go/src/cmd/compile/internal/gc/align.go

457 lines
9.9 KiB
Go
Raw Normal View History

// Copyright 2009 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package gc
import (
"cmd/compile/internal/types"
"sort"
)
// sizeCalculationDisabled indicates whether it is safe
// to calculate Types' widths and alignments. See dowidth.
var sizeCalculationDisabled bool
// machine size and rounding alignment is dictated around
// the size of a pointer, set in betypeinit (see ../amd64/galign.go).
var defercalc int
func Rnd(o int64, r int64) int64 {
if r < 1 || r > 8 || r&(r-1) != 0 {
Fatalf("rnd %d", r)
}
return (o + r - 1) &^ (r - 1)
}
// expandiface computes the method set for interface type t by
// expanding embedded interfaces.
func expandiface(t *types.Type) {
var fields []*types.Field
for _, m := range t.Methods().Slice() {
if m.Sym != nil {
fields = append(fields, m)
checkwidth(m.Type)
continue
}
if !m.Type.IsInterface() {
yyerrorl(asNode(m.Nname).Pos, "interface contains embedded non-interface %v", m.Type)
m.SetBroke(true)
t.SetBroke(true)
// Add to fields so that error messages
// include the broken embedded type when
// printing t.
// TODO(mdempsky): Revisit this.
fields = append(fields, m)
continue
}
// Embedded interface: duplicate all methods
// (including broken ones, if any) and add to t's
// method set.
for _, t1 := range m.Type.Fields().Slice() {
f := types.NewField()
f.Type = t1.Type
f.SetBroke(t1.Broke())
f.Sym = t1.Sym
f.Nname = m.Nname // preserve embedding position
fields = append(fields, f)
}
}
sort.Sort(methcmp(fields))
// Access fields directly to avoid recursively calling dowidth
// within Type.Fields().
t.Extra.(*types.Interface).Fields.Set(fields)
}
func offmod(t *types.Type) {
o := int32(0)
for _, f := range t.Fields().Slice() {
f.Offset = int64(o)
o += int32(Widthptr)
if int64(o) >= thearch.MAXWIDTH {
yyerror("interface too large")
o = int32(Widthptr)
}
}
}
func widstruct(errtype *types.Type, t *types.Type, o int64, flag int) int64 {
starto := o
maxalign := int32(flag)
if maxalign < 1 {
maxalign = 1
}
lastzero := int64(0)
for _, f := range t.Fields().Slice() {
if f.Type == nil {
// broken field, just skip it so that other valid fields
// get a width.
continue
}
dowidth(f.Type)
if int32(f.Type.Align) > maxalign {
maxalign = int32(f.Type.Align)
}
if f.Type.Align > 0 {
o = Rnd(o, int64(f.Type.Align))
}
f.Offset = o
if asNode(f.Nname) != nil {
cmd/compile: fix liveness computation for heap-escaped parameters The liveness computation of parameters generally was never correct, but forcing all parameters to be live throughout the function covered up that problem. The new SSA back end is too clever: even though it currently keeps the parameter values live throughout the function, it may find optimizations that mean the current values are not written back to the original parameter stack slots immediately or ever (for example if a parameter is set to nil, SSA constant propagation may replace all later uses of the parameter with a constant nil, eliminating the need to write the nil value back to the stack slot), so the liveness code must now track the actual operations on the stack slots, exposing these problems. One small problem in the handling of arguments is that nodarg can return ONAME PPARAM nodes with adjusted offsets, so that there are actually multiple *Node pointers for the same parameter in the instruction stream. This might be possible to correct, but not in this CL. For now, we fix this by using n.Orig instead of n when considering PPARAM and PPARAMOUT nodes. The major problem in the handling of arguments is general confusion in the liveness code about the meaning of PPARAM|PHEAP and PPARAMOUT|PHEAP nodes, especially as contrasted with PAUTO|PHEAP. The difference between these two is that when a local variable "moves" to the heap, it's really just allocated there to start with; in contrast, when an argument moves to the heap, the actual data has to be copied there from the stack at the beginning of the function, and when a result "moves" to the heap the value in the heap has to be copied back to the stack when the function returns This general confusion is also present in the SSA back end. The PHEAP bit worked decently when I first introduced it 7 years ago (!) in 391425ae. The back end did nothing sophisticated, and in particular there was no analysis at all: no escape analysis, no liveness analysis, and certainly no SSA back end. But the complications caused in the various downstream consumers suggest that this should be a detail kept mainly in the front end. This CL therefore eliminates both the PHEAP bit and even the idea of "heap variables" from the back ends. First, it replaces the PPARAM|PHEAP, PPARAMOUT|PHEAP, and PAUTO|PHEAP variable classes with the single PAUTOHEAP, a pseudo-class indicating a variable maintained on the heap and available by indirecting a local variable kept on the stack (a plain PAUTO). Second, walkexpr replaces all references to PAUTOHEAP variables with indirections of the corresponding PAUTO variable. The back ends and the liveness code now just see plain indirected variables. This may actually produce better code, but the real goal here is to eliminate these little-used and somewhat suspect code paths in the back end analyses. The OPARAM node type goes away too. A followup CL will do the same to PPARAMREF. I'm not sure that the back ends (SSA in particular) are handling those right either, and with the framework established in this CL that change is trivial and the result clearly more correct. Fixes #15747. Change-Id: I2770b1ce3cbc93981bfc7166be66a9da12013d74 Reviewed-on: https://go-review.googlesource.com/23393 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Russ Cox <rsc@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-05-25 01:33:24 -04:00
// addrescapes has similar code to update these offsets.
