884901871, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 FLAGS
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
},
[dev.ssa] cmd/compile/internal/gc: implement go and defer TODO: for now, just function calls. Do method and interface calls. Change-Id: Ib262dfa31cb753996cde899beaad4dc2e66705ac Reviewed-on: https://go-review.googlesource.com/14035 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-28 22:51:01 -07:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "CALLdefer",
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/gc: implement go and defer TODO: for now, just function calls. Do method and interface calls. Change-Id: Ib262dfa31cb753996cde899beaad4dc2e66705ac Reviewed-on: https://go-review.googlesource.com/14035 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-28 22:51:01 -07:00
reg: regInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 12884901871, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 FLAGS
[dev.ssa] cmd/compile/internal/gc: implement go and defer TODO: for now, just function calls. Do method and interface calls. Change-Id: Ib262dfa31cb753996cde899beaad4dc2e66705ac Reviewed-on: https://go-review.googlesource.com/14035 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-28 22:51:01 -07:00
},
},
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "CALLgo",
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/gc: implement go and defer TODO: for now, just function calls. Do method and interface calls. Change-Id: Ib262dfa31cb753996cde899beaad4dc2e66705ac Reviewed-on: https://go-review.googlesource.com/14035 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-28 22:51:01 -07:00
reg: regInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 12884901871, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 FLAGS
[dev.ssa] cmd/compile/internal/gc: implement go and defer TODO: for now, just function calls. Do method and interface calls. Change-Id: Ib262dfa31cb753996cde899beaad4dc2e66705ac Reviewed-on: https://go-review.googlesource.com/14035 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-28 22:51:01 -07:00
},
},
[dev.ssa] cmd/compile/internal/ssa: complete call ops OCALLINTER, as well as ODEFER/OPROC with OCALLMETH/OCALLINTER. Move all the call logic to its own routine, a lot of the code is shared. Change-Id: Ieac59596165e434cc6d1d7b5e46b78957e9c5ed3 Reviewed-on: https://go-review.googlesource.com/14464 Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2015-09-09 23:56:59 -07:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "CALLinter",
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: complete call ops OCALLINTER, as well as ODEFER/OPROC with OCALLMETH/OCALLINTER. Move all the call logic to its own routine, a lot of the code is shared. Change-Id: Ieac59596165e434cc6d1d7b5e46b78957e9c5ed3 Reviewed-on: https://go-review.googlesource.com/14464 Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2015-09-09 23:56:59 -07:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65519}, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile/internal/ssa: complete call ops OCALLINTER, as well as ODEFER/OPROC with OCALLMETH/OCALLINTER. Move all the call logic to its own routine, a lot of the code is shared. Change-Id: Ieac59596165e434cc6d1d7b5e46b78957e9c5ed3 Reviewed-on: https://go-review.googlesource.com/14464 Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2015-09-09 23:56:59 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 12884901871, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15 X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 FLAGS
[dev.ssa] cmd/compile/internal/ssa: complete call ops OCALLINTER, as well as ODEFER/OPROC with OCALLMETH/OCALLINTER. Move all the call logic to its own routine, a lot of the code is shared. Change-Id: Ieac59596165e434cc6d1d7b5e46b78957e9c5ed3 Reviewed-on: https://go-review.googlesource.com/14464 Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2015-09-09 23:56:59 -07:00
},
},
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
{
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
name: "DUFFCOPY",
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: better copying Replace REP MOVSB with all the copying techniques used by the old compiler. Copy in chunks, DUFFCOPY, etc. Introduces MOVO opcodes and an Int128 type to move around 16 bytes at a time. Change-Id: I1e73e68ca1d8b3dd58bb4af2f4c9e5d9bf13a502 Reviewed-on: https://go-review.googlesource.com/16174 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2015-10-21 17:18:07 -07:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 128}, // DI
{1, 64}, // SI
[dev.ssa] cmd/compile: better copying Replace REP MOVSB with all the copying techniques used by the old compiler. Copy in chunks, DUFFCOPY, etc. Introduces MOVO opcodes and an Int128 type to move around 16 bytes at a time. Change-Id: I1e73e68ca1d8b3dd58bb4af2f4c9e5d9bf13a502 Reviewed-on: https://go-review.googlesource.com/16174 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2015-10-21 17:18:07 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8590000320, // SI DI X0 FLAGS
[dev.ssa] cmd/compile: better copying Replace REP MOVSB with all the copying techniques used by the old compiler. Copy in chunks, DUFFCOPY, etc. Introduces MOVO opcodes and an Int128 type to move around 16 bytes at a time. Change-Id: I1e73e68ca1d8b3dd58bb4af2f4c9e5d9bf13a502 Reviewed-on: https://go-review.googlesource.com/16174 Reviewed-by: Todd Neal <todd@tneal.org> Run-TryBot: Keith Randall <khr@golang.org>
2015-10-21 17:18:07 -07:00
},
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "REPMOVSQ",
argLen: 4,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
reg: regInfo{
[dev.ssa] cmd/compile/internal/ssa: New register allocator Implement a global (whole function) register allocator. This replaces the local (per basic block) register allocator. Clobbering of registers by instructions is handled properly. A separate change will add the correct clobbers to all the instructions. Change-Id: I38ce4dc7dccb8303c1c0e0295fe70247b0a3f2ea Reviewed-on: https://go-review.googlesource.com/13622 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Todd Neal <todd@tneal.org>
2015-08-11 12:51:33 -07:00
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 128}, // DI
{1, 64}, // SI
{2, 2}, // CX
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 194, // CX SI DI
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
},
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "InvertFlags",
argLen: 1,
reg: regInfo{},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
},
[dev.ssa] cmd/compile: implement OGETG Change-Id: I7ecf62cf399c710b4a617803c43e83fce09b8a7d Reviewed-on: https://go-review.googlesource.com/13585 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 11:22:16 -07:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "LoweredGetG",
argLen: 1,
[dev.ssa] cmd/compile: add register spec to getg This omission was causing the new regalloc to fail. Change-Id: If7ba7be38a436dbd0dd443828ddd7ebf6e35be0e Reviewed-on: https://go-review.googlesource.com/13632 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-13 13:12:17 -07:00
reg: regInfo{
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: add register spec to getg This omission was causing the new regalloc to fail. Change-Id: If7ba7be38a436dbd0dd443828ddd7ebf6e35be0e Reviewed-on: https://go-review.googlesource.com/13632 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-13 13:12:17 -07:00
},
},
[dev.ssa] cmd/compile: implement OGETG Change-Id: I7ecf62cf399c710b4a617803c43e83fce09b8a7d Reviewed-on: https://go-review.googlesource.com/13585 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 11:22:16 -07:00
},
[dev.ssa] cmd/compile: addressed vars and closures Cleaned up first-block-in-function code. Added cases for |PHEAP for PPARAM and PAUTO. Made PPARAMOUT act more like PAUTO for purposes of address generation and vardef placement. Added cases for OCLOSUREVAR and Ops for getting closure pointer. Closure ops are scheduled at top of entry block to capture DX. Wrote test that seems to show proper behavior for addressed parameters, locals, and returns. Change-Id: Iee93ebf9e3d9f74cfb4d1c1da8038eb278d8a857 Reviewed-on: https://go-review.googlesource.com/14650 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com>
2015-09-11 16:40:05 -04:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "LoweredGetClosurePtr",
argLen: 0,
[dev.ssa] cmd/compile: addressed vars and closures Cleaned up first-block-in-function code. Added cases for |PHEAP for PPARAM and PAUTO. Made PPARAMOUT act more like PAUTO for purposes of address generation and vardef placement. Added cases for OCLOSUREVAR and Ops for getting closure pointer. Closure ops are scheduled at top of entry block to capture DX. Wrote test that seems to show proper behavior for addressed parameters, locals, and returns. Change-Id: Iee93ebf9e3d9f74cfb4d1c1da8038eb278d8a857 Reviewed-on: https://go-review.googlesource.com/14650 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com>
2015-09-11 16:40:05 -04:00
reg: regInfo{
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
4, // DX
[dev.ssa] cmd/compile: addressed vars and closures Cleaned up first-block-in-function code. Added cases for |PHEAP for PPARAM and PAUTO. Made PPARAMOUT act more like PAUTO for purposes of address generation and vardef placement. Added cases for OCLOSUREVAR and Ops for getting closure pointer. Closure ops are scheduled at top of entry block to capture DX. Wrote test that seems to show proper behavior for addressed parameters, locals, and returns. Change-Id: Iee93ebf9e3d9f74cfb4d1c1da8038eb278d8a857 Reviewed-on: https://go-review.googlesource.com/14650 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com>
2015-09-11 16:40:05 -04:00
},
},
},
[dev.ssa] cmd/compile: optimize nil checks Use faulting loads instead of test/jeq to do nil checks. Fold nil checks into a following load/store if possible. Makes binaries about 2% smaller. Change-Id: I54af0f0a93c853f37e34e0ce7e3f01dd2ac87f64 Reviewed-on: https://go-review.googlesource.com/16287 Reviewed-by: David Chase <drchase@google.com>
2015-10-23 19:12:49 -07:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "LoweredNilCheck",
argLen: 2,
[dev.ssa] cmd/compile: optimize nil checks Use faulting loads instead of test/jeq to do nil checks. Fold nil checks into a following load/store if possible. Makes binaries about 2% smaller. Change-Id: I54af0f0a93c853f37e34e0ce7e3f01dd2ac87f64 Reviewed-on: https://go-review.googlesource.com/16287 Reviewed-by: David Chase <drchase@google.com>
2015-10-23 19:12:49 -07:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: optimize nil checks Use faulting loads instead of test/jeq to do nil checks. Fold nil checks into a following load/store if possible. Makes binaries about 2% smaller. Change-Id: I54af0f0a93c853f37e34e0ce7e3f01dd2ac87f64 Reviewed-on: https://go-review.googlesource.com/16287 Reviewed-by: David Chase <drchase@google.com>
2015-10-23 19:12:49 -07:00
},
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
clobbers: 8589934592, // FLAGS
[dev.ssa] cmd/compile: optimize nil checks Use faulting loads instead of test/jeq to do nil checks. Fold nil checks into a following load/store if possible. Makes binaries about 2% smaller. Change-Id: I54af0f0a93c853f37e34e0ce7e3f01dd2ac87f64 Reviewed-on: https://go-review.googlesource.com/16287 Reviewed-by: David Chase <drchase@google.com>
2015-10-23 19:12:49 -07:00
},
},
[dev.ssa] cmd/compile: be safer about uintptr/unsafe.Pointer conversions Make sure that when a pointer value is live across a function call, we save it as a pointer. (And similarly a uintptr live across a function call should not be saved as a pointer.) Add a nasty test case. This is probably what is preventing the merge from master to dev.ssa. Signs point to something like this bug happening in mallocgc. Change-Id: Ib23fa1251b8d1c50d82c6a448cb4a4fc28219029 Reviewed-on: https://go-review.googlesource.com/16830 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: David Chase <drchase@google.com>
2015-11-10 15:35:36 -08:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "MOVQconvert",
argLen: 2,
asm: x86.AMOVQ,
[dev.ssa] cmd/compile: be safer about uintptr/unsafe.Pointer conversions Make sure that when a pointer value is live across a function call, we save it as a pointer. (And similarly a uintptr live across a function call should not be saved as a pointer.) Add a nasty test case. This is probably what is preventing the merge from master to dev.ssa. Signs point to something like this bug happening in mallocgc. Change-Id: Ib23fa1251b8d1c50d82c6a448cb4a4fc28219029 Reviewed-on: https://go-review.googlesource.com/16830 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: David Chase <drchase@google.com>
2015-11-10 15:35:36 -08:00
reg: regInfo{
inputs: []inputInfo{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
{0, 65535}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: be safer about uintptr/unsafe.Pointer conversions Make sure that when a pointer value is live across a function call, we save it as a pointer. (And similarly a uintptr live across a function call should not be saved as a pointer.) Add a nasty test case. This is probably what is preventing the merge from master to dev.ssa. Signs point to something like this bug happening in mallocgc. Change-Id: Ib23fa1251b8d1c50d82c6a448cb4a4fc28219029 Reviewed-on: https://go-review.googlesource.com/16830 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: David Chase <drchase@google.com>
2015-11-10 15:35:36 -08:00
},
outputs: []regMask{
cmd/compile: remove dots from register names They are kind of useless and are cluttering up https://go-review.googlesource.com/c/21000/ Change-Id: Iafdec75ada11c7ebdc40540d251fdc514bb00d3d Reviewed-on: https://go-review.googlesource.com/21001 Reviewed-by: Minux Ma <minux@golang.org>
2016-03-22 09:43:28 -07:00
65519, // AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15
[dev.ssa] cmd/compile: be safer about uintptr/unsafe.Pointer conversions Make sure that when a pointer value is live across a function call, we save it as a pointer. (And similarly a uintptr live across a function call should not be saved as a pointer.) Add a nasty test case. This is probably what is preventing the merge from master to dev.ssa. Signs point to something like this bug happening in mallocgc. Change-Id: Ib23fa1251b8d1c50d82c6a448cb4a4fc28219029 Reviewed-on: https://go-review.googlesource.com/16830 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: David Chase <drchase@google.com>
2015-11-10 15:35:36 -08:00
},
},
},
[dev.ssa] cmd/compile: clean up comparisons Add new constant-flags opcodes. These can be generated from comparisons that we know the result of, like x&31 < 32. Constant-fold the constant-flags opcodes into all flag users. Reorder some CMPxconst args so they read in the comparison direction. Reorg deadcode removal a bit - it needs to remove the OpCopy ops it generates when strength-reducing Phi ops. So it needs to splice out all the dead blocks and do a copy elimination before it computes live values. Change-Id: Ie922602033592ad8212efe4345394973d3b94d9f Reviewed-on: https://go-review.googlesource.com/18267 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-05 14:56:26 -08:00
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "FlagEQ",
argLen: 0,
reg: regInfo{},
[dev.ssa] cmd/compile: clean up comparisons Add new constant-flags opcodes. These can be generated from comparisons that we know the result of, like x&31 < 32. Constant-fold the constant-flags opcodes into all flag users. Reorder some CMPxconst args so they read in the comparison direction. Reorg deadcode removal a bit - it needs to remove the OpCopy ops it generates when strength-reducing Phi ops. So it needs to splice out all the dead blocks and do a copy elimination before it computes live values. Change-Id: Ie922602033592ad8212efe4345394973d3b94d9f Reviewed-on: https://go-review.googlesource.com/18267 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-05 14:56:26 -08:00
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "FlagLT_ULT",
argLen: 0,
reg: regInfo{},
[dev.ssa] cmd/compile: clean up comparisons Add new constant-flags opcodes. These can be generated from comparisons that we know the result of, like x&31 < 32. Constant-fold the constant-flags opcodes into all flag users. Reorder some CMPxconst args so they read in the comparison direction. Reorg deadcode removal a bit - it needs to remove the OpCopy ops it generates when strength-reducing Phi ops. So it needs to splice out all the dead blocks and do a copy elimination before it computes live values. Change-Id: Ie922602033592ad8212efe4345394973d3b94d9f Reviewed-on: https://go-review.googlesource.com/18267 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-05 14:56:26 -08:00
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "FlagLT_UGT",
argLen: 0,
reg: regInfo{},
[dev.ssa] cmd/compile: clean up comparisons Add new constant-flags opcodes. These can be generated from comparisons that we know the result of, like x&31 < 32. Constant-fold the constant-flags opcodes into all flag users. Reorder some CMPxconst args so they read in the comparison direction. Reorg deadcode removal a bit - it needs to remove the OpCopy ops it generates when strength-reducing Phi ops. So it needs to splice out all the dead blocks and do a copy elimination before it computes live values. Change-Id: Ie922602033592ad8212efe4345394973d3b94d9f Reviewed-on: https://go-review.googlesource.com/18267 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-05 14:56:26 -08:00
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "FlagGT_UGT",
argLen: 0,
reg: regInfo{},
[dev.ssa] cmd/compile: clean up comparisons Add new constant-flags opcodes. These can be generated from comparisons that we know the result of, like x&31 < 32. Constant-fold the constant-flags opcodes into all flag users. Reorder some CMPxconst args so they read in the comparison direction. Reorg deadcode removal a bit - it needs to remove the OpCopy ops it generates when strength-reducing Phi ops. So it needs to splice out all the dead blocks and do a copy elimination before it computes live values. Change-Id: Ie922602033592ad8212efe4345394973d3b94d9f Reviewed-on: https://go-review.googlesource.com/18267 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-05 14:56:26 -08:00
