Because NameOffsetExpr is always used with global variables, and SSA
backend only needs (*Name).Linksym() to generate value for them.
Passes toolstash -cmp.
Updates #43737
Change-Id: I17209e21383edb766070c0accd1fa4660659caef
Reviewed-on: https://go-review.googlesource.com/c/go/+/284119
Trust: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Dump uses reflection to print IR nodes, and it only knew how to print
out the Nodes slice type itself. This CL adds support for printing any
slice whose element type implements Node, such as SwitchStmt and
SelectStmt's clause lists.
Change-Id: I2fd8defe11868b564d1d389ea3cd9b8abcefac62
Reviewed-on: https://go-review.googlesource.com/c/go/+/281537
Trust: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
OTYPE and OMETHEXPR were missing from OpPrec. So add them with the
same precedences as OT{ARRAY,MAP,STRUCT,etc} and
ODOT{,METH,INTER,etc}, respectively. However, ODEREF (which is also
used for pointer types *T) has a lower precedence than other types, so
pointer types need to be specially handled to assign them their
correct, lower precedence.
Incidentally, this also improves the error messages in issue15055.go,
where we were adding unnecessary parentheses around the types in
conversion expressions.
Thanks to Cuong Manh Le for writing the test cases for #43428.
Fixes#43428.
Change-Id: I57e7979babe3ed9ef8a8b5a2a3745e3737dd785f
Reviewed-on: https://go-review.googlesource.com/c/go/+/280873
Trust: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
If you do two ir.Dumps in a row, there's no newline between them.
Change-Id: I1a80dd22da68cb677eb9abd7a50571ea33584010
Reviewed-on: https://go-review.googlesource.com/c/go/+/280672
Trust: Keith Randall <khr@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
It's only used inside package ir now.
[git-generate]
cd src/cmd/compile/internal/ir
rf 'mv FmtNode fmtNode'
sed -i 's/FmtNode/fmtNode/g' mknode.go
go generate
Change-Id: Ib8f6c6984905a4d4cfca1b23972a39c5ea30ff42
Reviewed-on: https://go-review.googlesource.com/c/go/+/279451
Trust: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
These three expression nodes all represent the same syntax, and so
they're represented the same within types2. And also they're not
handled that meaningfully differently throughout the rest of the
compiler to merit unique representations.
Method expressions are somewhat unique today that they're very
frequently turned into plain function names. But eventually that can
be handled by a post-typecheck desugaring phase that reduces the
number of redundant AST forms.
Passes toolstash -cmp.
Change-Id: I20df91bbd0d885c1f18ec67feb61ae1558670719
Reviewed-on: https://go-review.googlesource.com/c/go/+/280636
Trust: Matthew Dempsky <mdempsky@google.com>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
Rename these two AST nodes to match their cmd/compile/internal/syntax
and go/ast counterparts.
Passes toolstash -cmp.
[git-generate]
cd src/cmd/compile/internal/ir
rf '
mv CaseStmt CaseClause
mv CommStmt CommClause
'
sed -E -i -e 's/(Case|Comm)Stmt/\1Clause/g' mknode.go
Change-Id: I19fba0323a5de1e71346622857011b2f7879bcef
Reviewed-on: https://go-review.googlesource.com/c/go/+/280446
Trust: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
ODOTMETH is unique among SelectorExpr expressions, in that Sel gets
mangled so that it no longer has the original identifier that was
selected (e.g., just "Foo"), but instead the qualified symbol name for
the selected method (e.g., "pkg.Type.Foo"). This is rarely useful, and
instead results in a lot of compiler code needing to worry about
undoing this change.
This CL changes ODOTMETH to leave the original symbol in place. The
handful of code locations where the mangled symbol name is actually
wanted are updated to use ir.MethodExprName(n).Sym() or (equivalently)
ir.MethodExprName(n).Func.Sym() instead.
Historically, the compiler backend has mistakenly used types.Syms
where it should have used ir.Name/ir.Funcs. And this change in
particular may risk breaking something, as the SelectorExpr.Sel will
no longer point at a symbol that uniquely identifies the called
method. However, I expect CL 280294 (desugar OCALLMETH into OCALLFUNC)
to have substantially reduced this risk, as ODOTMETH expressions are
now replaced entirely earlier in the compiler.
Passes toolstash -cmp.
Change-Id: If3c9c3b7df78ea969f135840574cf89e1d263876
Reviewed-on: https://go-review.googlesource.com/c/go/+/280436
Trust: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Sets up for removing Func from Node interface.
That means that once the Name reorg is done,
which will let us remove Name, Sym, and Val,
Node will be basically a minimal interface.
Passes buildall w/ toolstash -cmp.