// Usually addrescapes runs after widstruct,
// in which case we could drop this,
// but function closure functions are the exception.
cmd/compile: fix liveness computation for heap-escaped parameters The liveness computation of parameters generally was never correct, but forcing all parameters to be live throughout the function covered up that problem. The new SSA back end is too clever: even though it currently keeps the parameter values live throughout the function, it may find optimizations that mean the current values are not written back to the original parameter stack slots immediately or ever (for example if a parameter is set to nil, SSA constant propagation may replace all later uses of the parameter with a constant nil, eliminating the need to write the nil value back to the stack slot), so the liveness code must now track the actual operations on the stack slots, exposing these problems. One small problem in the handling of arguments is that nodarg can return ONAME PPARAM nodes with adjusted offsets, so that there are actually multiple *Node pointers for the same parameter in the instruction stream. This might be possible to correct, but not in this CL. For now, we fix this by using n.Orig instead of n when considering PPARAM and PPARAMOUT nodes. The major problem in the handling of arguments is general confusion in the liveness code about the meaning of PPARAM|PHEAP and PPARAMOUT|PHEAP nodes, especially as contrasted with PAUTO|PHEAP. The difference between these two is that when a local variable "moves" to the heap, it's really just allocated there to start with; in contrast, when an argument moves to the heap, the actual data has to be copied there from the stack at the beginning of the function, and when a result "moves" to the heap the value in the heap has to be copied back to the stack when the function returns This general confusion is also present in the SSA back end. The PHEAP bit worked decently when I first introduced it 7 years ago (!) in 391425ae. The back end did nothing sophisticated, and in particular there was no analysis at all: no escape analysis, no liveness analysis, and certainly no SSA back end. But the complications caused in the various downstream consumers suggest that this should be a detail kept mainly in the front end. This CL therefore eliminates both the PHEAP bit and even the idea of "heap variables" from the back ends. First, it replaces the PPARAM|PHEAP, PPARAMOUT|PHEAP, and PAUTO|PHEAP variable classes with the single PAUTOHEAP, a pseudo-class indicating a variable maintained on the heap and available by indirecting a local variable kept on the stack (a plain PAUTO). Second, walkexpr replaces all references to PAUTOHEAP variables with indirections of the corresponding PAUTO variable. The back ends and the liveness code now just see plain indirected variables. This may actually produce better code, but the real goal here is to eliminate these little-used and somewhat suspect code paths in the back end analyses. The OPARAM node type goes away too. A followup CL will do the same to PPARAMREF. I'm not sure that the back ends (SSA in particular) are handling those right either, and with the framework established in this CL that change is trivial and the result clearly more correct. Fixes #15747. Change-Id: I2770b1ce3cbc93981bfc7166be66a9da12013d74 Reviewed-on: https://go-review.googlesource.com/23393 Reviewed-by: Keith Randall <khr@golang.org> Run-TryBot: Russ Cox <rsc@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-05-25 01:33:24 -04:00
// NOTE(rsc): This comment may be stale.
// It's possible the ordering has changed and this is
// now the common case. I'm not sure.
if asNode(f.Nname).Name.Param.Stackcopy != nil {
asNode(f.Nname).Name.Param.Stackcopy.Xoffset = o
asNode(f.Nname).Xoffset = 0
} else {
asNode(f.Nname).Xoffset = o
}
}
w := f.Type.Width
if w < 0 {
Fatalf("invalid width %d", f.Type.Width)
}
if w == 0 {
lastzero = o
}
o += w
maxwidth := thearch.MAXWIDTH
// On 32-bit systems, reflect tables impose an additional constraint
// that each field start offset must fit in 31 bits.
if maxwidth < 1<<32 {
maxwidth = 1<<31 - 1
}
if o >= maxwidth {
yyerror("type %L too large", errtype)
o = 8 // small but nonzero
}
}
// For nonzero-sized structs which end in a zero-sized thing, we add
// an extra byte of padding to the type. This padding ensures that
// taking the address of the zero-sized thing can't manufacture a
// pointer to the next object in the heap. See issue 9401.
if flag == 1 && o > starto && o == lastzero {
o++
}
// final width is rounded
if flag != 0 {
o = Rnd(o, int64(maxalign))
}
t.Align = uint8(maxalign)
// type width only includes back to first field's offset
t.Width = o - starto
return o
}
// dowidth calculates and stores the size and alignment for t.
// If sizeCalculationDisabled is set, and the size/alignment
// have not already been calculated, it calls Fatal.
// This is used to prevent data races in the back end.