},
{
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
name: "FlagGT_ULT",
argLen: 0,
reg: regInfo{},
[dev.ssa] cmd/compile: clean up comparisons Add new constant-flags opcodes. These can be generated from comparisons that we know the result of, like x&31 < 32. Constant-fold the constant-flags opcodes into all flag users. Reorder some CMPxconst args so they read in the comparison direction. Reorg deadcode removal a bit - it needs to remove the OpCopy ops it generates when strength-reducing Phi ops. So it needs to splice out all the dead blocks and do a copy elimination before it computes live values. Change-Id: Ie922602033592ad8212efe4345394973d3b94d9f Reviewed-on: https://go-review.googlesource.com/18267 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-05 14:56:26 -08:00
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
{
name: "ADD",
argLen: 2,
commutative: true,
asm: arm.AADD,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
},
outputs: []regMask{
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
},
},
},
{
name: "ADDconst",
[dev.ssa] cmd/compile: use MOVWaddr for address on ARM Introduce an op MOVWaddr for addresses on ARM, instead of overuse ADDconst. Mark MOVWaddr as rematerializable. This fixes a liveness problem: if it were not rematerializable, the address of a variable may be spilled and later use of the address may just load the spilled value without mentioning the variable, and the liveness code may think it is dead prematurely. Update #15365. Change-Id: Ib0b0fa826bdb75c9e6bb362b95c6cf132cc6b1c0 Reviewed-on: https://go-review.googlesource.com/23942 Reviewed-by: David Chase <drchase@google.com>
2016-06-06 22:36:45 -04:00
auxType: auxInt32,
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
argLen: 1,
asm: arm.AADD,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 14335}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 SP
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "SUB",
argLen: 2,
asm: arm.ASUB,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
},
outputs: []regMask{
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "SUBconst",
auxType: auxInt32,
argLen: 1,
asm: arm.ASUB,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "RSB",
argLen: 2,
asm: arm.ARSB,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "RSBconst",
auxType: auxInt32,
argLen: 1,
asm: arm.ARSB,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
{
name: "MUL",
argLen: 2,
commutative: true,
asm: arm.AMUL,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "HMUL",
argLen: 2,
commutative: true,
asm: arm.AMULL,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "HMULU",
argLen: 2,
commutative: true,
asm: arm.AMULLU,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
[dev.ssa] cmd/compile: handle Div, Convert, GetClosurePtr etc. on ARM This CL adds support of Div, Mod, Convert, GetClosurePtr and 64-bit indexing support to SSA backend for ARM. Add tests for 64-bit indexing to cmd/compile/internal/gc/testdata/string_ssa.go. Tests cmd/compile/internal/gc/testdata/*_ssa.go passed, except compound_ssa.go and fp_ssa.go. Progress on SSA for ARM. Still not complete. Essentially the only unsupported part is floating point. Updates #15365. Change-Id: I269e88b67f641c25e7a813d910c96d356d236bff Reviewed-on: https://go-review.googlesource.com/23542 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 09:49:28 -04:00
{
name: "DIV",
argLen: 2,
asm: arm.ADIV,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: handle Div, Convert, GetClosurePtr etc. on ARM This CL adds support of Div, Mod, Convert, GetClosurePtr and 64-bit indexing support to SSA backend for ARM. Add tests for 64-bit indexing to cmd/compile/internal/gc/testdata/string_ssa.go. Tests cmd/compile/internal/gc/testdata/*_ssa.go passed, except compound_ssa.go and fp_ssa.go. Progress on SSA for ARM. Still not complete. Essentially the only unsupported part is floating point. Updates #15365. Change-Id: I269e88b67f641c25e7a813d910c96d356d236bff Reviewed-on: https://go-review.googlesource.com/23542 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 09:49:28 -04:00
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
clobbers: 4294967296, // FLAGS
[dev.ssa] cmd/compile: handle Div, Convert, GetClosurePtr etc. on ARM This CL adds support of Div, Mod, Convert, GetClosurePtr and 64-bit indexing support to SSA backend for ARM. Add tests for 64-bit indexing to cmd/compile/internal/gc/testdata/string_ssa.go. Tests cmd/compile/internal/gc/testdata/*_ssa.go passed, except compound_ssa.go and fp_ssa.go. Progress on SSA for ARM. Still not complete. Essentially the only unsupported part is floating point. Updates #15365. Change-Id: I269e88b67f641c25e7a813d910c96d356d236bff Reviewed-on: https://go-review.googlesource.com/23542 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 09:49:28 -04:00
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "DIVU",
argLen: 2,
asm: arm.ADIVU,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: handle Div, Convert, GetClosurePtr etc. on ARM This CL adds support of Div, Mod, Convert, GetClosurePtr and 64-bit indexing support to SSA backend for ARM. Add tests for 64-bit indexing to cmd/compile/internal/gc/testdata/string_ssa.go. Tests cmd/compile/internal/gc/testdata/*_ssa.go passed, except compound_ssa.go and fp_ssa.go. Progress on SSA for ARM. Still not complete. Essentially the only unsupported part is floating point. Updates #15365. Change-Id: I269e88b67f641c25e7a813d910c96d356d236bff Reviewed-on: https://go-review.googlesource.com/23542 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 09:49:28 -04:00
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
clobbers: 4294967296, // FLAGS
[dev.ssa] cmd/compile: handle Div, Convert, GetClosurePtr etc. on ARM This CL adds support of Div, Mod, Convert, GetClosurePtr and 64-bit indexing support to SSA backend for ARM. Add tests for 64-bit indexing to cmd/compile/internal/gc/testdata/string_ssa.go. Tests cmd/compile/internal/gc/testdata/*_ssa.go passed, except compound_ssa.go and fp_ssa.go. Progress on SSA for ARM. Still not complete. Essentially the only unsupported part is floating point. Updates #15365. Change-Id: I269e88b67f641c25e7a813d910c96d356d236bff Reviewed-on: https://go-review.googlesource.com/23542 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 09:49:28 -04:00
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "MOD",
argLen: 2,
asm: arm.AMOD,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: handle Div, Convert, GetClosurePtr etc. on ARM This CL adds support of Div, Mod, Convert, GetClosurePtr and 64-bit indexing support to SSA backend for ARM. Add tests for 64-bit indexing to cmd/compile/internal/gc/testdata/string_ssa.go. Tests cmd/compile/internal/gc/testdata/*_ssa.go passed, except compound_ssa.go and fp_ssa.go. Progress on SSA for ARM. Still not complete. Essentially the only unsupported part is floating point. Updates #15365. Change-Id: I269e88b67f641c25e7a813d910c96d356d236bff Reviewed-on: https://go-review.googlesource.com/23542 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 09:49:28 -04:00
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
clobbers: 4294967296, // FLAGS
[dev.ssa] cmd/compile: handle Div, Convert, GetClosurePtr etc. on ARM This CL adds support of Div, Mod, Convert, GetClosurePtr and 64-bit indexing support to SSA backend for ARM. Add tests for 64-bit indexing to cmd/compile/internal/gc/testdata/string_ssa.go. Tests cmd/compile/internal/gc/testdata/*_ssa.go passed, except compound_ssa.go and fp_ssa.go. Progress on SSA for ARM. Still not complete. Essentially the only unsupported part is floating point. Updates #15365. Change-Id: I269e88b67f641c25e7a813d910c96d356d236bff Reviewed-on: https://go-review.googlesource.com/23542 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 09:49:28 -04:00
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "MODU",
argLen: 2,
asm: arm.AMODU,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: handle Div, Convert, GetClosurePtr etc. on ARM This CL adds support of Div, Mod, Convert, GetClosurePtr and 64-bit indexing support to SSA backend for ARM. Add tests for 64-bit indexing to cmd/compile/internal/gc/testdata/string_ssa.go. Tests cmd/compile/internal/gc/testdata/*_ssa.go passed, except compound_ssa.go and fp_ssa.go. Progress on SSA for ARM. Still not complete. Essentially the only unsupported part is floating point. Updates #15365. Change-Id: I269e88b67f641c25e7a813d910c96d356d236bff Reviewed-on: https://go-review.googlesource.com/23542 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 09:49:28 -04:00
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
clobbers: 4294967296, // FLAGS
[dev.ssa] cmd/compile: handle Div, Convert, GetClosurePtr etc. on ARM This CL adds support of Div, Mod, Convert, GetClosurePtr and 64-bit indexing support to SSA backend for ARM. Add tests for 64-bit indexing to cmd/compile/internal/gc/testdata/string_ssa.go. Tests cmd/compile/internal/gc/testdata/*_ssa.go passed, except compound_ssa.go and fp_ssa.go. Progress on SSA for ARM. Still not complete. Essentially the only unsupported part is floating point. Updates #15365. Change-Id: I269e88b67f641c25e7a813d910c96d356d236bff Reviewed-on: https://go-review.googlesource.com/23542 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 09:49:28 -04:00
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
{
name: "ADDS",
argLen: 2,
commutative: true,
asm: arm.AADD,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
clobbers: 4294967296, // FLAGS
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "ADC",
argLen: 3,
commutative: true,
asm: arm.AADC,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{2, 4294967296}, // FLAGS
{0, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
{1, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "SUBS",
argLen: 2,
asm: arm.ASUB,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
clobbers: 4294967296, // FLAGS
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "SBC",
argLen: 3,
asm: arm.ASBC,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{2, 4294967296}, // FLAGS
{0, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
{1, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "MULLU",
argLen: 2,
commutative: true,
asm: arm.AMULLU,
reg: regInfo{
inputs: []inputInfo{
{0, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
{1, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
clobbers: 1, // R0
outputs: []regMask{
5118, // R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "MULA",
argLen: 3,
asm: arm.AMULA,
reg: regInfo{
inputs: []inputInfo{
{0, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
{1, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
{2, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{
name: "ADDF",
argLen: 2,
commutative: true,
asm: arm.AADDF,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
{1, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "ADDD",
argLen: 2,
commutative: true,
asm: arm.AADDD,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
{1, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "SUBF",
argLen: 2,
asm: arm.ASUBF,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
{1, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "SUBD",
argLen: 2,
asm: arm.ASUBD,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
{1, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "MULF",
argLen: 2,
commutative: true,
asm: arm.AMULF,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
{1, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "MULD",
argLen: 2,
commutative: true,
asm: arm.AMULD,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
{1, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "DIVF",
argLen: 2,
asm: arm.ADIVF,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
{1, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "DIVD",
argLen: 2,
asm: arm.ADIVD,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
{1, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
{
name: "AND",
argLen: 2,
commutative: true,
asm: arm.AAND,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "ANDconst",
auxType: auxInt32,
argLen: 1,
asm: arm.AAND,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "OR",
argLen: 2,
commutative: true,
asm: arm.AORR,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "ORconst",
auxType: auxInt32,
argLen: 1,
asm: arm.AORR,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "XOR",
argLen: 2,
commutative: true,
asm: arm.AEOR,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "XORconst",
auxType: auxInt32,
argLen: 1,
asm: arm.AEOR,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "BIC",
argLen: 2,
asm: arm.ABIC,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "BICconst",
auxType: auxInt32,
argLen: 1,
asm: arm.ABIC,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
[dev.ssa] cmd/compile: handle boolean values for SSA on ARM Fix hardcoded flag register mask in ssa/flagalloc.go by auto-generating the mask. Also fix a mistake (in previous CL) about conditional branches. Progress on SSA backend for ARM. Still not complete. Now "container/ring" package compiles and tests passed. Updates #15365. Change-Id: Id7c8805c30dbb8107baedb485ed0f71f59ed6ea8 Reviewed-on: https://go-review.googlesource.com/23093 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 11:25:07 -04:00
{
name: "MVN",
argLen: 1,
asm: arm.AMVN,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: handle boolean values for SSA on ARM Fix hardcoded flag register mask in ssa/flagalloc.go by auto-generating the mask. Also fix a mistake (in previous CL) about conditional branches. Progress on SSA backend for ARM. Still not complete. Now "container/ring" package compiles and tests passed. Updates #15365. Change-Id: Id7c8805c30dbb8107baedb485ed0f71f59ed6ea8 Reviewed-on: https://go-review.googlesource.com/23093 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 11:25:07 -04:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{
name: "SQRTD",
argLen: 1,
asm: arm.ASQRTD,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
{
name: "SLL",
argLen: 2,
asm: arm.ASLL,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
clobbers: 4294967296, // FLAGS
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "SLLconst",
auxType: auxInt32,
argLen: 1,
asm: arm.ASLL,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "SRL",
argLen: 2,
asm: arm.ASRL,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
clobbers: 4294967296, // FLAGS
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "SRLconst",
auxType: auxInt32,
argLen: 1,
asm: arm.ASRL,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "SRA",
argLen: 2,
asm: arm.ASRA,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
clobbers: 4294967296, // FLAGS
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "SRAconst",
auxType: auxInt32,
argLen: 1,
asm: arm.ASRA,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement shifts & multiplications for SSA on ARM Implement shifts and multiplications for up to 32-bit values. Also handle Exit block. Progress on SSA backend for ARM. Still not complete. container/heap, crypto/subtle, hash/adler32 packages compile and tests passed. Updates #15365. Change-Id: I6bee4d5b0051e51d5de97e8a1938c4b87a36cbf8 Reviewed-on: https://go-review.googlesource.com/23096 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:22:56 -04:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
[dev.ssa] cmd/compile: handle 64-bit shifts on ARM Also fix a mistake in previous CL about x8 and x16 shifts: the shift needs ZeroExt. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: Ibc352760023d38bc6b9c5251e929fe26e016637a Reviewed-on: https://go-review.googlesource.com/23486 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 23:17:42 -04:00
{
name: "SRRconst",
auxType: auxInt32,
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: handle 64-bit shifts on ARM Also fix a mistake in previous CL about x8 and x16 shifts: the shift needs ZeroExt. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: Ibc352760023d38bc6b9c5251e929fe26e016637a Reviewed-on: https://go-review.googlesource.com/23486 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 23:17:42 -04:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
{
name: "CMP",
argLen: 2,
asm: arm.ACMP,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
4294967296, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
},
{
name: "CMPconst",
auxType: auxInt32,
argLen: 1,
asm: arm.ACMP,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
4294967296, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
},
{
name: "CMN",
argLen: 2,
asm: arm.ACMN,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
4294967296, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
},
{
name: "CMNconst",
auxType: auxInt32,
argLen: 1,
asm: arm.ACMN,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
4294967296, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
},
{
name: "TST",
argLen: 2,
commutative: true,
asm: arm.ATST,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
4294967296, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
},
{
name: "TSTconst",
auxType: auxInt32,
argLen: 1,
asm: arm.ATST,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
4294967296, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
},
{
name: "TEQ",
argLen: 2,
commutative: true,
asm: arm.ATEQ,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
4294967296, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
},
{
name: "TEQconst",
auxType: auxInt32,
argLen: 1,
asm: arm.ATEQ,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
4294967296, // FLAGS
},
},
},
{
name: "CMPF",
argLen: 2,
asm: arm.ACMPF,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
{1, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
4294967296, // FLAGS
},
},
},
{
name: "CMPD",
argLen: 2,
asm: arm.ACMPD,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
{1, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
4294967296, // FLAGS
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
},
},
},
{
name: "MOVWconst",
auxType: auxInt32,
argLen: 0,
rematerializeable: true,
asm: arm.AMOVW,
reg: regInfo{
outputs: []regMask{
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