Change-Id: I6e87897572debd7f8e29b4f5167763dc2792b408
Reviewed-on: https://go-review.googlesource.com/c/go/+/279484
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
For globals, Name.Offset is used as a way to address a field within
a global during static initialization. This CL replaces that use with
a separate NameOffsetExpr (ONAMEOFFSET) node.
For locals, Name.Offset is the stack frame offset. This CL calls it
that (FrameOffset, SetFrameOffset).
Now there is no longer any use of Name.Offset or Name.SetOffset.
And now that copies of Names are not being made to change their
offsets, we can lock down use of ir.Copy on Names. The only
remaining uses are during inlining and in handling generic system
functions. At both those times you do want to create a new name
and that can be made explicit by calling the new CloneName method
instead. ir.Copy on a name now panics.
Passes buildall w/ toolstash -cmp.
Change-Id: I0b0a25b9d93aeff7cf4e4025ac53faec7dc8603b
Reviewed-on: https://go-review.googlesource.com/c/go/+/278914
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
An automated rewrite will add concrete type assertions after
a test of n.Op(), when n can be safely type-asserted
(meaning, n is not reassigned a different type, n is not reassigned
and then used outside the scope of the type assertion,
and so on).
This sequence of CLs handles the code that the automated
rewrite does not: adding specific types to function arguments,
adjusting code not to call n.Left() etc when n may have multiple
representations, and so on.
This CL focuses on sinit.go.
Passes buildall w/ toolstash -cmp.
Change-Id: I3e9458e69a7a9b3f2fe139382bf961bc4473cc42
Reviewed-on: https://go-review.googlesource.com/c/go/+/277928
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
An automated rewrite will add concrete type assertions after
a test of n.Op(), when n can be safely type-asserted
(meaning, n is not reassigned a different type, n is not reassigned
and then used outside the scope of the type assertion,
and so on).
This sequence of CLs handles the code that the automated
rewrite does not: adding specific types to function arguments,
adjusting code not to call n.Left() etc when n may have multiple
representations, and so on.
This CL handles package fmt. There are various type assertions
but also some rewriting to lean more heavily on reflection.
Passes buildall w/ toolstash -cmp.
Change-Id: I503467468b42ace11bff2ba014b03cfa345e6d03
Reviewed-on: https://go-review.googlesource.com/c/go/+/277915
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
ir.Dump is the final (I think!) piece of the compiler that was walking
nodes using Left, Right etc without knowing what they meant.
This CL uses reflection to walk nodes without knowing what they mean instead.
One benefit is that we can print actual meanings (field names).
While we are here, I could not resist fixing a long-standing mental TODO:
make the line number more clearly a line number. I've forgotten where the
line number is in the dumps far too many times in the last decade.
As a small example, here is a fragment of go tool compile -W test/235.go:
. FOR l(28) tc(1)
. . LT-init
. . . AS l(28) tc(1)
. . . . NAME-main..autotmp_4 l(28) x(0) class(PAUTO) esc(N) tc(1) assigned used int
. . . . LEN l(28) tc(1) int
. . . . . NAME-main.xs g(2) l(26) x(0) class(PPARAM) esc(no) tc(1) used SLICE-[]uint64
. . LT l(28) tc(1) hascall bool
. . . NAME-main.i g(4) l(28) x(0) class(PAUTO) esc(no) tc(1) assigned used int
. . . NAME-main..autotmp_4 l(28) x(0) class(PAUTO) esc(N) tc(1) assigned used int
. . BLOCK l(28)
. . BLOCK-list
. . . ASOP-ADD l(28) tc(1) implicit(true) int
. . . . NAME-main.i g(4) l(28) x(0) class(PAUTO) esc(no) tc(1) assigned used int
. . . . LITERAL-1 l(28) tc(1) int
. FOR-body
. . VARKILL l(28) tc(1)
. . . NAME-main..autotmp_4 l(28) x(0) class(PAUTO) esc(N) tc(1) assigned used int
. . IF l(29) tc(1)
. . . LT l(29) tc(1) bool
. . . . INDEX l(29) tc(1) uint64
. . . . . NAME-main.xs g(2) l(26) x(0) class(PPARAM) esc(no) tc(1) used SLICE-[]uint64
. . . . . NAME-main.i g(4) l(28) x(0) class(PAUTO) esc(no) tc(1) assigned used int
. . . . NAME-main.m g(3) l(27) x(0) class(PAUTO) esc(no) tc(1) assigned used uint64
. . IF-body
. . . AS l(30) tc(1)
. . . . NAME-main.m g(3) l(27) x(0) class(PAUTO) esc(no) tc(1) assigned used uint64
. . . . INDEX l(30) tc(1) uint64