func dowidth(t *types.Type) {
if Widthptr == 0 {
Fatalf("dowidth without betypeinit")
}
if t == nil {
return
}
if t.Width == -2 {
cmd/compile: pack bool fields in Node, Name, Func and Type structs to bitsets This reduces compiler memory usage by up to 4% - see compilebench results below. name old time/op new time/op delta Template 245ms ± 4% 241ms ± 2% -1.88% (p=0.029 n=10+10) Unicode 126ms ± 3% 124ms ± 3% ~ (p=0.105 n=10+10) GoTypes 805ms ± 2% 813ms ± 3% ~ (p=0.515 n=8+10) Compiler 3.95s ± 2% 3.83s ± 1% -2.96% (p=0.000 n=9+10) MakeBash 47.4s ± 4% 46.6s ± 1% -1.59% (p=0.028 n=9+10) name old user-ns/op new user-ns/op delta Template 324M ± 5% 326M ± 3% ~ (p=0.935 n=10+10) Unicode 186M ± 5% 178M ±10% ~ (p=0.067 n=9+10) GoTypes 1.08G ± 7% 1.09G ± 4% ~ (p=0.956 n=10+10) Compiler 5.34G ± 4% 5.31G ± 1% ~ (p=0.501 n=10+8) name old alloc/op new alloc/op delta Template 41.0MB ± 0% 39.8MB ± 0% -3.03% (p=0.000 n=10+10) Unicode 32.3MB ± 0% 31.0MB ± 0% -4.13% (p=0.000 n=10+10) GoTypes 119MB ± 0% 116MB ± 0% -2.39% (p=0.000 n=10+10) Compiler 499MB ± 0% 487MB ± 0% -2.48% (p=0.000 n=10+10) name old allocs/op new allocs/op delta Template 380k ± 1% 379k ± 1% ~ (p=0.436 n=10+10) Unicode 324k ± 1% 324k ± 0% ~ (p=0.853 n=10+10) GoTypes 1.15M ± 0% 1.15M ± 0% ~ (p=0.481 n=10+10) Compiler 4.41M ± 0% 4.41M ± 0% -0.12% (p=0.007 n=10+10) name old text-bytes new text-bytes delta HelloSize 623k ± 0% 623k ± 0% ~ (all equal) CmdGoSize 6.64M ± 0% 6.64M ± 0% ~ (all equal) name old data-bytes new data-bytes delta HelloSize 5.81k ± 0% 5.81k ± 0% ~ (all equal) CmdGoSize 238k ± 0% 238k ± 0% ~ (all equal) name old bss-bytes new bss-bytes delta HelloSize 134k ± 0% 134k ± 0% ~ (all equal) CmdGoSize 152k ± 0% 152k ± 0% ~ (all equal) name old exe-bytes new exe-bytes delta HelloSize 967k ± 0% 967k ± 0% ~ (all equal) CmdGoSize 10.2M ± 0% 10.2M ± 0% ~ (all equal) Change-Id: I1f40af738254892bd6c8ba2eb43390b175753d52 Reviewed-on: https://go-review.googlesource.com/37445 Reviewed-by: Matthew Dempsky <mdempsky@google.com> Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2017-02-27 19:56:38 +02:00
if !t.Broke() {
t.SetBroke(true)
yyerrorl(asNode(t.Nod).Pos, "invalid recursive type %v", t)
}
t.Width = 0
return
}
if t.WidthCalculated() {
return
}
if sizeCalculationDisabled {
if t.Broke() {
// break infinite recursion from Fatal call below
return
}
t.SetBroke(true)
Fatalf("width not calculated: %v", t)
}
// break infinite recursion if the broken recursive type
// is referenced again
cmd/compile: pack bool fields in Node, Name, Func and Type structs to bitsets This reduces compiler memory usage by up to 4% - see compilebench results below. name old time/op new time/op delta Template 245ms ± 4% 241ms ± 2% -1.88% (p=0.029 n=10+10) Unicode 126ms ± 3% 124ms ± 3% ~ (p=0.105 n=10+10) GoTypes 805ms ± 2% 813ms ± 3% ~ (p=0.515 n=8+10) Compiler 3.95s ± 2% 3.83s ± 1% -2.96% (p=0.000 n=9+10) MakeBash 47.4s ± 4% 46.6s ± 1% -1.59% (p=0.028 n=9+10) name old user-ns/op new user-ns/op delta Template 324M ± 5% 326M ± 3% ~ (p=0.935 n=10+10) Unicode 186M ± 5% 178M ±10% ~ (p=0.067 n=9+10) GoTypes 1.08G ± 7% 1.09G ± 4% ~ (p=0.956 n=10+10) Compiler 5.34G ± 4% 5.31G ± 1% ~ (p=0.501 n=10+8) name old alloc/op new alloc/op delta Template 41.0MB ± 0% 39.8MB ± 0% -3.03% (p=0.000 n=10+10) Unicode 32.3MB ± 0% 31.0MB ± 0% -4.13% (p=0.000 n=10+10) GoTypes 119MB ± 0% 116MB ± 0% -2.39% (p=0.000 n=10+10) Compiler 499MB ± 0% 487MB ± 0% -2.48% (p=0.000 n=10+10) name old allocs/op new allocs/op delta Template 380k ± 1% 379k ± 1% ~ (p=0.436 n=10+10) Unicode 324k ± 1% 324k ± 0% ~ (p=0.853 n=10+10) GoTypes 1.15M ± 0% 1.15M ± 0% ~ (p=0.481 n=10+10) Compiler 4.41M ± 0% 4.41M ± 0% -0.12% (p=0.007 n=10+10) name old text-bytes new text-bytes delta HelloSize 623k ± 0% 623k ± 0% ~ (all equal) CmdGoSize 6.64M ± 0% 6.64M ± 0% ~ (all equal) name old data-bytes new data-bytes delta HelloSize 5.81k ± 0% 5.81k ± 0% ~ (all equal) CmdGoSize 238k ± 0% 238k ± 0% ~ (all equal) name old bss-bytes new bss-bytes delta HelloSize 134k ± 0% 134k ± 0% ~ (all equal) CmdGoSize 152k ± 0% 152k ± 0% ~ (all equal) name old exe-bytes new exe-bytes delta HelloSize 967k ± 0% 967k ± 0% ~ (all equal) CmdGoSize 10.2M ± 0% 10.2M ± 0% ~ (all equal) Change-Id: I1f40af738254892bd6c8ba2eb43390b175753d52 Reviewed-on: https://go-review.googlesource.com/37445 Reviewed-by: Matthew Dempsky <mdempsky@google.com> Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2017-02-27 19:56:38 +02:00
if t.Broke() && t.Width == 0 {
return
}
// defer checkwidth calls until after we're done
defercalc++
lno := lineno
if asNode(t.Nod) != nil {
lineno = asNode(t.Nod).Pos
}
t.Width = -2
t.Align = 0
et := t.Etype
switch et {
case TFUNC, TCHAN, TMAP, TSTRING:
break
// simtype == 0 during bootstrap
default:
if simtype[t.Etype] != 0 {
et = simtype[t.Etype]
}
}
w := int64(0)
switch et {
default:
Fatalf("dowidth: unknown type: %v", t)
// compiler-specific stuff
case TINT8, TUINT8, TBOOL:
// bool is int8
w = 1
case TINT16, TUINT16:
w = 2
case TINT32, TUINT32, TFLOAT32:
w = 4
case TINT64, TUINT64, TFLOAT64:
w = 8
t.Align = uint8(Widthreg)
case TCOMPLEX64:
w = 8
t.Align = 4
case TCOMPLEX128:
w = 16
t.Align = uint8(Widthreg)
case TPTR32:
w = 4
checkwidth(t.Elem())
case TPTR64:
w = 8
checkwidth(t.Elem())
case TUNSAFEPTR:
w = int64(Widthptr)
case TINTER: // implemented as 2 pointers
w = 2 * int64(Widthptr)
t.Align = uint8(Widthptr)
expandiface(t)
case TCHAN: // implemented as pointer
w = int64(Widthptr)
checkwidth(t.Elem())
// make fake type to check later to
// trigger channel argument check.