},
},
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{
name: "MOVFconst",
auxType: auxFloat64,
argLen: 0,
rematerializeable: true,
asm: arm.AMOVF,
reg: regInfo{
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "MOVDconst",
auxType: auxFloat64,
argLen: 0,
rematerializeable: true,
asm: arm.AMOVD,
reg: regInfo{
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
[dev.ssa] cmd/compile: use MOVWaddr for address on ARM Introduce an op MOVWaddr for addresses on ARM, instead of overuse ADDconst. Mark MOVWaddr as rematerializable. This fixes a liveness problem: if it were not rematerializable, the address of a variable may be spilled and later use of the address may just load the spilled value without mentioning the variable, and the liveness code may think it is dead prematurely. Update #15365. Change-Id: Ib0b0fa826bdb75c9e6bb362b95c6cf132cc6b1c0 Reviewed-on: https://go-review.googlesource.com/23942 Reviewed-by: David Chase <drchase@google.com>
2016-06-06 22:36:45 -04:00
{
name: "MOVWaddr",
auxType: auxSymOff,
argLen: 1,
rematerializeable: true,
asm: arm.AMOVW,
reg: regInfo{
inputs: []inputInfo{
{0, 8589942784}, // SP SB
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
{
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
name: "MOVBload",
auxType: auxSymOff,
argLen: 2,
asm: arm.AMOVB,
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 8589948927}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 SP SB
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
},
outputs: []regMask{
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "MOVBUload",
auxType: auxSymOff,
argLen: 2,
asm: arm.AMOVBU,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 8589948927}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 SP SB
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "MOVHload",
auxType: auxSymOff,
argLen: 2,
asm: arm.AMOVH,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 8589948927}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 SP SB
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "MOVHUload",
auxType: auxSymOff,
argLen: 2,
asm: arm.AMOVHU,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 8589948927}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 SP SB
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
},
},
},
{
cmd/compile: more sanity checks on rewrite rules Make sure ops have the right number of args, set aux and auxint only if allowed, etc. Normalize error reporting format. Change-Id: Ie545fcc5990c8c7d62d40d9a0a55885f941eb645 Reviewed-on: https://go-review.googlesource.com/22320 Reviewed-by: David Chase <drchase@google.com>
2016-04-20 11:17:41 -07:00
name: "MOVWload",
auxType: auxSymOff,
argLen: 2,
asm: arm.AMOVW,
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 8589948927}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 SP SB
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
},
outputs: []regMask{
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{
name: "MOVFload",
auxType: auxSymOff,
argLen: 2,
asm: arm.AMOVF,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 8589948927}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 SP SB
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "MOVDload",
auxType: auxSymOff,
argLen: 2,
asm: arm.AMOVD,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 8589948927}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 SP SB
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
{
name: "MOVBstore",
auxType: auxSymOff,
argLen: 3,
asm: arm.AMOVB,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{0, 8589948927}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 SP SB
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
},
{
name: "MOVHstore",
auxType: auxSymOff,
argLen: 3,
asm: arm.AMOVH,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{0, 8589948927}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 SP SB
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
},
},
},
{
cmd/compile: more sanity checks on rewrite rules Make sure ops have the right number of args, set aux and auxint only if allowed, etc. Normalize error reporting format. Change-Id: Ie545fcc5990c8c7d62d40d9a0a55885f941eb645 Reviewed-on: https://go-review.googlesource.com/22320 Reviewed-by: David Chase <drchase@google.com>
2016-04-20 11:17:41 -07:00
name: "MOVWstore",
auxType: auxSymOff,
argLen: 3,
asm: arm.AMOVW,
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{1, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
{0, 8589948927}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 SP SB
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
},
},
},
{
name: "MOVFstore",
auxType: auxSymOff,
argLen: 3,
asm: arm.AMOVF,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 8589948927}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 SP SB
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{1, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "MOVDstore",
auxType: auxSymOff,
argLen: 3,
asm: arm.AMOVD,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 8589948927}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 SP SB
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{1, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
},
{
name: "MOVBreg",
argLen: 1,
asm: arm.AMOVBS,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "MOVBUreg",
argLen: 1,
asm: arm.AMOVBU,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "MOVHreg",
argLen: 1,
asm: arm.AMOVHS,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "MOVHUreg",
argLen: 1,
asm: arm.AMOVHU,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
},
},
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{
name: "MOVWF",
argLen: 1,
asm: arm.AMOVWF,
reg: regInfo{
inputs: []inputInfo{
{0, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "MOVWD",
argLen: 1,
asm: arm.AMOVWD,
reg: regInfo{
inputs: []inputInfo{
{0, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "MOVWUF",
argLen: 1,
asm: arm.AMOVWF,
reg: regInfo{
inputs: []inputInfo{
{0, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "MOVWUD",
argLen: 1,
asm: arm.AMOVWD,
reg: regInfo{
inputs: []inputInfo{
{0, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "MOVFW",
argLen: 1,
asm: arm.AMOVFW,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "MOVDW",
argLen: 1,
asm: arm.AMOVDW,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "MOVFWU",
argLen: 1,
asm: arm.AMOVFW,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "MOVDWU",
argLen: 1,
asm: arm.AMOVDW,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "MOVFD",
argLen: 1,
asm: arm.AMOVFD,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
{
name: "MOVDF",
argLen: 1,
asm: arm.AMOVDF,
reg: regInfo{
inputs: []inputInfo{
{0, 4294901760}, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
outputs: []regMask{
4294901760, // F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15
},
},
},
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
{
name: "CALLstatic",
auxType: auxSymOff,
argLen: 1,
reg: regInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
clobbers: 8589875199, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
{
name: "CALLclosure",
auxType: auxInt64,
argLen: 3,
reg: regInfo{
inputs: []inputInfo{
{1, 128}, // R7
{0, 13311}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12 SP
},
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
clobbers: 8589875199, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
{
name: "CALLdefer",
auxType: auxInt64,
argLen: 1,
reg: regInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
clobbers: 8589875199, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
{
name: "CALLgo",
auxType: auxInt64,
argLen: 1,
reg: regInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
clobbers: 8589875199, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
{
name: "CALLinter",
auxType: auxInt64,
argLen: 2,
reg: regInfo{
inputs: []inputInfo{
{0, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
clobbers: 8589875199, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12 F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
{
name: "LoweredNilCheck",
argLen: 2,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
},
},
{
name: "Equal",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{0, 4294967296}, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "NotEqual",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{0, 4294967296}, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
},
},
{
name: "LessThan",
cmd/compile: more sanity checks on rewrite rules Make sure ops have the right number of args, set aux and auxint only if allowed, etc. Normalize error reporting format. Change-Id: Ie545fcc5990c8c7d62d40d9a0a55885f941eb645 Reviewed-on: https://go-review.googlesource.com/22320 Reviewed-by: David Chase <drchase@google.com>
2016-04-20 11:17:41 -07:00
argLen: 1,
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{0, 4294967296}, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "LessEqual",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{0, 4294967296}, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "GreaterThan",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{0, 4294967296}, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "GreaterEqual",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{0, 4294967296}, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "LessThanU",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{0, 4294967296}, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "LessEqualU",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{0, 4294967296}, // FLAGS
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
},
outputs: []regMask{
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "GreaterThanU",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{0, 4294967296}, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
{
name: "GreaterEqualU",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{0, 4294967296}, // FLAGS
[dev.ssa] cmd/compile: implement the following for SSA on ARM - generic Ops: Phi, CALL variants, NilCheck - generic Blocks: Plain, Check - 32-bit arithmetics - CMP and conditional branches - load/store - zero/sign-extensions (8 to 16, 8 to 32, 16 to 32) Progress on SSA backend for ARM. Still not complete. Now "errors" package compiles and tests passed. Updates #15365. Change-Id: If126fd17f8695cbf55d64085bb3f1a4a53205701 Reviewed-on: https://go-review.googlesource.com/22856 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-06 10:13:31 -07:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
},
},
},
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
{
name: "Carry",
argLen: 1,
reg: regInfo{
outputs: []regMask{
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
4294967296, // FLAGS
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
},
},
},
{
name: "LoweredSelect0",
argLen: 1,
reg: regInfo{
outputs: []regMask{
1, // R0
},
},
},
{
name: "LoweredSelect1",
argLen: 1,
resultInArg0: true,
reg: regInfo{
inputs: []inputInfo{
{0, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
[dev.ssa] cmd/compile: handle 64-bit shifts on ARM Also fix a mistake in previous CL about x8 and x16 shifts: the shift needs ZeroExt. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: Ibc352760023d38bc6b9c5251e929fe26e016637a Reviewed-on: https://go-review.googlesource.com/23486 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 23:17:42 -04:00
{
name: "LoweredZeromask",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: handle 64-bit shifts on ARM Also fix a mistake in previous CL about x8 and x16 shifts: the shift needs ZeroExt. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: Ibc352760023d38bc6b9c5251e929fe26e016637a Reviewed-on: https://go-review.googlesource.com/23486 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 23:17:42 -04:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
[dev.ssa] cmd/compile: implement Zero, Move, Copy for SSA on ARM Generate load/stores for small zeroing/move, DUFFZERO/DUFFCOPY for medium zeroing/move, and loops for large zeroing/move. cmd/compile/internal/gc/testdata/{copy_ssa.go,zero_ssa.go} tests passed. Progress on SSA backend for ARM. Still not complete. A few packages in the standard library compile and tests passed, including container/list, hash/crc32, unicode/utf8, etc. Updates #15365. Change-Id: Ieb4b68b44ee7de66bf7b68f5f33a605349fcc6fa Reviewed-on: https://go-review.googlesource.com/23097 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:31:14 -04:00
{
name: "DUFFZERO",
auxType: auxInt64,
argLen: 3,
reg: regInfo{
inputs: []inputInfo{
{0, 2}, // R1
{1, 1}, // R0
},
clobbers: 2, // R1
},
},
{
name: "DUFFCOPY",
auxType: auxInt64,
argLen: 3,
reg: regInfo{
inputs: []inputInfo{
{0, 4}, // R2
{1, 2}, // R1
},
clobbers: 7, // R0 R1 R2
},
},
{
name: "LoweredZero",
argLen: 4,
reg: regInfo{
inputs: []inputInfo{
{0, 2}, // R1
{1, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
{2, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
clobbers: 4294967298, // R1 FLAGS
[dev.ssa] cmd/compile: implement Zero, Move, Copy for SSA on ARM Generate load/stores for small zeroing/move, DUFFZERO/DUFFCOPY for medium zeroing/move, and loops for large zeroing/move. cmd/compile/internal/gc/testdata/{copy_ssa.go,zero_ssa.go} tests passed. Progress on SSA backend for ARM. Still not complete. A few packages in the standard library compile and tests passed, including container/list, hash/crc32, unicode/utf8, etc. Updates #15365. Change-Id: Ieb4b68b44ee7de66bf7b68f5f33a605349fcc6fa Reviewed-on: https://go-review.googlesource.com/23097 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:31:14 -04:00
},
},
{
name: "LoweredMove",
argLen: 4,
reg: regInfo{
inputs: []inputInfo{
{0, 4}, // R2
{1, 2}, // R1
{2, 5119}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
clobbers: 4294967302, // R1 R2 FLAGS
[dev.ssa] cmd/compile: implement Zero, Move, Copy for SSA on ARM Generate load/stores for small zeroing/move, DUFFZERO/DUFFCOPY for medium zeroing/move, and loops for large zeroing/move. cmd/compile/internal/gc/testdata/{copy_ssa.go,zero_ssa.go} tests passed. Progress on SSA backend for ARM. Still not complete. A few packages in the standard library compile and tests passed, including container/list, hash/crc32, unicode/utf8, etc. Updates #15365. Change-Id: Ieb4b68b44ee7de66bf7b68f5f33a605349fcc6fa Reviewed-on: https://go-review.googlesource.com/23097 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-13 15:31:14 -04:00
},
},
[dev.ssa] cmd/compile: handle Div, Convert, GetClosurePtr etc. on ARM This CL adds support of Div, Mod, Convert, GetClosurePtr and 64-bit indexing support to SSA backend for ARM. Add tests for 64-bit indexing to cmd/compile/internal/gc/testdata/string_ssa.go. Tests cmd/compile/internal/gc/testdata/*_ssa.go passed, except compound_ssa.go and fp_ssa.go. Progress on SSA for ARM. Still not complete. Essentially the only unsupported part is floating point. Updates #15365. Change-Id: I269e88b67f641c25e7a813d910c96d356d236bff Reviewed-on: https://go-review.googlesource.com/23542 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 09:49:28 -04:00
{
name: "LoweredGetClosurePtr",
argLen: 0,
reg: regInfo{
outputs: []regMask{
128, // R7
},
},
},
{
name: "MOVWconvert",
argLen: 2,
asm: arm.AMOVW,
reg: regInfo{
inputs: []inputInfo{
[dev.ssa] cmd/compile: handle GetG on ARM Use hardware g register (R10) for GetG, allow g to appear at LHS of some ops. Progress on SSA backend for ARM. Now everything compiles and runs. Updates #15365. Change-Id: Icdf93585579faa86cc29b1e17ab7c90f0119fc4e Reviewed-on: https://go-review.googlesource.com/23952 Reviewed-by: David Chase <drchase@google.com>
2016-05-31 14:01:34 -04:00
{0, 6143}, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 g R12
[dev.ssa] cmd/compile: handle Div, Convert, GetClosurePtr etc. on ARM This CL adds support of Div, Mod, Convert, GetClosurePtr and 64-bit indexing support to SSA backend for ARM. Add tests for 64-bit indexing to cmd/compile/internal/gc/testdata/string_ssa.go. Tests cmd/compile/internal/gc/testdata/*_ssa.go passed, except compound_ssa.go and fp_ssa.go. Progress on SSA for ARM. Still not complete. Essentially the only unsupported part is floating point. Updates #15365. Change-Id: I269e88b67f641c25e7a813d910c96d356d236bff Reviewed-on: https://go-review.googlesource.com/23542 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 09:49:28 -04:00
},
outputs: []regMask{
5119, // R0 R1 R2 R3 R4 R5 R6 R7 R8 R9 R12
},
},
},
cmd/compile: start on ARM port Start working on arm port. Gets close to correct code for fibonacci: func fib(n int) int { if n < 2 { return n } return fib(n-1) + fib(n-2) } Still a lot to do, but this is a good starting point. Cleaned up some arch-specific dependencies in regalloc. Change-Id: I4301c6c31a8402168e50dcfee8bcf7aee73ea9d5 Reviewed-on: https://go-review.googlesource.com/21000 Reviewed-by: David Chase <drchase@google.com>
2016-03-21 22:57:26 -07:00
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Add8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Add16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Add32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Add64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "AddPtr",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
{
name: "Add32F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
generic: true,
},
{
name: "Add64F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Sub8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Sub16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Sub32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Sub64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
[dev.ssa] cmd/compile: implement OSLICESTR Add a new function and generic operation to handle bounds checking for slices. Unlike the index bounds checking the index can be equal to the upper bound. Do gc-friendly slicing that generates proper code for 0-length result slices. This is a takeover of Alexandru's original change, (https://go-review.googlesource.com/#/c/12764/) submittable now that the decompose phase is in. Change-Id: I17d164cf42ed7839f84ca949c6ad3289269c9160 Reviewed-on: https://go-review.googlesource.com/13903 Reviewed-by: David Chase <drchase@google.com>
2015-08-24 23:52:03 -07:00
{