. . . . . NAME-main.xs g(2) l(26) x(0) class(PPARAM) esc(no) tc(1) used SLICE-[]uint64
. . . . . NAME-main.i g(4) l(28) x(0) class(PAUTO) esc(no) tc(1) assigned used int
and here it is after this CL:
. FOR tc(1) # 235.go:28
. FOR-Cond
. . LT-init
. . . AS tc(1) # 235.go:28
. . . . NAME-main..autotmp_4 x(0) class(PAUTO) esc(N) tc(1) assigned used int # 235.go:28
. . . . LEN tc(1) int # 235.go:28 int
. . . . . NAME-main.xs g(2) x(0) class(PPARAM) esc(no) tc(1) used SLICE-[]uint64 # 235.go:26
. . LT tc(1) hascall bool # 235.go:28 bool
. . . NAME-main.i g(4) x(0) class(PAUTO) esc(no) tc(1) assigned used int # 235.go:28
. . . NAME-main..autotmp_4 x(0) class(PAUTO) esc(N) tc(1) assigned used int # 235.go:28
. FOR-Post
. . BLOCK # 235.go:28
. . BLOCK-List
. . . ASOP-ADD tc(1) implicit(true) int # 235.go:28 int
. . . . NAME-main.i g(4) x(0) class(PAUTO) esc(no) tc(1) assigned used int # 235.go:28
. . . . LITERAL-1 tc(1) int # 235.go:28
. FOR-Body
. . VARKILL tc(1) # 235.go:28
. . . NAME-main..autotmp_4 x(0) class(PAUTO) esc(N) tc(1) assigned used int # 235.go:28
. . IF tc(1) # 235.go:29
. . IF-Cond
. . . LT tc(1) bool # 235.go:29 bool
. . . . INDEX tc(1) uint64 # 235.go:29 uint64
. . . . . NAME-main.xs g(2) x(0) class(PPARAM) esc(no) tc(1) used SLICE-[]uint64 # 235.go:26
. . . . . NAME-main.i g(4) x(0) class(PAUTO) esc(no) tc(1) assigned used int # 235.go:28
. . . . NAME-main.m g(3) x(0) class(PAUTO) esc(no) tc(1) assigned used uint64 # 235.go:27
. . IF-Body
. . . AS tc(1) # 235.go:30
. . . . NAME-main.m g(3) x(0) class(PAUTO) esc(no) tc(1) assigned used uint64 # 235.go:27
. . . . INDEX tc(1) uint64 # 235.go:30 uint64
. . . . . NAME-main.xs g(2) x(0) class(PPARAM) esc(no) tc(1) used SLICE-[]uint64 # 235.go:26
. . . . . NAME-main.i g(4) x(0) class(PAUTO) esc(no) tc(1) assigned used int # 235.go:28
Note in particular the clear marking of FOR-Cond, FOR-Post, FOR-Body compared to the original.
The only changes to a few test files are the improved field name lines, and of course the line numbers.
Passes buildall w/ toolstash -cmp.
Change-Id: I5b654d9d8ee898976d4c387742ea688a082bac78
Reviewed-on: https://go-review.googlesource.com/c/go/+/275785
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
On ir.Node, ir.Nodes, and ir.Op, # is ignored, so %#v is %v.
On ir.Node, %S is the same as %v.
On types.Type, # is ignored, so %#L is %L, %#v is %v.
On types.Type, 0 is ignored, so %0S is %S.
Rewrite all these using go test cmd/compile -r, plus a
few multiline formats mentioning %0S on types updated by hand.
Now the formats used in the compiler match the documentation
for the format methods, a minor miracle.
Passes buildall w/ toolstash -cmp.
Change-Id: I3d4a3fae543145a68da13eede91166632c5b1ceb
Reviewed-on: https://go-review.googlesource.com/c/go/+/275782
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Some cleanup left over from moving the Type and Sym formatting to types.
And then document what the type formats are, now that it's clear.
Passes buildall w/ toolstash -cmp.
Change-Id: I35cb8978f1627db1056cb8ab343ce6ba6c99afad
Reviewed-on: https://go-review.googlesource.com/c/go/+/275780
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Move the printing of types.Type and types.Sym out of ir
into package types, where it properly belongs. This wasn't
done originally (when the code was in gc) because the Type
and Sym printing was a bit tangled up with the Node printing.
But now they are untangled and can move into the correct
package.
This CL is automatically generated.
A followup CL will clean up a little bit more by hand.
Passes buildall w/ toolstash -cmp.
[git-generate]
cd src/cmd/compile/internal/ir
rf '
mv FmtMode fmtMode
mv FErr fmtGo
mv FDbg fmtDebug
mv FTypeId fmtTypeID
mv FTypeIdName fmtTypeIDName
mv methodSymName SymMethodName
mv BuiltinPkg LocalPkg BlankSym OrigSym NumImport \
fmtMode fmtGo symFormat sconv sconv2 symfmt SymMethodName \
BasicTypeNames fmtBufferPool InstallTypeFormats typeFormat tconv tconv2 fldconv FmtConst \
typefmt.go
mv typefmt.go cmd/compile/internal/types
'
cd ../types
mv typefmt.go fmt.go
Change-Id: I6f3fd818323733ab8446f00594937c1628760b27
Reviewed-on: https://go-review.googlesource.com/c/go/+/275779
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
It turns out that the FmtFlag is really only tracking the FmtLong and FmtShort bits,
and the others simply mirror the state of the FmtMode and are copied out and
back in repeatedly.