t1 := types.NewChanArgs(t)
checkwidth(t1)
case TCHANARGS:
cmd/compile: shrink gc.Type in half Many of Type's fields are etype-specific. This CL organizes them into their own auxiliary types, duplicating a few fields as necessary, and adds an Extra field to hold them. It also sorts the remaining fields for better struct packing. It also improves documentation for most fields. This reduces the size of Type at the cost of some extra allocations. There's no CPU impact; memory impact below. It also makes the natural structure of Type clearer. Passes toolstash -cmp on all architectures. Ideas for future work in this vein: (1) Width and Align probably only need to be stored for Struct and Array types. The refactoring to accomplish this would hopefully also eliminate TFUNCARGS and TCHANARGS entirely. (2) Maplineno is sparsely used and could probably better be stored in a separate map[*Type]int32, with mapqueue updated to store both a Node and a line number. (3) The Printed field may be removable once the old (non-binary) importer/exported has been removed. (4) StructType's fields field could be changed from *[]*Field to []*Field, which would remove a common allocation. (5) I believe that Type.Nod can be moved to ForwardType. Separate CL. name old alloc/op new alloc/op delta Template 57.9MB ± 0% 55.9MB ± 0% -3.43% (p=0.000 n=50+50) Unicode 38.3MB ± 0% 37.8MB ± 0% -1.39% (p=0.000 n=50+50) GoTypes 185MB ± 0% 180MB ± 0% -2.56% (p=0.000 n=50+50) Compiler 824MB ± 0% 806MB ± 0% -2.19% (p=0.000 n=50+50) name old allocs/op new allocs/op delta Template 486k ± 0% 497k ± 0% +2.25% (p=0.000 n=50+50) Unicode 377k ± 0% 379k ± 0% +0.55% (p=0.000 n=50+50) GoTypes 1.39M ± 0% 1.42M ± 0% +1.63% (p=0.000 n=50+50) Compiler 5.52M ± 0% 5.57M ± 0% +0.84% (p=0.000 n=47+50) Change-Id: I828488eeb74902b013d5ae4cf844de0b6c0dfc87 Reviewed-on: https://go-review.googlesource.com/21611 Reviewed-by: Matthew Dempsky <mdempsky@google.com> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
2016-04-01 20:11:30 -07:00
t1 := t.ChanArgs()
dowidth(t1) // just in case
if t1.Elem().Width >= 1<<16 {
yyerror("channel element type too large (>64kB)")
}
w = 1 // anything will do
case TMAP: // implemented as pointer
w = int64(Widthptr)
checkwidth(t.Val())
checkwidth(t.Key())
case TFORW: // should have been filled in
cmd/compile: pack bool fields in Node, Name, Func and Type structs to bitsets This reduces compiler memory usage by up to 4% - see compilebench results below. name old time/op new time/op delta Template 245ms ± 4% 241ms ± 2% -1.88% (p=0.029 n=10+10) Unicode 126ms ± 3% 124ms ± 3% ~ (p=0.105 n=10+10) GoTypes 805ms ± 2% 813ms ± 3% ~ (p=0.515 n=8+10) Compiler 3.95s ± 2% 3.83s ± 1% -2.96% (p=0.000 n=9+10) MakeBash 47.4s ± 4% 46.6s ± 1% -1.59% (p=0.028 n=9+10) name old user-ns/op new user-ns/op delta Template 324M ± 5% 326M ± 3% ~ (p=0.935 n=10+10) Unicode 186M ± 5% 178M ±10% ~ (p=0.067 n=9+10) GoTypes 1.08G ± 7% 1.09G ± 4% ~ (p=0.956 n=10+10) Compiler 5.34G ± 4% 5.31G ± 1% ~ (p=0.501 n=10+8) name old alloc/op new alloc/op delta Template 41.0MB ± 0% 39.8MB ± 0% -3.03% (p=0.000 n=10+10) Unicode 32.3MB ± 0% 31.0MB ± 0% -4.13% (p=0.000 n=10+10) GoTypes 119MB ± 0% 116MB ± 0% -2.39% (p=0.000 n=10+10) Compiler 499MB ± 0% 487MB ± 0% -2.48% (p=0.000 n=10+10) name old allocs/op new allocs/op delta Template 380k ± 1% 379k ± 1% ~ (p=0.436 n=10+10) Unicode 324k ± 1% 324k ± 0% ~ (p=0.853 n=10+10) GoTypes 1.15M ± 0% 1.15M ± 0% ~ (p=0.481 n=10+10) Compiler 4.41M ± 0% 4.41M ± 0% -0.12% (p=0.007 n=10+10) name old text-bytes new text-bytes delta HelloSize 623k ± 0% 623k ± 0% ~ (all equal) CmdGoSize 6.64M ± 0% 6.64M ± 0% ~ (all equal) name old data-bytes new data-bytes delta HelloSize 5.81k ± 0% 5.81k ± 0% ~ (all equal) CmdGoSize 238k ± 0% 238k ± 0% ~ (all equal) name old bss-bytes new bss-bytes delta HelloSize 134k ± 0% 134k ± 0% ~ (all equal) CmdGoSize 152k ± 0% 152k ± 0% ~ (all equal) name old exe-bytes new exe-bytes delta HelloSize 967k ± 0% 967k ± 0% ~ (all equal) CmdGoSize 10.2M ± 0% 10.2M ± 0% ~ (all equal) Change-Id: I1f40af738254892bd6c8ba2eb43390b175753d52 Reviewed-on: https://go-review.googlesource.com/37445 Reviewed-by: Matthew Dempsky <mdempsky@google.com> Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2017-02-27 19:56:38 +02:00
if !t.Broke() {
yyerror("invalid recursive type %v", t)
}
w = 1 // anything will do
case TANY:
// dummy type; should be replaced before use.