name: "SubPtr",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: implement OSLICESTR Add a new function and generic operation to handle bounds checking for slices. Unlike the index bounds checking the index can be equal to the upper bound. Do gc-friendly slicing that generates proper code for 0-length result slices. This is a takeover of Alexandru's original change, (https://go-review.googlesource.com/#/c/12764/) submittable now that the decompose phase is in. Change-Id: I17d164cf42ed7839f84ca949c6ad3289269c9160 Reviewed-on: https://go-review.googlesource.com/13903 Reviewed-by: David Chase <drchase@google.com>
2015-08-24 23:52:03 -07:00
generic: true,
},
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
{
name: "Sub32F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
generic: true,
},
{
name: "Sub64F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Mul8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Mul16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Mul32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Mul64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: implement multiplies Use width-and-signed-specific multiply opcodes. Implement OMUL. A few other cleanups. Fixes #11467 Change-Id: Ib0fe80a1a9b7208dbb8a2b6b652a478847f5d244 Reviewed-on: https://go-review.googlesource.com/12540 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-22 13:46:15 -07:00
},
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
{
name: "Mul32F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
generic: true,
},
{
name: "Mul64F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
generic: true,
},
{
name: "Div32F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
generic: true,
},
{
name: "Div64F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
{
name: "Hmul8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
generic: true,
},
{
name: "Hmul8u",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
generic: true,
},
{
name: "Hmul16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
generic: true,
},
{
name: "Hmul16u",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
generic: true,
},
{
name: "Hmul32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
generic: true,
},
{
name: "Hmul32u",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement OHMUL Adds support for high multiply which is used by the frontend when rewriting const division. The frontend currently only does this for 8, 16, and 32 bit integer arithmetic. Change-Id: I9b6c6018f3be827a50ee6c185454ebc79b3094c8 Reviewed-on: https://go-review.googlesource.com/13696 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:14:47 -05:00
generic: true,
},
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
{
name: "Hmul64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
generic: true,
},
{
name: "Hmul64u",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
generic: true,
},
{
name: "Avg64u",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: strength-reduce 64-bit constant divides The frontend does this for 32 bits and below, but SSA needs to do it for 64 bits. The algorithms are all copied from cgen.go:cgen_div. Speeds up TimeFormat substantially: ~40% slower to ~10% slower. Change-Id: I023ea2eb6040df98ccd9105e15ca6ea695610a7a Reviewed-on: https://go-review.googlesource.com/19302 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-02-05 20:26:18 -08:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
{
name: "Div8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
generic: true,
},
{
name: "Div8u",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
generic: true,
},
{
name: "Div16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
generic: true,
},
{
name: "Div16u",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
generic: true,
},
{
name: "Div32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
generic: true,
},
{
name: "Div32u",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
generic: true,
},
{
name: "Div64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
generic: true,
},
{
name: "Div64u",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement ODIV Implement integer division for non-consts. Change-Id: If40cbde20e5f0ebb9993064def7be468e4eca076 Reviewed-on: https://go-review.googlesource.com/13644 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-17 17:46:06 -05:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
{
name: "Mod8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
generic: true,
},
{
name: "Mod8u",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
generic: true,
},
{
name: "Mod16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
generic: true,
},
{
name: "Mod16u",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
generic: true,
},
{
name: "Mod32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
generic: true,
},
{
name: "Mod32u",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
generic: true,
},
{
name: "Mod64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
generic: true,
},
{
name: "Mod64u",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: implement OMOD Change-Id: Iec954c4daefef4ab3fa2c98bfb2c70b2dea8dffb Reviewed-on: https://go-review.googlesource.com/13743 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 19:51:44 -05:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "And8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "And16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "And32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "And64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OAND. Before this patch there was only partial support for ANDQconst which was not lowered. This patch added support for AND operations for all bit sizes and signs. Change-Id: I3a6b2cddfac5361b27e85fcd97f7f3537ebfbcb6 Reviewed-on: https://go-review.googlesource.com/12761 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-28 14:58:49 +02:00
},
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Or8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Or16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Or32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Or64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OOR. From compiling go there were 761 functions where OR was needed. Change-Id: Ied8bf59cec50a3175273387bc7416bd042def6d8 Reviewed-on: https://go-review.googlesource.com/12766 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:52:25 +02:00
},
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Xor8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Xor16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Xor32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Xor64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: implement lots of small (<8byte) ops. Lots and lots of ops! Also XOR for good measure. Add a pass to the compiler generator to check that all of the architecture-specific opcodes are handled by genValue. We will catch any missing ones if we come across them during compilation, but probably better to catch them statically. Change-Id: Ic4adfbec55c8257f88117bc732fa664486262868 Reviewed-on: https://go-review.googlesource.com/12813 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 16:04:50 -07:00
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
name: "Lsh8x8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Lsh8x16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Lsh8x32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Lsh8x64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Lsh16x8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Lsh16x16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Lsh16x32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Lsh16x64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Lsh32x8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Lsh32x16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Lsh32x32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
name: "Lsh32x64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
name: "Lsh64x8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
name: "Lsh64x16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
name: "Lsh64x32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
name: "Lsh64x64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
name: "Rsh8x8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
name: "Rsh8x16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
name: "Rsh8x32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
name: "Rsh8x64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
name: "Rsh16x8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
name: "Rsh16x16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh16x32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh16x64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh32x8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh32x16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh32x32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh32x64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh64x8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh64x16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh64x32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh64x64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh8Ux8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh8Ux16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh8Ux32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh8Ux64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh16Ux8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh16Ux16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh16Ux32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh16Ux64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh32Ux8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh32Ux16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh32Ux32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh32Ux64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh64Ux8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh64Ux16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh64Ux32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Rsh64Ux64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
{
name: "Lrot8",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
generic: true,
},
{
name: "Lrot16",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
generic: true,
},
{
name: "Lrot32",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
generic: true,
},
{
name: "Lrot64",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add support for LROT, and tests Hardcoded the limit on constants only allowed. Change-Id: Idb9b07b4871db7a752a79e492671e9b41207b956 Reviewed-on: https://go-review.googlesource.com/13257 Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-05 22:11:14 -04:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Eq8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Eq16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Eq32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Eq64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
},
[dev.ssa] cmd/compile: implement non-numeric comparisons The only slice/interface comparisons that reach the backend are comparisons to nil. Funcs, maps, and channels are references types, so pointer equality is enough. Change-Id: I60a71da46a36202e9bd62ed370ab7d7f2e2800e7 Reviewed-on: https://go-review.googlesource.com/12715 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-27 13:17:45 -07:00
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "EqPtr",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile: implement non-numeric comparisons The only slice/interface comparisons that reach the backend are comparisons to nil. Funcs, maps, and channels are references types, so pointer equality is enough. Change-Id: I60a71da46a36202e9bd62ed370ab7d7f2e2800e7 Reviewed-on: https://go-review.googlesource.com/12715 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-27 13:17:45 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: fix iface and slice comparisons A simpler way to do iface/slice comparisons. Fixes some cases of failed lowerings. Change-Id: Ia252bc8648293a2d460f63c41f1591785543a1e9 Reviewed-on: https://go-review.googlesource.com/14493 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-09-10 13:53:27 -07:00
name: "EqInter",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix iface and slice comparisons A simpler way to do iface/slice comparisons. Fixes some cases of failed lowerings. Change-Id: Ia252bc8648293a2d460f63c41f1591785543a1e9 Reviewed-on: https://go-review.googlesource.com/14493 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-09-10 13:53:27 -07:00
generic: true,
},
{
name: "EqSlice",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: implement non-numeric comparisons The only slice/interface comparisons that reach the backend are comparisons to nil. Funcs, maps, and channels are references types, so pointer equality is enough. Change-Id: I60a71da46a36202e9bd62ed370ab7d7f2e2800e7 Reviewed-on: https://go-review.googlesource.com/12715 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-27 13:17:45 -07:00
generic: true,
},
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
{
name: "Eq32F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
generic: true,
},
{
name: "Eq64F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Neq8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Neq16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Neq32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "Neq64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
},
[dev.ssa] cmd/compile: implement non-numeric comparisons The only slice/interface comparisons that reach the backend are comparisons to nil. Funcs, maps, and channels are references types, so pointer equality is enough. Change-Id: I60a71da46a36202e9bd62ed370ab7d7f2e2800e7 Reviewed-on: https://go-review.googlesource.com/12715 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-27 13:17:45 -07:00
{
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
name: "NeqPtr",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: handle commutative operations in cse * If a operation is commutative order the parameters in a canonical way. Size of pkg/tool/linux_amd64/* excluding compile: before: 95882288 after: 95868152 change: 14136 ~0.015% I tried something similar with Leq and Geq, but the results were not great because it confuses the 'lowered cse' pass too much which can no longer remove redundant comparisons from IsInBounds. Change-Id: I2f928663a11320bfc51c7fa47e384b7411c420ba Reviewed-on: https://go-review.googlesource.com/19727 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-22 11:19:15 +01:00
commutative: true,
generic: true,
[dev.ssa] cmd/compile: implement non-numeric comparisons The only slice/interface comparisons that reach the backend are comparisons to nil. Funcs, maps, and channels are references types, so pointer equality is enough. Change-Id: I60a71da46a36202e9bd62ed370ab7d7f2e2800e7 Reviewed-on: https://go-review.googlesource.com/12715 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-27 13:17:45 -07:00
},
{
[dev.ssa] cmd/compile/internal/ssa: fix iface and slice comparisons A simpler way to do iface/slice comparisons. Fixes some cases of failed lowerings. Change-Id: Ia252bc8648293a2d460f63c41f1591785543a1e9 Reviewed-on: https://go-review.googlesource.com/14493 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-09-10 13:53:27 -07:00
name: "NeqInter",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: fix iface and slice comparisons A simpler way to do iface/slice comparisons. Fixes some cases of failed lowerings. Change-Id: Ia252bc8648293a2d460f63c41f1591785543a1e9 Reviewed-on: https://go-review.googlesource.com/14493 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-09-10 13:53:27 -07:00
generic: true,
},
{
name: "NeqSlice",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: implement non-numeric comparisons The only slice/interface comparisons that reach the backend are comparisons to nil. Funcs, maps, and channels are references types, so pointer equality is enough. Change-Id: I60a71da46a36202e9bd62ed370ab7d7f2e2800e7 Reviewed-on: https://go-review.googlesource.com/12715 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-27 13:17:45 -07:00
generic: true,
},
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
{
name: "Neq32F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
generic: true,
},
{
name: "Neq64F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Less8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Less8U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Less16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Less16U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Less32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Less32U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Less64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Less64U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
{
name: "Less32F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
generic: true,
},
{
name: "Less64F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Leq8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Leq8U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Leq16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Leq16U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Leq32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Leq32U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Leq64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Leq64U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
{
name: "Leq32F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
generic: true,
},
{
name: "Leq64F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Greater8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Greater8U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Greater16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Greater16U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Greater32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Greater32U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Greater64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Greater64U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