Fatalf("dowidth any")
case TSTRING:
if sizeof_String == 0 {
Fatalf("early dowidth string")
}
w = int64(sizeof_String)
t.Align = uint8(Widthptr)
case TARRAY:
if t.Elem() == nil {
break
}
if t.IsDDDArray() {
cmd/compile: pack bool fields in Node, Name, Func and Type structs to bitsets This reduces compiler memory usage by up to 4% - see compilebench results below. name old time/op new time/op delta Template 245ms ± 4% 241ms ± 2% -1.88% (p=0.029 n=10+10) Unicode 126ms ± 3% 124ms ± 3% ~ (p=0.105 n=10+10) GoTypes 805ms ± 2% 813ms ± 3% ~ (p=0.515 n=8+10) Compiler 3.95s ± 2% 3.83s ± 1% -2.96% (p=0.000 n=9+10) MakeBash 47.4s ± 4% 46.6s ± 1% -1.59% (p=0.028 n=9+10) name old user-ns/op new user-ns/op delta Template 324M ± 5% 326M ± 3% ~ (p=0.935 n=10+10) Unicode 186M ± 5% 178M ±10% ~ (p=0.067 n=9+10) GoTypes 1.08G ± 7% 1.09G ± 4% ~ (p=0.956 n=10+10) Compiler 5.34G ± 4% 5.31G ± 1% ~ (p=0.501 n=10+8) name old alloc/op new alloc/op delta Template 41.0MB ± 0% 39.8MB ± 0% -3.03% (p=0.000 n=10+10) Unicode 32.3MB ± 0% 31.0MB ± 0% -4.13% (p=0.000 n=10+10) GoTypes 119MB ± 0% 116MB ± 0% -2.39% (p=0.000 n=10+10) Compiler 499MB ± 0% 487MB ± 0% -2.48% (p=0.000 n=10+10) name old allocs/op new allocs/op delta Template 380k ± 1% 379k ± 1% ~ (p=0.436 n=10+10) Unicode 324k ± 1% 324k ± 0% ~ (p=0.853 n=10+10) GoTypes 1.15M ± 0% 1.15M ± 0% ~ (p=0.481 n=10+10) Compiler 4.41M ± 0% 4.41M ± 0% -0.12% (p=0.007 n=10+10) name old text-bytes new text-bytes delta HelloSize 623k ± 0% 623k ± 0% ~ (all equal) CmdGoSize 6.64M ± 0% 6.64M ± 0% ~ (all equal) name old data-bytes new data-bytes delta HelloSize 5.81k ± 0% 5.81k ± 0% ~ (all equal) CmdGoSize 238k ± 0% 238k ± 0% ~ (all equal) name old bss-bytes new bss-bytes delta HelloSize 134k ± 0% 134k ± 0% ~ (all equal) CmdGoSize 152k ± 0% 152k ± 0% ~ (all equal) name old exe-bytes new exe-bytes delta HelloSize 967k ± 0% 967k ± 0% ~ (all equal) CmdGoSize 10.2M ± 0% 10.2M ± 0% ~ (all equal) Change-Id: I1f40af738254892bd6c8ba2eb43390b175753d52 Reviewed-on: https://go-review.googlesource.com/37445 Reviewed-by: Matthew Dempsky <mdempsky@google.com> Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2017-02-27 19:56:38 +02:00
if !t.Broke() {
yyerror("use of [...] array outside of array literal")
cmd/compile: pack bool fields in Node, Name, Func and Type structs to bitsets This reduces compiler memory usage by up to 4% - see compilebench results below. name old time/op new time/op delta Template 245ms ± 4% 241ms ± 2% -1.88% (p=0.029 n=10+10) Unicode 126ms ± 3% 124ms ± 3% ~ (p=0.105 n=10+10) GoTypes 805ms ± 2% 813ms ± 3% ~ (p=0.515 n=8+10) Compiler 3.95s ± 2% 3.83s ± 1% -2.96% (p=0.000 n=9+10) MakeBash 47.4s ± 4% 46.6s ± 1% -1.59% (p=0.028 n=9+10) name old user-ns/op new user-ns/op delta Template 324M ± 5% 326M ± 3% ~ (p=0.935 n=10+10) Unicode 186M ± 5% 178M ±10% ~ (p=0.067 n=9+10) GoTypes 1.08G ± 7% 1.09G ± 4% ~ (p=0.956 n=10+10) Compiler 5.34G ± 4% 5.31G ± 1% ~ (p=0.501 n=10+8) name old alloc/op new alloc/op delta Template 41.0MB ± 0% 39.8MB ± 0% -3.03% (p=0.000 n=10+10) Unicode 32.3MB ± 0% 31.0MB ± 0% -4.13% (p=0.000 n=10+10) GoTypes 119MB ± 0% 116MB ± 0% -2.39% (p=0.000 n=10+10) Compiler 499MB ± 0% 487MB ± 0% -2.48% (p=0.000 n=10+10) name old allocs/op new allocs/op delta Template 380k ± 1% 379k ± 1% ~ (p=0.436 n=10+10) Unicode 324k ± 1% 324k ± 0% ~ (p=0.853 n=10+10) GoTypes 1.15M ± 0% 1.15M ± 0% ~ (p=0.481 n=10+10) Compiler 4.41M ± 0% 4.41M ± 0% -0.12% (p=0.007 n=10+10) name old text-bytes new text-bytes delta HelloSize 623k ± 0% 623k ± 0% ~ (all equal) CmdGoSize 6.64M ± 0% 6.64M ± 0% ~ (all equal) name old data-bytes new data-bytes delta HelloSize 5.81k ± 0% 5.81k ± 0% ~ (all equal) CmdGoSize 238k ± 0% 238k ± 0% ~ (all equal) name old bss-bytes new bss-bytes delta HelloSize 134k ± 0% 134k ± 0% ~ (all equal) CmdGoSize 152k ± 0% 152k ± 0% ~ (all equal) name old exe-bytes new exe-bytes delta HelloSize 967k ± 0% 967k ± 0% ~ (all equal) CmdGoSize 10.2M ± 0% 10.2M ± 0% ~ (all equal) Change-Id: I1f40af738254892bd6c8ba2eb43390b175753d52 Reviewed-on: https://go-review.googlesource.com/37445 Reviewed-by: Matthew Dempsky <mdempsky@google.com> Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2017-02-27 19:56:38 +02:00
t.SetBroke(true)
}
break
}
dowidth(t.Elem())
if t.Elem().Width != 0 {
cap := (uint64(thearch.MAXWIDTH) - 1) / uint64(t.Elem().Width)
if uint64(t.NumElem()) > cap {
yyerror("type %L larger than address space", t)
}
}
w = t.NumElem() * t.Elem().Width
t.Align = t.Elem().Align
case TSLICE:
if t.Elem() == nil {
break
}
w = int64(sizeof_Array)
checkwidth(t.Elem())
t.Align = uint8(Widthptr)
case TSTRUCT:
if t.IsFuncArgStruct() {
Fatalf("dowidth fn struct %v", t)
}
w = widstruct(t, t, 0, 1)
// make fake type to check later to
// trigger function argument computation.
case TFUNC:
t1 := types.NewFuncArgs(t)
checkwidth(t1)
w = int64(Widthptr) // width of func type is pointer
// function is 3 cated structures;
// compute their widths as side-effect.
case TFUNCARGS:
cmd/compile: shrink gc.Type in half Many of Type's fields are etype-specific. This CL organizes them into their own auxiliary types, duplicating a few fields as necessary, and adds an Extra field to hold them. It also sorts the remaining fields for better struct packing. It also improves documentation for most fields. This reduces the size of Type at the cost of some extra allocations. There's no CPU impact; memory impact below. It also makes the natural structure of Type clearer. Passes toolstash -cmp on all architectures. Ideas for future work in this vein: (1) Width and Align probably only need to be stored for Struct and Array types. The refactoring to accomplish this would hopefully also eliminate TFUNCARGS and TCHANARGS entirely. (2) Maplineno is sparsely used and could probably better be stored in a separate map[*Type]int32, with mapqueue updated to store both a Node and a line number. (3) The Printed field may be removable once the old (non-binary) importer/exported has been removed. (4) StructType's fields field could be changed from *[]*Field to []*Field, which would remove a common allocation. (5) I believe that Type.Nod can be moved to ForwardType. Separate CL. name old alloc/op new alloc/op delta Template 57.9MB ± 0% 55.9MB ± 0% -3.43% (p=0.000 n=50+50) Unicode 38.3MB ± 0% 37.8MB ± 0% -1.39% (p=0.000 n=50+50) GoTypes 185MB ± 0% 180MB ± 0% -2.56% (p=0.000 n=50+50) Compiler 824MB ± 0% 806MB ± 0% -2.19% (p=0.000 n=50+50) name old allocs/op new allocs/op delta Template 486k ± 0% 497k ± 0% +2.25% (p=0.000 n=50+50) Unicode 377k ± 0% 379k ± 0% +0.55% (p=0.000 n=50+50) GoTypes 1.39M ± 0% 1.42M ± 0% +1.63% (p=0.000 n=50+50) Compiler 5.52M ± 0% 5.57M ± 0% +0.84% (p=0.000 n=47+50) Change-Id: I828488eeb74902b013d5ae4cf844de0b6c0dfc87 Reviewed-on: https://go-review.googlesource.com/21611 Reviewed-by: Matthew Dempsky <mdempsky@google.com> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
2016-04-01 20:11:30 -07:00
t1 := t.FuncArgs()
w = widstruct(t1, t1.Recvs(), 0, 0)
w = widstruct(t1, t1.Params(), w, Widthreg)
w = widstruct(t1, t1.Results(), w, Widthreg)
t1.Extra.(*types.Func).Argwid = w
if w%int64(Widthreg) != 0 {
Warn("bad type %v %d\n", t1, w)
}
t.Align = 1
}
if Widthptr == 4 && w != int64(int32(w)) {
yyerror("type %v too large", t)
}
t.Width = w
if t.Align == 0 {
if w > 8 || w&(w-1) != 0 || w == 0 {
Fatalf("invalid alignment for %v", t)
}
t.Align = uint8(w)
}
if t.Etype == TINTER {
// We defer calling these functions until after
// setting t.Width and t.Align so the recursive calls
// to dowidth within t.Fields() will succeed.