generic: true,
},
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
{
name: "Greater32F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
generic: true,
},
{
name: "Greater64F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: use width and sign specific opcodes Bake the bit width and signedness into opcodes. Pro: Rewrite rules become easier. Less chance for confusion. Con: Lots more opcodes. Let me know what you think. I'm leaning towards this, but I could be convinced otherwise if people think this is too ugly. Update #11467 Change-Id: Icf1b894268cdf73515877bb123839800d97b9df9 Reviewed-on: https://go-review.googlesource.com/12362 Reviewed-by: Alan Donovan <adonovan@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-19 15:48:20 -07:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Geq8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Geq8U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
[dev.ssa] cmd/compile/ssa: add comparison ops Increase SSA coverage of functions in the standard library from 20.79% to 27.81%. The most significant unimplemented items are now: 10.16% 2597 SSA unimplemented: zero for type error not implemented 8.44% 2157 SSA unimplemented: addr: bad op DOTPTR 7.98% 2039 SSA unimplemented: unhandled OLITERAL 7 6.29% 1607 SSA unimplemented: unhandled expr OROR 4.73% 1209 SSA unimplemented: unhandled expr LEN 4.55% 1163 SSA unimplemented: unhandled expr LROT 3.42% 874 SSA unimplemented: unhandled OLITERAL 6 2.46% 629 SSA unimplemented: unhandled expr DOT 2.41% 615 SSA unimplemented: zero for type []byte not implemented 2.02% 516 SSA unimplemented: unhandled stmt CALLMETH 1.90% 486 SSA unimplemented: unhandled expr ANDAND 1.79% 458 SSA unimplemented: unhandled expr CALLINTER 1.69% 433 SSA unimplemented: unhandled stmt SWITCH 1.67% 428 SSA unimplemented: unhandled expr CALLMETH 1.67% 426 SSA unimplemented: unhandled expr CLOSUREVAR Change-Id: I40959b22993c4f70784b4eca472cae752347879c Reviewed-on: https://go-review.googlesource.com/11452 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-24 17:48:22 -07:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Geq16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/ssa: add comparison ops Increase SSA coverage of functions in the standard library from 20.79% to 27.81%. The most significant unimplemented items are now: 10.16% 2597 SSA unimplemented: zero for type error not implemented 8.44% 2157 SSA unimplemented: addr: bad op DOTPTR 7.98% 2039 SSA unimplemented: unhandled OLITERAL 7 6.29% 1607 SSA unimplemented: unhandled expr OROR 4.73% 1209 SSA unimplemented: unhandled expr LEN 4.55% 1163 SSA unimplemented: unhandled expr LROT 3.42% 874 SSA unimplemented: unhandled OLITERAL 6 2.46% 629 SSA unimplemented: unhandled expr DOT 2.41% 615 SSA unimplemented: zero for type []byte not implemented 2.02% 516 SSA unimplemented: unhandled stmt CALLMETH 1.90% 486 SSA unimplemented: unhandled expr ANDAND 1.79% 458 SSA unimplemented: unhandled expr CALLINTER 1.69% 433 SSA unimplemented: unhandled stmt SWITCH 1.67% 428 SSA unimplemented: unhandled expr CALLMETH 1.67% 426 SSA unimplemented: unhandled expr CLOSUREVAR Change-Id: I40959b22993c4f70784b4eca472cae752347879c Reviewed-on: https://go-review.googlesource.com/11452 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-24 17:48:22 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Geq16U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/ssa: add comparison ops Increase SSA coverage of functions in the standard library from 20.79% to 27.81%. The most significant unimplemented items are now: 10.16% 2597 SSA unimplemented: zero for type error not implemented 8.44% 2157 SSA unimplemented: addr: bad op DOTPTR 7.98% 2039 SSA unimplemented: unhandled OLITERAL 7 6.29% 1607 SSA unimplemented: unhandled expr OROR 4.73% 1209 SSA unimplemented: unhandled expr LEN 4.55% 1163 SSA unimplemented: unhandled expr LROT 3.42% 874 SSA unimplemented: unhandled OLITERAL 6 2.46% 629 SSA unimplemented: unhandled expr DOT 2.41% 615 SSA unimplemented: zero for type []byte not implemented 2.02% 516 SSA unimplemented: unhandled stmt CALLMETH 1.90% 486 SSA unimplemented: unhandled expr ANDAND 1.79% 458 SSA unimplemented: unhandled expr CALLINTER 1.69% 433 SSA unimplemented: unhandled stmt SWITCH 1.67% 428 SSA unimplemented: unhandled expr CALLMETH 1.67% 426 SSA unimplemented: unhandled expr CLOSUREVAR Change-Id: I40959b22993c4f70784b4eca472cae752347879c Reviewed-on: https://go-review.googlesource.com/11452 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-24 17:48:22 -07:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Geq32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/ssa: add comparison ops Increase SSA coverage of functions in the standard library from 20.79% to 27.81%. The most significant unimplemented items are now: 10.16% 2597 SSA unimplemented: zero for type error not implemented 8.44% 2157 SSA unimplemented: addr: bad op DOTPTR 7.98% 2039 SSA unimplemented: unhandled OLITERAL 7 6.29% 1607 SSA unimplemented: unhandled expr OROR 4.73% 1209 SSA unimplemented: unhandled expr LEN 4.55% 1163 SSA unimplemented: unhandled expr LROT 3.42% 874 SSA unimplemented: unhandled OLITERAL 6 2.46% 629 SSA unimplemented: unhandled expr DOT 2.41% 615 SSA unimplemented: zero for type []byte not implemented 2.02% 516 SSA unimplemented: unhandled stmt CALLMETH 1.90% 486 SSA unimplemented: unhandled expr ANDAND 1.79% 458 SSA unimplemented: unhandled expr CALLINTER 1.69% 433 SSA unimplemented: unhandled stmt SWITCH 1.67% 428 SSA unimplemented: unhandled expr CALLMETH 1.67% 426 SSA unimplemented: unhandled expr CLOSUREVAR Change-Id: I40959b22993c4f70784b4eca472cae752347879c Reviewed-on: https://go-review.googlesource.com/11452 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-24 17:48:22 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Geq32U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/ssa: add comparison ops Increase SSA coverage of functions in the standard library from 20.79% to 27.81%. The most significant unimplemented items are now: 10.16% 2597 SSA unimplemented: zero for type error not implemented 8.44% 2157 SSA unimplemented: addr: bad op DOTPTR 7.98% 2039 SSA unimplemented: unhandled OLITERAL 7 6.29% 1607 SSA unimplemented: unhandled expr OROR 4.73% 1209 SSA unimplemented: unhandled expr LEN 4.55% 1163 SSA unimplemented: unhandled expr LROT 3.42% 874 SSA unimplemented: unhandled OLITERAL 6 2.46% 629 SSA unimplemented: unhandled expr DOT 2.41% 615 SSA unimplemented: zero for type []byte not implemented 2.02% 516 SSA unimplemented: unhandled stmt CALLMETH 1.90% 486 SSA unimplemented: unhandled expr ANDAND 1.79% 458 SSA unimplemented: unhandled expr CALLINTER 1.69% 433 SSA unimplemented: unhandled stmt SWITCH 1.67% 428 SSA unimplemented: unhandled expr CALLMETH 1.67% 426 SSA unimplemented: unhandled expr CLOSUREVAR Change-Id: I40959b22993c4f70784b4eca472cae752347879c Reviewed-on: https://go-review.googlesource.com/11452 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-24 17:48:22 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Geq64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/ssa: add comparison ops Increase SSA coverage of functions in the standard library from 20.79% to 27.81%. The most significant unimplemented items are now: 10.16% 2597 SSA unimplemented: zero for type error not implemented 8.44% 2157 SSA unimplemented: addr: bad op DOTPTR 7.98% 2039 SSA unimplemented: unhandled OLITERAL 7 6.29% 1607 SSA unimplemented: unhandled expr OROR 4.73% 1209 SSA unimplemented: unhandled expr LEN 4.55% 1163 SSA unimplemented: unhandled expr LROT 3.42% 874 SSA unimplemented: unhandled OLITERAL 6 2.46% 629 SSA unimplemented: unhandled expr DOT 2.41% 615 SSA unimplemented: zero for type []byte not implemented 2.02% 516 SSA unimplemented: unhandled stmt CALLMETH 1.90% 486 SSA unimplemented: unhandled expr ANDAND 1.79% 458 SSA unimplemented: unhandled expr CALLINTER 1.69% 433 SSA unimplemented: unhandled stmt SWITCH 1.67% 428 SSA unimplemented: unhandled expr CALLMETH 1.67% 426 SSA unimplemented: unhandled expr CLOSUREVAR Change-Id: I40959b22993c4f70784b4eca472cae752347879c Reviewed-on: https://go-review.googlesource.com/11452 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-24 17:48:22 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Geq64U",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
{
name: "Geq32F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
generic: true,
},
{
name: "Geq64F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add FP comparison ops Basic ops, no particular optimization in the pattern matching yet (e.g. x!=x for Nan detection, x cmp constant, etc.) Change-Id: I0043564081d6dc0eede876c4a9eb3c33cbd1521c Reviewed-on: https://go-review.googlesource.com/13704 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-18 14:39:26 -04:00
generic: true,
},
cmd/compile: introduce bool operations. Introduce OrB, EqB, NeqB, AndB to handle bool operations. Change-Id: I53e4d5125a8090d5eeb4576db619103f19fff58d Reviewed-on: https://go-review.googlesource.com/22412 Reviewed-by: Keith Randall <khr@golang.org>
2016-04-24 21:21:07 +02:00
{
name: "AndB",
argLen: 2,
generic: true,
},
{
name: "OrB",
argLen: 2,
generic: true,
},
{
name: "EqB",
argLen: 2,
generic: true,
},
{
name: "NeqB",
argLen: 2,
generic: true,
},
[dev.ssa] cmd/compile: implement ONOT Co-hacking with josharian at Gophercon. Change-Id: Ia59dfab676c6ed598c2c25483439cd1395a4ea87 Reviewed-on: https://go-review.googlesource.com/12029 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
2015-07-10 11:25:48 -06:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Not",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: implement ONOT Co-hacking with josharian at Gophercon. Change-Id: Ia59dfab676c6ed598c2c25483439cd1395a4ea87 Reviewed-on: https://go-review.googlesource.com/12029 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
2015-07-10 11:25:48 -06:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
{
name: "Neg8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
generic: true,
},
{
name: "Neg16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
generic: true,
},
{
name: "Neg32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
generic: true,
},
{
name: "Neg64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa/gen: implement OMINUS Change-Id: Ibc645d6cf229ecc18af3549dd3750be9d7451abe Reviewed-on: https://go-review.googlesource.com/12472 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 16:58:18 +02:00
generic: true,
},
[dev.ssa] cmd/compile: add complex arithmetic Still to do: details, more testing corner cases. (e.g. negative zero) Includes small cleanups for previous CL. Note: complex division is currently done in the runtime, so the division code here is apparently not yet necessary and also not tested. Seems likely better to open code division and expose the widening/narrowing to optimization. Complex64 multiplication and division is done in wide format to avoid cancellation errors; for division, this also happens to be compatible with pre-SSA practice (which uses a single complex128 division function). It would-be-nice to widen for complex128 multiplication intermediates as well, but that is trickier to implement without a handy wider-precision format. Change-Id: I595a4300f68868fb7641852a54674c6b2b78855e Reviewed-on: https://go-review.googlesource.com/14028 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-28 14:24:10 -04:00
{
name: "Neg32F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add complex arithmetic Still to do: details, more testing corner cases. (e.g. negative zero) Includes small cleanups for previous CL. Note: complex division is currently done in the runtime, so the division code here is apparently not yet necessary and also not tested. Seems likely better to open code division and expose the widening/narrowing to optimization. Complex64 multiplication and division is done in wide format to avoid cancellation errors; for division, this also happens to be compatible with pre-SSA practice (which uses a single complex128 division function). It would-be-nice to widen for complex128 multiplication intermediates as well, but that is trickier to implement without a handy wider-precision format. Change-Id: I595a4300f68868fb7641852a54674c6b2b78855e Reviewed-on: https://go-review.googlesource.com/14028 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-28 14:24:10 -04:00
generic: true,
},
{
name: "Neg64F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add complex arithmetic Still to do: details, more testing corner cases. (e.g. negative zero) Includes small cleanups for previous CL. Note: complex division is currently done in the runtime, so the division code here is apparently not yet necessary and also not tested. Seems likely better to open code division and expose the widening/narrowing to optimization. Complex64 multiplication and division is done in wide format to avoid cancellation errors; for division, this also happens to be compatible with pre-SSA practice (which uses a single complex128 division function). It would-be-nice to widen for complex128 multiplication intermediates as well, but that is trickier to implement without a handy wider-precision format. Change-Id: I595a4300f68868fb7641852a54674c6b2b78855e Reviewed-on: https://go-review.googlesource.com/14028 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-28 14:24:10 -04:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
{
name: "Com8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Com16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Com32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
{
name: "Com64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: fix shift operations Convert shift ops to also encode the size of the shift amount. Change signed right shift from using CMOV to using bit twiddles. It is a little bit better (5 instructions instead of 4, but fewer bytes and slightly faster code). It's also a bit faster than the 4-instruction branch version, even with a very predictable branch. As tested on my machine, YMMV. Implement OCOM while we are here. Change-Id: I8ca12dd62fae5d626dc0e6da5d4bbd34fd9640d2 Reviewed-on: https://go-review.googlesource.com/12867 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-29 17:07:09 -07:00
generic: true,
},
cmd/compile: added some intrinsics to SSA back end One intrinsic was needed to help get the very best performance out of a future GC; as long as that one was being added, I also added Bswap since that is sometimes a handy thing to have. I had intended to fill out the bit-scan intrinsic family, but the mismatch between the "scan forward" instruction and "count leading zeroes" was large enough to cause me to leave it out -- it poses a dilemma that I'd rather dodge right now. These intrinsics are not exposed for general use. That's a separate issue requiring an API proposal change ( https://github.com/golang/proposal ) All intrinsics are tested, both that they are substituted on the appropriate architecture, and that they produce the expected result. Change-Id: I5848037cfd97de4f75bdc33bdd89bba00af4a8ee Reviewed-on: https://go-review.googlesource.com/20564 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-11 00:10:52 -05:00
{
name: "Ctz16",
argLen: 1,
generic: true,
},
{
name: "Ctz32",
argLen: 1,
generic: true,
},
{
name: "Ctz64",
argLen: 1,
generic: true,
},
{
name: "Clz16",
argLen: 1,
generic: true,
},
{
name: "Clz32",
argLen: 1,
generic: true,
},
{
name: "Clz64",
argLen: 1,
generic: true,
},
{
name: "Bswap32",
argLen: 1,
generic: true,
},
{
name: "Bswap64",
argLen: 1,
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: implement OSQRT Change-Id: Iec61ca1bdc064c29ceca6d47f600d5643d0a64dd Reviewed-on: https://go-review.googlesource.com/14533 Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
2015-09-12 13:26:57 -07:00
{
name: "Sqrt",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: implement OSQRT Change-Id: Iec61ca1bdc064c29ceca6d47f600d5643d0a64dd Reviewed-on: https://go-review.googlesource.com/14533 Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
2015-09-12 13:26:57 -07:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Phi",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: -1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Copy",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
[dev.ssa] cmd/compile: enhance SSA filtering, add OpConvert Modified GOSSA{HASH.PKG} environment variable filters to make it easier to make/run with all SSA for testing. Disable attempts at SSA for architectures that are not amd64 (avoid spurious errors/unimplementeds.) Removed easy out for unimplemented features. Add convert op for proper liveness in presence of uintptr to/from unsafe.Pointer conversions. Tweaked stack sizes to get a pass on windows; 1024 instead 768, was observed to pass at least once. Change-Id: Ida3800afcda67d529e3b1cf48ca4a3f0fa48b2c5 Reviewed-on: https://go-review.googlesource.com/16201 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com>
2015-10-19 11:36:07 -04:00
{
name: "Convert",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: enhance SSA filtering, add OpConvert Modified GOSSA{HASH.PKG} environment variable filters to make it easier to make/run with all SSA for testing. Disable attempts at SSA for architectures that are not amd64 (avoid spurious errors/unimplementeds.) Removed easy out for unimplemented features. Add convert op for proper liveness in presence of uintptr to/from unsafe.Pointer conversions. Tweaked stack sizes to get a pass on windows; 1024 instead 768, was observed to pass at least once. Change-Id: Ida3800afcda67d529e3b1cf48ca4a3f0fa48b2c5 Reviewed-on: https://go-review.googlesource.com/16201 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com>
2015-10-19 11:36:07 -04:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
name: "ConstBool",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxBool,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
generic: true,
},
{
name: "ConstString",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxString,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
generic: true,
},
{
name: "ConstNil",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
generic: true,
},
{