checkdupfields("method", t)
offmod(t)
}
lineno = lno
if defercalc == 1 {
resumecheckwidth()
} else {
defercalc--
}
}
// when a type's width should be known, we call checkwidth
// to compute it. during a declaration like
//
// type T *struct { next T }
//
// it is necessary to defer the calculation of the struct width
// until after T has been initialized to be a pointer to that struct.
// similarly, during import processing structs may be used
// before their definition. in those situations, calling
// defercheckwidth() stops width calculations until
// resumecheckwidth() is called, at which point all the
// checkwidths that were deferred are executed.
// dowidth should only be called when the type's size
// is needed immediately. checkwidth makes sure the
// size is evaluated eventually.
var deferredTypeStack []*types.Type
func checkwidth(t *types.Type) {
if t == nil {
return
}
// function arg structs should not be checked
// outside of the enclosing function.
if t.IsFuncArgStruct() {
Fatalf("checkwidth %v", t)
}
if defercalc == 0 {
dowidth(t)
return
}
cmd/compile: pack bool fields in Node, Name, Func and Type structs to bitsets This reduces compiler memory usage by up to 4% - see compilebench results below. name old time/op new time/op delta Template 245ms ± 4% 241ms ± 2% -1.88% (p=0.029 n=10+10) Unicode 126ms ± 3% 124ms ± 3% ~ (p=0.105 n=10+10) GoTypes 805ms ± 2% 813ms ± 3% ~ (p=0.515 n=8+10) Compiler 3.95s ± 2% 3.83s ± 1% -2.96% (p=0.000 n=9+10) MakeBash 47.4s ± 4% 46.6s ± 1% -1.59% (p=0.028 n=9+10) name old user-ns/op new user-ns/op delta Template 324M ± 5% 326M ± 3% ~ (p=0.935 n=10+10) Unicode 186M ± 5% 178M ±10% ~ (p=0.067 n=9+10) GoTypes 1.08G ± 7% 1.09G ± 4% ~ (p=0.956 n=10+10) Compiler 5.34G ± 4% 5.31G ± 1% ~ (p=0.501 n=10+8) name old alloc/op new alloc/op delta Template 41.0MB ± 0% 39.8MB ± 0% -3.03% (p=0.000 n=10+10) Unicode 32.3MB ± 0% 31.0MB ± 0% -4.13% (p=0.000 n=10+10) GoTypes 119MB ± 0% 116MB ± 0% -2.39% (p=0.000 n=10+10) Compiler 499MB ± 0% 487MB ± 0% -2.48% (p=0.000 n=10+10) name old allocs/op new allocs/op delta Template 380k ± 1% 379k ± 1% ~ (p=0.436 n=10+10) Unicode 324k ± 1% 324k ± 0% ~ (p=0.853 n=10+10) GoTypes 1.15M ± 0% 1.15M ± 0% ~ (p=0.481 n=10+10) Compiler 4.41M ± 0% 4.41M ± 0% -0.12% (p=0.007 n=10+10) name old text-bytes new text-bytes delta HelloSize 623k ± 0% 623k ± 0% ~ (all equal) CmdGoSize 6.64M ± 0% 6.64M ± 0% ~ (all equal) name old data-bytes new data-bytes delta HelloSize 5.81k ± 0% 5.81k ± 0% ~ (all equal) CmdGoSize 238k ± 0% 238k ± 0% ~ (all equal) name old bss-bytes new bss-bytes delta HelloSize 134k ± 0% 134k ± 0% ~ (all equal) CmdGoSize 152k ± 0% 152k ± 0% ~ (all equal) name old exe-bytes new exe-bytes delta HelloSize 967k ± 0% 967k ± 0% ~ (all equal) CmdGoSize 10.2M ± 0% 10.2M ± 0% ~ (all equal) Change-Id: I1f40af738254892bd6c8ba2eb43390b175753d52 Reviewed-on: https://go-review.googlesource.com/37445 Reviewed-by: Matthew Dempsky <mdempsky@google.com> Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2017-02-27 19:56:38 +02:00
if t.Deferwidth() {
return
}
cmd/compile: pack bool fields in Node, Name, Func and Type structs to bitsets This reduces compiler memory usage by up to 4% - see compilebench results below. name old time/op new time/op delta Template 245ms ± 4% 241ms ± 2% -1.88% (p=0.029 n=10+10) Unicode 126ms ± 3% 124ms ± 3% ~ (p=0.105 n=10+10) GoTypes 805ms ± 2% 813ms ± 3% ~ (p=0.515 n=8+10) Compiler 3.95s ± 2% 3.83s ± 1% -2.96% (p=0.000 n=9+10) MakeBash 47.4s ± 4% 46.6s ± 1% -1.59% (p=0.028 n=9+10) name old user-ns/op new user-ns/op delta Template 324M ± 5% 326M ± 3% ~ (p=0.935 n=10+10) Unicode 186M ± 5% 178M ±10% ~ (p=0.067 n=9+10) GoTypes 1.08G ± 7% 1.09G ± 4% ~ (p=0.956 n=10+10) Compiler 5.34G ± 4% 5.31G ± 1% ~ (p=0.501 