name: "Const8",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt8,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
generic: true,
},
{
name: "Const16",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt16,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
generic: true,
},
{
name: "Const32",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt32,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
generic: true,
},
{
name: "Const64",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile/internal/ssa: Split OpConst into an OpConst8, OpConst16, ... Convert the polymorphic OpConst into monomorphic variants. Change-Id: I90bb8894fbac04ca5e5484ea260c131ef8b506fb Reviewed-on: https://go-review.googlesource.com/12798 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-28 14:19:20 -07:00
generic: true,
},
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
{
name: "Const32F",
cmd/compile: const folding for float32/64 Split the auxFloat type into 32/64 bit versions and perform checking for exactly representable float32 values. Perform const folding on float32/64. Comment out some const negation rules that the frontend already performs. Change-Id: Ib3f8d59fa8b30e50fe0267786cfb3c50a06169d2 Reviewed-on: https://go-review.googlesource.com/20568 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
2016-03-11 19:36:54 -06:00
auxType: auxFloat32,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
generic: true,
},
{
name: "Const64F",
cmd/compile: const folding for float32/64 Split the auxFloat type into 32/64 bit versions and perform checking for exactly representable float32 values. Perform const folding on float32/64. Comment out some const negation rules that the frontend already performs. Change-Id: Ib3f8d59fa8b30e50fe0267786cfb3c50a06169d2 Reviewed-on: https://go-review.googlesource.com/20568 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
2016-03-11 19:36:54 -06:00
auxType: auxFloat64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile: first unoptimized cut at adding FP support Added F32 and F64 load, store, and addition. Added F32 and F64 multiply. Added F32 and F64 subtraction and division. Added X15 to "clobber" for FP sub/div Added FP constants Added separate FP test in gc/testdata Change-Id: Ifa60dbad948a40011b478d9605862c4b0cc9134c Reviewed-on: https://go-review.googlesource.com/13612 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 16:38:11 -04:00
generic: true,
},
[dev.ssa] cmd/compile: add decompose pass Decompose breaks compound objects up into pieces that can be operated on by the target architecture. The decompose pass only does phi ops, the rest is done by the rewrite rules in generic.rules. Compound objects include strings,slices,interfaces,structs,arrays. Arrays aren't decomposed because of indexing (we could support constant indexes, but dynamic indexes can't be handled using SSA). Structs will come in a subsequent CL. TODO: after this pass we have lost the association between, e.g., a string's pointer and its size. It would be nice if we could keep that information around for debugging info somehow. Change-Id: I6379ab962a7beef62297d0f68c421f22aa0a0901 Reviewed-on: https://go-review.googlesource.com/13683 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-18 10:26:28 -07:00
{
name: "ConstInterface",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile: add decompose pass Decompose breaks compound objects up into pieces that can be operated on by the target architecture. The decompose pass only does phi ops, the rest is done by the rewrite rules in generic.rules. Compound objects include strings,slices,interfaces,structs,arrays. Arrays aren't decomposed because of indexing (we could support constant indexes, but dynamic indexes can't be handled using SSA). Structs will come in a subsequent CL. TODO: after this pass we have lost the association between, e.g., a string's pointer and its size. It would be nice if we could keep that information around for debugging info somehow. Change-Id: I6379ab962a7beef62297d0f68c421f22aa0a0901 Reviewed-on: https://go-review.googlesource.com/13683 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-18 10:26:28 -07:00
generic: true,
},
{
name: "ConstSlice",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile: add decompose pass Decompose breaks compound objects up into pieces that can be operated on by the target architecture. The decompose pass only does phi ops, the rest is done by the rewrite rules in generic.rules. Compound objects include strings,slices,interfaces,structs,arrays. Arrays aren't decomposed because of indexing (we could support constant indexes, but dynamic indexes can't be handled using SSA). Structs will come in a subsequent CL. TODO: after this pass we have lost the association between, e.g., a string's pointer and its size. It would be nice if we could keep that information around for debugging info somehow. Change-Id: I6379ab962a7beef62297d0f68c421f22aa0a0901 Reviewed-on: https://go-review.googlesource.com/13683 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-18 10:26:28 -07:00
generic: true,
},
[dev.ssa] cmd/compile: start arguments as spilled Declare a function's arguments as having already been spilled so their use just requires a restore. Allow spill locations to be portions of larger objects the stack. Required to load portions of compound input arguments. Rename the memory input to InputMem. Use Arg for the pre-spilled argument values. Change-Id: I8fe2a03ffbba1022d98bfae2052b376b96d32dda Reviewed-on: https://go-review.googlesource.com/16536 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: David Chase <drchase@google.com>
2015-11-02 08:10:26 -08:00
{
name: "InitMem",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile: start arguments as spilled Declare a function's arguments as having already been spilled so their use just requires a restore. Allow spill locations to be portions of larger objects the stack. Required to load portions of compound input arguments. Rename the memory input to InputMem. Use Arg for the pre-spilled argument values. Change-Id: I8fe2a03ffbba1022d98bfae2052b376b96d32dda Reviewed-on: https://go-review.googlesource.com/16536 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: David Chase <drchase@google.com>
2015-11-02 08:10:26 -08:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Arg",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Addr",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxSym,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "SP",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "SB",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Func",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxSym,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Load",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Store",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Move",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: Initial implementation of memory zeroing This will make it possible for us to start implementing interfaces and other stack allocated types which are more than one machine word. Change-Id: I52b187a791cf1919cb70ed6dabdc9f57b317ea83 Reviewed-on: https://go-review.googlesource.com/11631 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-27 15:45:20 +01:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "Zero",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: Initial implementation of memory zeroing This will make it possible for us to start implementing interfaces and other stack allocated types which are more than one machine word. Change-Id: I52b187a791cf1919cb70ed6dabdc9f57b317ea83 Reviewed-on: https://go-review.googlesource.com/11631 Reviewed-by: Keith Randall <khr@golang.org>
2015-06-27 15:45:20 +01:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "ClosureCall",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "StaticCall",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxSymOff,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
[dev.ssa] cmd/compile/internal/gc: implement go and defer TODO: for now, just function calls. Do method and interface calls. Change-Id: Ib262dfa31cb753996cde899beaad4dc2e66705ac Reviewed-on: https://go-review.googlesource.com/14035 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-28 22:51:01 -07:00
{
name: "DeferCall",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/gc: implement go and defer TODO: for now, just function calls. Do method and interface calls. Change-Id: Ib262dfa31cb753996cde899beaad4dc2e66705ac Reviewed-on: https://go-review.googlesource.com/14035 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-28 22:51:01 -07:00
generic: true,
},
{
name: "GoCall",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/gc: implement go and defer TODO: for now, just function calls. Do method and interface calls. Change-Id: Ib262dfa31cb753996cde899beaad4dc2e66705ac Reviewed-on: https://go-review.googlesource.com/14035 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-28 22:51:01 -07:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: complete call ops OCALLINTER, as well as ODEFER/OPROC with OCALLMETH/OCALLINTER. Move all the call logic to its own routine, a lot of the code is shared. Change-Id: Ieac59596165e434cc6d1d7b5e46b78957e9c5ed3 Reviewed-on: https://go-review.googlesource.com/14464 Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2015-09-09 23:56:59 -07:00
{
name: "InterCall",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: complete call ops OCALLINTER, as well as ODEFER/OPROC with OCALLMETH/OCALLINTER. Move all the call logic to its own routine, a lot of the code is shared. Change-Id: Ieac59596165e434cc6d1d7b5e46b78957e9c5ed3 Reviewed-on: https://go-review.googlesource.com/14464 Reviewed-by: Todd Neal <todd@tneal.org> Reviewed-by: David Chase <drchase@google.com>
2015-09-09 23:56:59 -07:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
name: "SignExt8to16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "SignExt8to32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "SignExt8to64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "SignExt16to32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "SignExt16to64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "SignExt32to64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "ZeroExt8to16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "ZeroExt8to32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "ZeroExt8to64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "ZeroExt16to32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "ZeroExt16to64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "ZeroExt32to64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "Trunc16to8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "Trunc32to8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "Trunc32to16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "Trunc64to8",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "Trunc64to16",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: redo how sign extension is handled For integer types less than a machine register, we have to decide what the invariants are for the high bits of the register. We used to set the high bits to the correct extension (sign or zero, as determined by the type) of the low bits. This CL makes the compiler ignore the high bits of the register altogether (they are junk). On this plus side, this means ops that generate subword results don't have to worry about correctly extending them. On the minus side, ops that consume subword arguments have to deal with the input registers not being correctly extended. For x86, this tradeoff is probably worth it. Almost all opcodes have versions that use only the correct subword piece of their inputs. (The one big exception is array indexing.) Not many opcodes can correctly sign extend on output. For other architectures, the tradeoff is probably not so clear, as they don't have many subword-safe opcodes (e.g. 16-bit compare, ignoring the high 16/48 bits). Fortunately we can decide whether we do this per-architecture. For the machine-independent opcodes, we pretend that the "register" size is equal to the type width, so sign extension is immaterial. Opcodes that care about the signedness of the input (e.g. compare, right shift) have two different variants. Change-Id: I465484c5734545ee697afe83bc8bf4b53bd9df8d Reviewed-on: https://go-review.googlesource.com/12600 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-07-23 14:35:02 -07:00
generic: true,
},
{
name: "Trunc64to32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
{
name: "Cvt32to32F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
generic: true,
},
{
name: "Cvt32to64F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
generic: true,
},
{
name: "Cvt64to32F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
generic: true,
},
{
name: "Cvt64to64F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
generic: true,
},
{
name: "Cvt32Fto32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
generic: true,
},
{
name: "Cvt32Fto64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
generic: true,
},
{
name: "Cvt64Fto32",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
generic: true,
},
{
name: "Cvt64Fto64",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
generic: true,
},
{
name: "Cvt32Fto64F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
generic: true,
},
{
name: "Cvt64Fto32F",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add [u]int and FP to FP conversions Change-Id: I8c17f706a3e0f1fa2d754bfb4ccd1f7a027cb3db Reviewed-on: https://go-review.googlesource.com/13744 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-20 15:14:20 -04:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "IsNonNil",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "IsInBounds",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
[dev.ssa] cmd/compile: implement OSLICESTR Add a new function and generic operation to handle bounds checking for slices. Unlike the index bounds checking the index can be equal to the upper bound. Do gc-friendly slicing that generates proper code for 0-length result slices. This is a takeover of Alexandru's original change, (https://go-review.googlesource.com/#/c/12764/) submittable now that the decompose phase is in. Change-Id: I17d164cf42ed7839f84ca949c6ad3289269c9160 Reviewed-on: https://go-review.googlesource.com/13903 Reviewed-by: David Chase <drchase@google.com>
2015-08-24 23:52:03 -07:00
{
name: "IsSliceInBounds",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: implement OSLICESTR Add a new function and generic operation to handle bounds checking for slices. Unlike the index bounds checking the index can be equal to the upper bound. Do gc-friendly slicing that generates proper code for 0-length result slices. This is a takeover of Alexandru's original change, (https://go-review.googlesource.com/#/c/12764/) submittable now that the decompose phase is in. Change-Id: I17d164cf42ed7839f84ca949c6ad3289269c9160 Reviewed-on: https://go-review.googlesource.com/13903 Reviewed-by: David Chase <drchase@google.com>
2015-08-24 23:52:03 -07:00
generic: true,
},
[dev.ssa] cmd/compile: make failed nil checks panic Introduce pseudo-ops PanicMem and LoweredPanicMem. PanicMem could be rewritten directly into MOVL during lowering, but then we couldn't log nil checks. With this change, runnable nil check tests pass: GOSSAPKG=main go run run.go -- nil*.go Compiler output nil check tests fail: GOSSAPKG=p go run run.go -- nil*.go This is due to several factors: * SSA has improved elimination of unnecessary nil checks. * SSA is missing elimination of implicit nil checks. * SSA is missing extra logging about why nil checks were removed. I'm not sure how best to resolve these failures, particularly in a world in which the two backends will live side by side for some time. For now, punt on the problem. Change-Id: Ib2ca6824551671f92e0e1800b036f5ca0905e2a3 Reviewed-on: https://go-review.googlesource.com/13474 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-11 09:47:45 -07:00
{
[dev.ssa] cmd/compile: optimize nil checks Use faulting loads instead of test/jeq to do nil checks. Fold nil checks into a following load/store if possible. Makes binaries about 2% smaller. Change-Id: I54af0f0a93c853f37e34e0ce7e3f01dd2ac87f64 Reviewed-on: https://go-review.googlesource.com/16287 Reviewed-by: David Chase <drchase@google.com>
2015-10-23 19:12:49 -07:00
name: "NilCheck",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: make failed nil checks panic Introduce pseudo-ops PanicMem and LoweredPanicMem. PanicMem could be rewritten directly into MOVL during lowering, but then we couldn't log nil checks. With this change, runnable nil check tests pass: GOSSAPKG=main go run run.go -- nil*.go Compiler output nil check tests fail: GOSSAPKG=p go run run.go -- nil*.go This is due to several factors: * SSA has improved elimination of unnecessary nil checks. * SSA is missing elimination of implicit nil checks. * SSA is missing extra logging about why nil checks were removed. I'm not sure how best to resolve these failures, particularly in a world in which the two backends will live side by side for some time. For now, punt on the problem. Change-Id: Ib2ca6824551671f92e0e1800b036f5ca0905e2a3 Reviewed-on: https://go-review.googlesource.com/13474 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-11 09:47:45 -07:00
generic: true,
},
[dev.ssa] cmd/compile: implement OGETG Change-Id: I7ecf62cf399c710b4a617803c43e83fce09b8a7d Reviewed-on: https://go-review.googlesource.com/13585 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 11:22:16 -07:00
{
name: "GetG",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: implement OGETG Change-Id: I7ecf62cf399c710b4a617803c43e83fce09b8a7d Reviewed-on: https://go-review.googlesource.com/13585 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-12 11:22:16 -07:00
generic: true,
},
[dev.ssa] cmd/compile: addressed vars and closures Cleaned up first-block-in-function code. Added cases for |PHEAP for PPARAM and PAUTO. Made PPARAMOUT act more like PAUTO for purposes of address generation and vardef placement. Added cases for OCLOSUREVAR and Ops for getting closure pointer. Closure ops are scheduled at top of entry block to capture DX. Wrote test that seems to show proper behavior for addressed parameters, locals, and returns. Change-Id: Iee93ebf9e3d9f74cfb4d1c1da8038eb278d8a857 Reviewed-on: https://go-review.googlesource.com/14650 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com>
2015-09-11 16:40:05 -04:00
{
name: "GetClosurePtr",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile: addressed vars and closures Cleaned up first-block-in-function code. Added cases for |PHEAP for PPARAM and PAUTO. Made PPARAMOUT act more like PAUTO for purposes of address generation and vardef placement. Added cases for OCLOSUREVAR and Ops for getting closure pointer. Closure ops are scheduled at top of entry block to capture DX. Wrote test that seems to show proper behavior for addressed parameters, locals, and returns. Change-Id: Iee93ebf9e3d9f74cfb4d1c1da8038eb278d8a857 Reviewed-on: https://go-review.googlesource.com/14650 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: David Chase <drchase@google.com>
2015-09-11 16:40:05 -04:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "ArrayIndex",
cmd/compile: some SSA cleanup Do some easy TODOs. Move a bunch of other TODOs into bugs. Change-Id: Iaba9dad6221a2af11b3cbcc512875f4a85842873 Reviewed-on: https://go-review.googlesource.com/20114 Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-03-01 15:59:15 -08:00
auxType: auxInt64,
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "PtrIndex",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "OffPtr",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "SliceMake",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "SlicePtr",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "SliceLen",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "SliceCap",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
[dev.ssa] cmd/compile: add compose/decompose for complex, phi, constants Still to do: arithmetic Change-Id: I31fd23b34980c9ed4b4e304b8597134b2ba6ca5c Reviewed-on: https://go-review.googlesource.com/14024 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-28 14:24:10 -04:00
{
name: "ComplexMake",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add compose/decompose for complex, phi, constants Still to do: arithmetic Change-Id: I31fd23b34980c9ed4b4e304b8597134b2ba6ca5c Reviewed-on: https://go-review.googlesource.com/14024 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-28 14:24:10 -04:00
generic: true,
},
{
name: "ComplexReal",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add compose/decompose for complex, phi, constants Still to do: arithmetic Change-Id: I31fd23b34980c9ed4b4e304b8597134b2ba6ca5c Reviewed-on: https://go-review.googlesource.com/14024 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-28 14:24:10 -04:00
generic: true,
},
{
name: "ComplexImag",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add compose/decompose for complex, phi, constants Still to do: arithmetic Change-Id: I31fd23b34980c9ed4b4e304b8597134b2ba6ca5c Reviewed-on: https://go-review.googlesource.com/14024 Reviewed-by: Keith Randall <khr@golang.org>
2015-08-28 14:24:10 -04:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "StringMake",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "StringPtr",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "StringLen",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
[dev.ssa] cmd/compile: add decompose pass Decompose breaks compound objects up into pieces that can be operated on by the target architecture. The decompose pass only does phi ops, the rest is done by the rewrite rules in generic.rules. Compound objects include strings,slices,interfaces,structs,arrays. Arrays aren't decomposed because of indexing (we could support constant indexes, but dynamic indexes can't be handled using SSA). Structs will come in a subsequent CL. TODO: after this pass we have lost the association between, e.g., a string's pointer and its size. It would be nice if we could keep that information around for debugging info somehow. Change-Id: I6379ab962a7beef62297d0f68c421f22aa0a0901 Reviewed-on: https://go-review.googlesource.com/13683 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-18 10:26:28 -07:00
{
name: "IMake",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: add decompose pass Decompose breaks compound objects up into pieces that can be operated on by the target architecture. The decompose pass only does phi ops, the rest is done by the rewrite rules in generic.rules. Compound objects include strings,slices,interfaces,structs,arrays. Arrays aren't decomposed because of indexing (we could support constant indexes, but dynamic indexes can't be handled using SSA). Structs will come in a subsequent CL. TODO: after this pass we have lost the association between, e.g., a string's pointer and its size. It would be nice if we could keep that information around for debugging info somehow. Change-Id: I6379ab962a7beef62297d0f68c421f22aa0a0901 Reviewed-on: https://go-review.googlesource.com/13683 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-18 10:26:28 -07:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: implement ITAB Implement ITAB, selecting the itable field of an interface. Soften the lowering check to allow lowerings that leave generic but dead ops behind. (The ITAB lowering does this.) Change-Id: Icc84961dd4060d143602f001311aa1d8be0d7fc0 Reviewed-on: https://go-review.googlesource.com/13144 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-04 15:47:22 -07:00
{
name: "ITab",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: implement ITAB Implement ITAB, selecting the itable field of an interface. Soften the lowering check to allow lowerings that leave generic but dead ops behind. (The ITAB lowering does this.) Change-Id: Icc84961dd4060d143602f001311aa1d8be0d7fc0 Reviewed-on: https://go-review.googlesource.com/13144 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-04 15:47:22 -07:00
generic: true,
},
[dev.ssa] cmd/compile: add decompose pass Decompose breaks compound objects up into pieces that can be operated on by the target architecture. The decompose pass only does phi ops, the rest is done by the rewrite rules in generic.rules. Compound objects include strings,slices,interfaces,structs,arrays. Arrays aren't decomposed because of indexing (we could support constant indexes, but dynamic indexes can't be handled using SSA). Structs will come in a subsequent CL. TODO: after this pass we have lost the association between, e.g., a string's pointer and its size. It would be nice if we could keep that information around for debugging info somehow. Change-Id: I6379ab962a7beef62297d0f68c421f22aa0a0901 Reviewed-on: https://go-review.googlesource.com/13683 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-18 10:26:28 -07:00
{
name: "IData",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: add decompose pass Decompose breaks compound objects up into pieces that can be operated on by the target architecture. The decompose pass only does phi ops, the rest is done by the rewrite rules in generic.rules. Compound objects include strings,slices,interfaces,structs,arrays. Arrays aren't decomposed because of indexing (we could support constant indexes, but dynamic indexes can't be handled using SSA). Structs will come in a subsequent CL. TODO: after this pass we have lost the association between, e.g., a string's pointer and its size. It would be nice if we could keep that information around for debugging info somehow. Change-Id: I6379ab962a7beef62297d0f68c421f22aa0a0901 Reviewed-on: https://go-review.googlesource.com/13683 Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-18 10:26:28 -07:00
generic: true,
},
[dev.ssa] cmd/compile: Allow structs to be SSAd Break small structs up into their components so they can be registerized. Change StructSelect to use field indexes instead of field offsets, as field offsets aren't unique in the presence of zero-sized fields. Change-Id: I2f1dc89f7fa58e1cf58aa1a32b238959d53f62e4 Reviewed-on: https://go-review.googlesource.com/18570 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-11 21:05:33 -08:00
{
name: "StructMake0",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile: Allow structs to be SSAd Break small structs up into their components so they can be registerized. Change StructSelect to use field indexes instead of field offsets, as field offsets aren't unique in the presence of zero-sized fields. Change-Id: I2f1dc89f7fa58e1cf58aa1a32b238959d53f62e4 Reviewed-on: https://go-review.googlesource.com/18570 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-11 21:05:33 -08:00
generic: true,
},
{
name: "StructMake1",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: Allow structs to be SSAd Break small structs up into their components so they can be registerized. Change StructSelect to use field indexes instead of field offsets, as field offsets aren't unique in the presence of zero-sized fields. Change-Id: I2f1dc89f7fa58e1cf58aa1a32b238959d53f62e4 Reviewed-on: https://go-review.googlesource.com/18570 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-11 21:05:33 -08:00
generic: true,
},
{
name: "StructMake2",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 2,
[dev.ssa] cmd/compile: Allow structs to be SSAd Break small structs up into their components so they can be registerized. Change StructSelect to use field indexes instead of field offsets, as field offsets aren't unique in the presence of zero-sized fields. Change-Id: I2f1dc89f7fa58e1cf58aa1a32b238959d53f62e4 Reviewed-on: https://go-review.googlesource.com/18570 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-11 21:05:33 -08:00
generic: true,
},
{
name: "StructMake3",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 3,
[dev.ssa] cmd/compile: Allow structs to be SSAd Break small structs up into their components so they can be registerized. Change StructSelect to use field indexes instead of field offsets, as field offsets aren't unique in the presence of zero-sized fields. Change-Id: I2f1dc89f7fa58e1cf58aa1a32b238959d53f62e4 Reviewed-on: https://go-review.googlesource.com/18570 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-11 21:05:33 -08:00
generic: true,
},
{
name: "StructMake4",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 4,
[dev.ssa] cmd/compile: Allow structs to be SSAd Break small structs up into their components so they can be registerized. Change StructSelect to use field indexes instead of field offsets, as field offsets aren't unique in the presence of zero-sized fields. Change-Id: I2f1dc89f7fa58e1cf58aa1a32b238959d53f62e4 Reviewed-on: https://go-review.googlesource.com/18570 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-11 21:05:33 -08:00
generic: true,
},
{
name: "StructSelect",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxInt64,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile: Allow structs to be SSAd Break small structs up into their components so they can be registerized. Change StructSelect to use field indexes instead of field offsets, as field offsets aren't unique in the presence of zero-sized fields. Change-Id: I2f1dc89f7fa58e1cf58aa1a32b238959d53f62e4 Reviewed-on: https://go-review.googlesource.com/18570 Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: David Chase <drchase@google.com>
2016-01-11 21:05:33 -08:00
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
{
[dev.ssa] cmd/compile: respect stack slot width when storing/loading registers Prior to this, we were smashing our own stack, which caused the crypto/sha256 tests to fail. Change-Id: I7dd94cf466d175b3be0cd65f9c4fe8b1223081fe Reviewed-on: https://go-review.googlesource.com/12660 Reviewed-by: Daniel Morsing <daniel.morsing@gmail.com> Reviewed-by: Keith Randall <khr@golang.org>
2015-07-24 14:51:51 -07:00
name: "StoreReg",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: respect stack slot width when storing/loading registers Prior to this, we were smashing our own stack, which caused the crypto/sha256 tests to fail. Change-Id: I7dd94cf466d175b3be0cd65f9c4fe8b1223081fe Reviewed-on: https://go-review.googlesource.com/12660 Reviewed-by: Daniel Morsing <daniel.morsing@gmail.com> Reviewed-by: Keith Randall <khr@golang.org>
2015-07-24 14:51:51 -07:00
name: "LoadReg",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
{
[dev.ssa] cmd/compile: don't generate zero values for ssa ops Shorter code, easier to read, no pointless empty slices. Change-Id: Id410364b4f6924b5665188af3373a5e914117c38 Reviewed-on: https://go-review.googlesource.com/12480 Reviewed-by: Keith Randall <khr@golang.org>
2015-07-21 07:10:56 -07:00
name: "FwdRef",
cmd/compile: keep pointer input arguments live throughout function Introduce a KeepAlive op which makes sure that its argument is kept live until the KeepAlive. Use KeepAlive to mark pointer input arguments as live after each function call and at each return. We do this change only for pointer arguments. Those are the critical ones to handle because they might have finalizers. Doing compound arguments (slices, structs, ...) is more complicated because we would need to track field liveness individually (we do that for auto variables now, but inputs requires extra trickery). Turn off the automatic marking of args as live. That way, when args are explicitly nulled, plive will know that the original argument is dead. The KeepAlive op will be the eventual implementation of runtime.KeepAlive. Fixes #15277 Change-Id: I5f223e65d99c9f8342c03fbb1512c4d363e903e5 Reviewed-on: https://go-review.googlesource.com/22365 Reviewed-by: David Chase <drchase@google.com> Reviewed-by: Russ Cox <rsc@golang.org>
2016-04-21 19:28:28 -07:00
auxType: auxSym,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
generic: true,
},
[dev.ssa] cmd/compile: optimize phi ops Redo how we keep track of forward references when building SSA. When the forward reference is resolved, update the Value node in place. Improve the phi elimination pass so it can simplify phis of phis. Give SSA package access to decoded line numbers. Fix line numbers for constant booleans. Change-Id: I3dc9896148d260be2f3dd14cbe5db639ec9fa6b7 Reviewed-on: https://go-review.googlesource.com/18674 Reviewed-by: David Chase <drchase@google.com> Run-TryBot: Keith Randall <khr@golang.org>
2016-01-14 16:02:23 -08:00
{
name: "Unknown",
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 0,
[dev.ssa] cmd/compile: optimize phi ops Redo how we keep track of forward references when building SSA. When the forward reference is resolved, update the Value node in place. Improve the phi elimination pass so it can simplify phis of phis. Give SSA package access to decoded line numbers. Fix line numbers for constant booleans. Change-Id: I3dc9896148d260be2f3dd14cbe5db639ec9fa6b7 Reviewed-on: https://go-review.googlesource.com/18674 Reviewed-by: David Chase <drchase@google.com> Run-TryBot: Keith Randall <khr@golang.org>
2016-01-14 16:02:23 -08:00
generic: true,
},
[dev.ssa] cmd/runtime: generate gc bitmaps for SSA-compiled code This change is all about leveraging the gc bitmap generation that is already done by the current compiler. We rearrange how stack allocation is done so that we generate a variable declaration for each spill. We also reorganize how args/locals are recorded during SSA. Then we can use the existing allocauto/defframe to allocate the stack frame and liveness to make the gc bitmaps. With this change, stack copying works correctly and we no longer need hacks in runtime/stack*.go to make tests work. GC is close to working, it just needs write barriers. Change-Id: I990fb4e3fbe98850c6be35c3185a1c85d9e1a6ba Reviewed-on: https://go-review.googlesource.com/13894 Reviewed-by: David Chase <drchase@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-24 02:16:19 -07:00
{
name: "VarDef",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxSym,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/runtime: generate gc bitmaps for SSA-compiled code This change is all about leveraging the gc bitmap generation that is already done by the current compiler. We rearrange how stack allocation is done so that we generate a variable declaration for each spill. We also reorganize how args/locals are recorded during SSA. Then we can use the existing allocauto/defframe to allocate the stack frame and liveness to make the gc bitmaps. With this change, stack copying works correctly and we no longer need hacks in runtime/stack*.go to make tests work. GC is close to working, it just needs write barriers. Change-Id: I990fb4e3fbe98850c6be35c3185a1c85d9e1a6ba Reviewed-on: https://go-review.googlesource.com/13894 Reviewed-by: David Chase <drchase@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-24 02:16:19 -07:00
generic: true,
},
{
name: "VarKill",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxSym,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] cmd/runtime: generate gc bitmaps for SSA-compiled code This change is all about leveraging the gc bitmap generation that is already done by the current compiler. We rearrange how stack allocation is done so that we generate a variable declaration for each spill. We also reorganize how args/locals are recorded during SSA. Then we can use the existing allocauto/defframe to allocate the stack frame and liveness to make the gc bitmaps. With this change, stack copying works correctly and we no longer need hacks in runtime/stack*.go to make tests work. GC is close to working, it just needs write barriers. Change-Id: I990fb4e3fbe98850c6be35c3185a1c85d9e1a6ba Reviewed-on: https://go-review.googlesource.com/13894 Reviewed-by: David Chase <drchase@google.com> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
2015-08-24 02:16:19 -07:00
generic: true,
},
[dev.ssa] Merge remote-tracking branch 'origin/master' into mergebranch Semi-regular merge from tip to dev.ssa. Conflicts: src/runtime/sys_windows_amd64.s Change-Id: I5f733130049c810e6ceacd46dad85faebca52b29
2016-01-19 09:59:21 -08:00
{
name: "VarLive",
[dev.ssa] cmd/compile: add aux typing, flags to ops Add the aux type to opcodes. Add rematerializeable as a flag. Change-Id: I906e19281498f3ee51bb136299bf26e13a54b2ec Reviewed-on: https://go-review.googlesource.com/19088 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Todd Neal <todd@tneal.org>
2016-01-31 11:39:39 -08:00
auxType: auxSym,