n=10+8) name old alloc/op new alloc/op delta Template 41.0MB ± 0% 39.8MB ± 0% -3.03% (p=0.000 n=10+10) Unicode 32.3MB ± 0% 31.0MB ± 0% -4.13% (p=0.000 n=10+10) GoTypes 119MB ± 0% 116MB ± 0% -2.39% (p=0.000 n=10+10) Compiler 499MB ± 0% 487MB ± 0% -2.48% (p=0.000 n=10+10) name old allocs/op new allocs/op delta Template 380k ± 1% 379k ± 1% ~ (p=0.436 n=10+10) Unicode 324k ± 1% 324k ± 0% ~ (p=0.853 n=10+10) GoTypes 1.15M ± 0% 1.15M ± 0% ~ (p=0.481 n=10+10) Compiler 4.41M ± 0% 4.41M ± 0% -0.12% (p=0.007 n=10+10) name old text-bytes new text-bytes delta HelloSize 623k ± 0% 623k ± 0% ~ (all equal) CmdGoSize 6.64M ± 0% 6.64M ± 0% ~ (all equal) name old data-bytes new data-bytes delta HelloSize 5.81k ± 0% 5.81k ± 0% ~ (all equal) CmdGoSize 238k ± 0% 238k ± 0% ~ (all equal) name old bss-bytes new bss-bytes delta HelloSize 134k ± 0% 134k ± 0% ~ (all equal) CmdGoSize 152k ± 0% 152k ± 0% ~ (all equal) name old exe-bytes new exe-bytes delta HelloSize 967k ± 0% 967k ± 0% ~ (all equal) CmdGoSize 10.2M ± 0% 10.2M ± 0% ~ (all equal) Change-Id: I1f40af738254892bd6c8ba2eb43390b175753d52 Reviewed-on: https://go-review.googlesource.com/37445 Reviewed-by: Matthew Dempsky <mdempsky@google.com> Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2017-02-27 19:56:38 +02:00
t.SetDeferwidth(true)
deferredTypeStack = append(deferredTypeStack, t)
}
func defercheckwidth() {
// we get out of sync on syntax errors, so don't be pedantic.
if defercalc != 0 && nerrors == 0 {
Fatalf("defercheckwidth")
}
defercalc = 1
}
func resumecheckwidth() {
if defercalc == 0 {
Fatalf("resumecheckwidth")
}
for len(deferredTypeStack) > 0 {
t := deferredTypeStack[len(deferredTypeStack)-1]
deferredTypeStack = deferredTypeStack[:len(deferredTypeStack)-1]
cmd/compile: pack bool fields in Node, Name, Func and Type structs to bitsets This reduces compiler memory usage by up to 4% - see compilebench results below. name old time/op new time/op delta Template 245ms ± 4% 241ms ± 2% -1.88% (p=0.029 n=10+10) Unicode 126ms ± 3% 124ms ± 3% ~ (p=0.105 n=10+10) GoTypes 805ms ± 2% 813ms ± 3% ~ (p=0.515 n=8+10) Compiler 3.95s ± 2% 3.83s ± 1% -2.96% (p=0.000 n=9+10) MakeBash 47.4s ± 4% 46.6s ± 1% -1.59% (p=0.028 n=9+10) name old user-ns/op new user-ns/op delta Template 324M ± 5% 326M ± 3% ~ (p=0.935 n=10+10) Unicode 186M ± 5% 178M ±10% ~ (p=0.067 n=9+10) GoTypes 1.08G ± 7% 1.09G ± 4% ~ (p=0.956 n=10+10) Compiler 5.34G ± 4% 5.31G ± 1% ~ (p=0.501 n=10+8) name old alloc/op new alloc/op delta Template 41.0MB ± 0% 39.8MB ± 0% -3.03% (p=0.000 n=10+10) Unicode 32.3MB ± 0% 31.0MB ± 0% -4.13% (p=0.000 n=10+10) GoTypes 119MB ± 0% 116MB ± 0% -2.39% (p=0.000 n=10+10) Compiler 499MB ± 0% 487MB ± 0% -2.48% (p=0.000 n=10+10) name old allocs/op new allocs/op delta Template 380k ± 1% 379k ± 1% ~ (p=0.436 n=10+10) Unicode 324k ± 1% 324k ± 0% ~ (p=0.853 n=10+10) GoTypes 1.15M ± 0% 1.15M ± 0% ~ (p=0.481 n=10+10) Compiler 4.41M ± 0% 4.41M ± 0% -0.12% (p=0.007 n=10+10) name old text-bytes new text-bytes delta HelloSize 623k ± 0% 623k ± 0% ~ (all equal) CmdGoSize 6.64M ± 0% 6.64M ± 0% ~ (all equal) name old data-bytes new data-bytes delta HelloSize 5.81k ± 0% 5.81k ± 0% ~ (all equal) CmdGoSize 238k ± 0% 238k ± 0% ~ (all equal) name old bss-bytes new bss-bytes delta HelloSize 134k ± 0% 134k ± 0% ~ (all equal) CmdGoSize 152k ± 0% 152k ± 0% ~ (all equal) name old exe-bytes new exe-bytes delta HelloSize 967k ± 0% 967k ± 0% ~ (all equal) CmdGoSize 10.2M ± 0% 10.2M ± 0% ~ (all equal) Change-Id: I1f40af738254892bd6c8ba2eb43390b175753d52 Reviewed-on: https://go-review.googlesource.com/37445 Reviewed-by: Matthew Dempsky <mdempsky@google.com> Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
2017-02-27 19:56:38 +02:00
t.SetDeferwidth(false)
dowidth(t)
}
defercalc = 0
}