[dev.ssa] cmd/compile: add max arg length to opcodes Add the max arg length to opcodes and use it in zcse. Doesn't affect speed, but allows better checking in checkFunc and removes the need to keep a list of zero arg opcodes up to date. Change-Id: I157c6587154604119720ec6228b767b6e52bb5c7 Reviewed-on: https://go-review.googlesource.com/19994 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Todd Neal <todd@tneal.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-02-27 08:04:48 -06:00
argLen: 1,
[dev.ssa] Merge remote-tracking branch 'origin/master' into mergebranch Semi-regular merge from tip to dev.ssa. Conflicts: src/runtime/sys_windows_amd64.s Change-Id: I5f733130049c810e6ceacd46dad85faebca52b29
2016-01-19 09:59:21 -08:00
generic: true,
},
cmd/compile: keep pointer input arguments live throughout function Introduce a KeepAlive op which makes sure that its argument is kept live until the KeepAlive. Use KeepAlive to mark pointer input arguments as live after each function call and at each return. We do this change only for pointer arguments. Those are the critical ones to handle because they might have finalizers. Doing compound arguments (slices, structs, ...) is more complicated because we would need to track field liveness individually (we do that for auto variables now, but inputs requires extra trickery). Turn off the automatic marking of args as live. That way, when args are explicitly nulled, plive will know that the original argument is dead. The KeepAlive op will be the eventual implementation of runtime.KeepAlive. Fixes #15277 Change-Id: I5f223e65d99c9f8342c03fbb1512c4d363e903e5 Reviewed-on: https://go-review.googlesource.com/22365 Reviewed-by: David Chase <drchase@google.com> Reviewed-by: Russ Cox <rsc@golang.org>
2016-04-21 19:28:28 -07:00
{
name: "KeepAlive",
argLen: 2,
generic: true,
},
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
{
name: "Int64Make",
argLen: 2,
generic: true,
},
{
name: "Int64Hi",
argLen: 1,
generic: true,
},
{
name: "Int64Lo",
argLen: 1,
generic: true,
},
{
name: "Add32carry",
argLen: 2,
commutative: true,
generic: true,
},
{
name: "Add32withcarry",
argLen: 3,
commutative: true,
generic: true,
},
{
name: "Sub32carry",
argLen: 2,
generic: true,
},
{
name: "Sub32withcarry",
argLen: 3,
generic: true,
},
{
name: "Mul32uhilo",
argLen: 2,
generic: true,
},
{
name: "Signmask",
argLen: 1,
generic: true,
},
[dev.ssa] cmd/compile: handle 64-bit shifts on ARM Also fix a mistake in previous CL about x8 and x16 shifts: the shift needs ZeroExt. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: Ibc352760023d38bc6b9c5251e929fe26e016637a Reviewed-on: https://go-review.googlesource.com/23486 Reviewed-by: David Chase <drchase@google.com>
2016-05-25 23:17:42 -04:00
{
name: "Zeromask",
argLen: 1,
generic: true,
},
[dev.ssa] cmd/compile: handle floating point on ARM Machine supports (or the runtime simulates in soft float mode) (u)int32<->float conversions. The frontend rewrites int64<->float conversions to call to runtime function. For int64->float32 conversion, the frontend generates . . AS u(100) l(10) tc(1) . . . NAME-main.~r1 u(1) a(true) g(1) l(9) x(8+0) class(PPARAMOUT) f(1) float32 . . . CALLFUNC u(100) l(10) tc(1) float32 . . . . NAME-runtime.int64tofloat64 u(1) a(true) x(0+0) class(PFUNC) tc(1) used(true) FUNC-func(int64) float64 The CALLFUNC node has type float32, whereas runtime.int64tofloat64 returns float64. The legacy backend implicitly makes a float64->float32 conversion. The SSA backend does not do implicit conversion, so we insert an explicit CONV here. All cmd/compile/internal/gc/testdata/*_ssa.go tests passed. Progress on SSA for ARM. Still not complete. Update #15365. Change-Id: I30937c8ff977271246b068f48224693776804339 Reviewed-on: https://go-review.googlesource.com/23652 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-31 11:27:16 -04:00
{
name: "Cvt32Uto32F",
argLen: 1,
generic: true,
},
{
name: "Cvt32Uto64F",
argLen: 1,
generic: true,
},
{
name: "Cvt32Fto32U",
argLen: 1,
generic: true,
},
{
name: "Cvt64Fto32U",
argLen: 1,
generic: true,
},
[dev.ssa] cmd/compile: decompose 64-bit integer on ARM Introduce dec64 rules to (generically) decompose 64-bit integer on 32-bit architectures. 64-bit integer is composed/decomposed with Int64Make/Hi/Lo ops, as for complex types. The idea of dealing with Add64 is the following: (Add64 (Int64Make xh xl) (Int64Make yh yl)) -> (Int64Make (Add32withcarry xh yh (Select0 (Add32carry xl yl))) (Select1 (Add32carry xl yl))) where Add32carry returns a tuple (flags,uint32). Select0 and Select1 read the first and the second component of the tuple, respectively. The two Add32carry will be CSE'd. Similarly for multiplication, Mul32uhilo returns a tuple (hi, lo). Also add support of KeepAlive, to fix build after merge. Tests addressed_ssa.go, array_ssa.go, break_ssa.go, chan_ssa.go, cmp_ssa.go, ctl_ssa.go, map_ssa.go, and string_ssa.go in cmd/compile/internal/gc/testdata passed. Progress on SSA for ARM. Still not complete. Updates #15365. Change-Id: I7867c76785a456312de5d8398a6b3f7ca5a4f7ec Reviewed-on: https://go-review.googlesource.com/23213 Reviewed-by: Keith Randall <khr@golang.org>
2016-05-18 18:14:36 -04:00
{
name: "Select0",
argLen: 1,
generic: true,
},
{
name: "Select1",
argLen: 1,
generic: true,
},
[dev.ssa] cmd/compile/internal/ssa: autogenerate opcodes Revamp autogeneration. Get rid of gogenerate commands, they are more trouble than they are worth. (If the code won't compile, gogenerate doesn't work.) Generate opcode enums & tables. This means we only have to specify opcodes in one place instead of two. Add arch prefixes to opcodes so they will be globally unique. Change-Id: I175d0a89b701b2377bbe699f3756731b7c9f5a9f Reviewed-on: https://go-review.googlesource.com/10812 Reviewed-by: Alan Donovan <adonovan@google.com>
2015-06-06 16:03:33 -07:00
[dev.ssa] cmd/compile: Add initial SSA configuration for PPC64 This adds the initial SSA implementation for PPC64. Builds golang and all.bash runs correctly. Simple hello.go builds but does not run. Change-Id: I7cec211b934cd7a2dd75a6cdfaf9f71867063466 Reviewed-on: https://go-review.googlesource.com/24453 Reviewed-by: David Chase <drchase@google.com> Reviewed-by: Cherry Zhang <cherryyz@google.com> Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-06-24 14:37:17 -05:00
{
name: "ADD",
argLen: 2,
commutative: true,
asm: ppc64.AADD,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "ADDconst",
auxType: auxSymOff,
argLen: 1,
asm: ppc64.AADD,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "FADD",
argLen: 2,
commutative: true,
asm: ppc64.AFADD,
reg: regInfo{
inputs: []inputInfo{
{0, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
{1, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
outputs: []regMask{
576460743713488896, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
},
},
{
name: "FADDS",
argLen: 2,
commutative: true,
asm: ppc64.AFADDS,
reg: regInfo{
inputs: []inputInfo{
{0, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
{1, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
outputs: []regMask{
576460743713488896, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
},
},
{
name: "SUB",
argLen: 2,
asm: ppc64.ASUB,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "FSUB",
argLen: 2,
asm: ppc64.AFSUB,
reg: regInfo{
inputs: []inputInfo{
{0, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
{1, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
outputs: []regMask{
576460743713488896, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
},
},
{
name: "FSUBS",
argLen: 2,
asm: ppc64.AFSUBS,
reg: regInfo{
inputs: []inputInfo{
{0, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
{1, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
outputs: []regMask{
576460743713488896, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
},
},
{
name: "MULLD",
argLen: 2,
commutative: true,
asm: ppc64.AMULLD,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MULLW",
argLen: 2,
commutative: true,
asm: ppc64.AMULLW,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "FMUL",
argLen: 2,
commutative: true,
asm: ppc64.AFMUL,
reg: regInfo{
inputs: []inputInfo{
{0, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
{1, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
outputs: []regMask{
576460743713488896, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
},
},
{
name: "FMULS",
argLen: 2,
commutative: true,
asm: ppc64.AFMULS,
reg: regInfo{
inputs: []inputInfo{
{0, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
{1, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
outputs: []regMask{
576460743713488896, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
},
},
{
name: "FDIV",
argLen: 2,
asm: ppc64.AFDIV,
reg: regInfo{
inputs: []inputInfo{
{0, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
{1, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
outputs: []regMask{
576460743713488896, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
},
},
{
name: "FDIVS",
argLen: 2,
asm: ppc64.AFDIVS,
reg: regInfo{
inputs: []inputInfo{
{0, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
{1, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
outputs: []regMask{
576460743713488896, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
},
},
{
name: "AND",
argLen: 2,
commutative: true,
asm: ppc64.AAND,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "ANDconst",
auxType: auxInt32,
argLen: 1,
asm: ppc64.AAND,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "OR",
argLen: 2,
commutative: true,
asm: ppc64.AOR,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "ORconst",
auxType: auxInt32,
argLen: 1,
asm: ppc64.AOR,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "XOR",
argLen: 2,
commutative: true,
asm: ppc64.AXOR,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "XORconst",
auxType: auxInt32,
argLen: 1,
asm: ppc64.AXOR,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "NEG",
argLen: 1,
asm: ppc64.ANEG,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVBreg",
argLen: 1,
asm: ppc64.AMOVB,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVBZreg",
argLen: 1,
asm: ppc64.AMOVBZ,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVHreg",
argLen: 1,
asm: ppc64.AMOVH,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVHZreg",
argLen: 1,
asm: ppc64.AMOVHZ,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVWreg",
argLen: 1,
asm: ppc64.AMOVW,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVWZreg",
argLen: 1,
asm: ppc64.AMOVWZ,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVBload",
argLen: 2,
asm: ppc64.AMOVB,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVBZload",
argLen: 2,
asm: ppc64.AMOVBZ,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVHload",
argLen: 2,
asm: ppc64.AMOVH,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVHZload",
argLen: 2,
asm: ppc64.AMOVHZ,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVWload",
argLen: 2,
asm: ppc64.AMOVW,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVWZload",
argLen: 2,
asm: ppc64.AMOVWZ,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVDload",
argLen: 2,
asm: ppc64.AMOVD,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "FMOVDload",
argLen: 2,
asm: ppc64.AFMOVD,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
576460743713488896, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
},
},
{
name: "FMOVSload",
argLen: 2,
asm: ppc64.AFMOVS,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
576460743713488896, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
},
},
{
name: "MOVBstore",
auxType: auxSymOff,
argLen: 3,
asm: ppc64.AMOVB,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVHstore",
auxType: auxSymOff,
argLen: 3,
asm: ppc64.AMOVH,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVWstore",
auxType: auxSymOff,
argLen: 3,
asm: ppc64.AMOVW,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVDstore",
auxType: auxSymOff,
argLen: 3,
asm: ppc64.AMOVD,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "FMOVDstore",
auxType: auxSymOff,
argLen: 3,
asm: ppc64.AFMOVD,
reg: regInfo{
inputs: []inputInfo{
{0, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "FMOVSstore",
auxType: auxSymOff,
argLen: 3,
asm: ppc64.AFMOVS,
reg: regInfo{
inputs: []inputInfo{
{0, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVBstoreconst",
auxType: auxSymValAndOff,
argLen: 2,
asm: ppc64.AMOVB,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVHstoreconst",
auxType: auxSymValAndOff,
argLen: 2,
asm: ppc64.AMOVH,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVWstoreconst",
auxType: auxSymValAndOff,
argLen: 2,
asm: ppc64.AMOVW,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVDstoreconst",
auxType: auxSymValAndOff,
argLen: 2,
asm: ppc64.AMOVD,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVDconst",
auxType: auxInt64,
argLen: 0,
rematerializeable: true,
asm: ppc64.AMOVD,
reg: regInfo{
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVWconst",
auxType: auxInt32,
argLen: 0,
rematerializeable: true,
asm: ppc64.AMOVW,
reg: regInfo{
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVHconst",
auxType: auxInt16,
argLen: 0,
rematerializeable: true,
asm: ppc64.AMOVH,
reg: regInfo{
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "MOVBconst",
auxType: auxInt8,
argLen: 0,
rematerializeable: true,
asm: ppc64.AMOVB,
reg: regInfo{
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "FMOVDconst",
auxType: auxFloat64,
argLen: 0,
rematerializeable: true,
asm: ppc64.AFMOVD,
reg: regInfo{
outputs: []regMask{
576460743713488896, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
},
},
{
name: "FMOVSconst",
auxType: auxFloat32,
argLen: 0,
rematerializeable: true,
asm: ppc64.AFMOVS,
reg: regInfo{
outputs: []regMask{
576460743713488896, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
},
},
{
name: "FCMPU",
argLen: 2,
asm: ppc64.AFCMPU,
reg: regInfo{
inputs: []inputInfo{
{0, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
{1, 576460743713488896}, // F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
outputs: []regMask{
576460752303423488, // CR
},
},
},
{
name: "CMP",
argLen: 2,
asm: ppc64.ACMP,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
576460752303423488, // CR
},
},
},
{
name: "CMPU",
argLen: 2,
asm: ppc64.ACMPU,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
576460752303423488, // CR
},
},
},
{
name: "CMPW",
argLen: 2,
asm: ppc64.ACMPW,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
576460752303423488, // CR
},
},
},
{
name: "CMPWU",
argLen: 2,
asm: ppc64.ACMPWU,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
{1, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
576460752303423488, // CR
},
},
},
{
name: "CMPconst",
auxType: auxInt32,
argLen: 1,
asm: ppc64.ACMP,
reg: regInfo{
inputs: []inputInfo{
{0, 1073731578}, // SP R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
outputs: []regMask{
576460752303423488, // CR
},
},
},
{
name: "CALLstatic",
auxType: auxSymOff,
argLen: 1,
reg: regInfo{
clobbers: 576460744787220472, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26
},
},
{
name: "Equal",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
{0, 576460752303423488}, // CR
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "NotEqual",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
{0, 576460752303423488}, // CR
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "LessThan",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
{0, 576460752303423488}, // CR
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "LessEqual",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
{0, 576460752303423488}, // CR
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "GreaterThan",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
{0, 576460752303423488}, // CR
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
{
name: "GreaterEqual",
argLen: 1,
reg: regInfo{
inputs: []inputInfo{
{0, 576460752303423488}, // CR
},
outputs: []regMask{
1073731576, // R3 R4 R5 R6 R7 R8 R9 R10 R12 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29
},
},
},
}
func (o Op) Asm() obj.As { return opcodeTable[o].asm }
func (o Op) String() string { return opcodeTable[o].name }
var registersAMD64 = [...]Register{
{0, "AX"},
{1, "CX"},
{2, "DX"},
{3, "BX"},
{4, "SP"},
{5, "BP"},
{6, "SI"},
{7, "DI"},
{8, "R8"},
{9, "R9"},
{10, "R10"},
{11, "R11"},
{12, "R12"},
{13, "R13"},
{14, "R14"},
{15, "R15"},
{16, "X0"},
{17, "X1"},
{18, "X2"},
{19, "X3"},
{20, "X4"},
{21, "X5"},
{22, "X6"},
{23, "X7"},
{24, "X8"},
{25, "X9"},
{26, "X10"},
{27, "X11"},
{28, "X12"},
{29, "X13"},
{30, "X14"},
{31, "X15"},
{32, "SB"},
{33, "FLAGS"},
}
var gpRegMaskAMD64 = regMask(65519)
var fpRegMaskAMD64 = regMask(4294901760)
var flagRegMaskAMD64 = regMask(8589934592)
var framepointerRegAMD64 = int8(5)
var registersARM = [...]Register{
{0, "R0"},
{1, "R1"},
{2, "R2"},
{3, "R3"},
{4, "R4"},
{5, "R5"},
{6, "R6"},
{7, "R7"},
{8, "R8"},
{9, "R9"},
{10, "g"},
{11, "R11"},
{12, "R12"},
{13, "SP"},
{14, "R14"},
{15, "R15"},
{16, "F0"},
{17, "F1"},
{18, "F2"},
{19, "F3"},
{20, "F4"},
{21, "F5"},
{22, "F6"},
{23, "F7"},
{24, "F8"},
{25, "F9"},
{26, "F10"},
{27, "F11"},
{28, "F12"},
{29, "F13"},
{30, "F14"},
{31, "F15"},
{32, "FLAGS"},
{33, "SB"},
}
var gpRegMaskARM = regMask(5119)
var fpRegMaskARM = regMask(4294901760)
var flagRegMaskARM = regMask(4294967296)
var framepointerRegARM = int8(-1)
var registersPPC64 = [...]Register{
{0, "R0"},
{1, "SP"},
{2, "SB"},
{3, "R3"},
{4, "R4"},
{5, "R5"},
{6, "R6"},
{7, "R7"},
{8, "R8"},
{9, "R9"},
{10, "R10"},
{11, "R11"},
{12, "R12"},
{13, "R13"},
{14, "R14"},
{15, "R15"},
{16, "R16"},
{17, "R17"},
{18, "R18"},
{19, "R19"},
{20, "R20"},
{21, "R21"},
{22, "R22"},
{23, "R23"},
{24, "R24"},
{25, "R25"},
{26, "R26"},
{27, "R27"},
{28, "R28"},
{29, "R29"},
{30, "R30"},
{31, "R31"},
{32, "F0"},
{33, "F1"},
{34, "F2"},
{35, "F3"},
{36, "F4"},
{37, "F5"},
{38, "F6"},
{39, "F7"},
{40, "F8"},
{41, "F9"},
{42, "F10"},
{43, "F11"},
{44, "F12"},
{45, "F13"},
{46, "F14"},
{47, "F15"},
{48, "F16"},
{49, "F17"},
{50, "F18"},
{51, "F19"},
{52, "F20"},
{53, "F21"},
{54, "F22"},
{55, "F23"},
{56, "F24"},
{57, "F25"},
{58, "F26"},
{59, "CR"},
}
var gpRegMaskPPC64 = regMask(1073731576)
var fpRegMaskPPC64 = regMask(576460743713488896)
var flagRegMaskPPC64 = regMask(0)
var framepointerRegPPC64 = int8(1)
Reference in a new issue Copy permalink