Simplify to FmtFlag being the verb itself ('S', 'L', or 'v').
Now there is only one formatting enumeration, making it a bit
easier to understand what's going on.
Passes buildall w/ toolstash -cmp.
Change-Id: I85bde2183eb22228fcf46d19d003401d588d9825
Reviewed-on: https://go-review.googlesource.com/c/go/+/275778
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
This code is now hardly used and not worth the complexity.
It also tangles together Nodes and Types in a way that keeps
this code in package ir instead of package types.
Passes buildall w/ toolstash -cmp.
Change-Id: I2e829c1f6b602acbdc8ab4aac3b798f9ded762ef
Reviewed-on: https://go-review.googlesource.com/c/go/+/275777
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Narrow the interface between package ir and package types
to make it easier to clean up the type formatting code all in one place.
Also introduce ir.BlankSym for use by OrigSym, so that later
OrigSym can move to package types without needing to reference
a variable of type ir.Node.
Passes buildall w/ toolstash -cmp.
Change-Id: I39fa419a1c8fb3318203e31cacc8d06399deeff9
Reviewed-on: https://go-review.googlesource.com/c/go/+/275776
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
nconvFmt calls base.Fatalf if mode is anything but FErr,
proving that the only formats that matter for nodes are
plain %v, %S, and %L. And the nodes formatter can only get to %v.
(%S and %v are the same; we'll clean that up separately.)
Node and Nodes can therefore ignore mode, and all the mode
code can be removed from those implementations, removing
quite a few layers of abstraction.
Op similarly only runs in one mode and can be simplified.
Passes buildall w/ toolstash -cmp.
Change-Id: Ibfd845033e9c68181a20fb81c8f3dd428463920a
Reviewed-on: https://go-review.googlesource.com/c/go/+/275775
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
The Node printing code is tangled up due to the multiple printing modes.
Split out the Dump mode into its own code, which clarifies it considerably.
We are going to have to change the code for the new Node representations,
so it is nice to have it in an understandable form first.
The output of Dump is unchanged except for the removal of spurious
mid-Dump blank lines that have been printed for a while but don't
really make sense and appear to be a bug.
The %+v verb on Op prints the name ("ADD" not "+"), matching
%+v on Node and %+v on Nodes to get Dump and DumpList formats.
Passes buildall w/ toolstash -cmp.
Change-Id: I07f0f245859f1f785e10bdd671855ca43c51b545
Reviewed-on: https://go-review.googlesource.com/c/go/+/275774
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
This code is a few layer of abstraction stacked up on top
of each other, and they're hard to see all at the same time
because the file is pretty mixed up. As much as I try to avoid
code rearrangement to keep history, this one is long overdue.
A followup CL will cut out some of the layers, and the diff will be
much clearer what's going on with the code ordered with
callers near callees, as it is now.
Passes buildall w/ toolstash -cmp.
Change-Id: Iffc49d43cf4be9fab47e2dd59a5f98930573350f
Reviewed-on: https://go-review.googlesource.com/c/go/+/275773
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
This is unreachable code - the only way short can be true is
if verb == 'S', but jconv is only called when verb == 'j'.
Simplify by removing.
Passes buildall w/ toolstash -cmp.
Change-Id: I27bd38319f72215069e940b320b5c82608e2651a
Reviewed-on: https://go-review.googlesource.com/c/go/+/275772
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
This CL consolidates and cleans up fmt.go's logic for skipping past
Nodes introduced during typechecking. This allows eliminating SetOrig
on ConvExpr and Name. Also changes ConstExpr.SetOrig to a panic for
good measure.
The only remaining SetOrig uses now are for rewriting multi-value
"f(g())" calls and "return g()" statements, and type-checking
composite literals. It should be possible to eliminate both of those
as well.
Passes buildall w/ toolstash -cmp.
Change-Id: I478aea1a17dfb7a784293b930bf9081637eb2d7a
Reviewed-on: https://go-review.googlesource.com/c/go/+/275179
Trust: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>
It's useful to have quick access to the types.Field that a given
selector or method value expression refer to. Previously we abused Opt
for this, but couldn't do that for OCALLPART because escape analysis
uses Opt.
Now that we have more flexibility, we can simply add additional
pointer fields for this. This also allows getting rid of an unneeded
ONAME node for OCALLPART.
Passes buildall w/ toolstash -cmp.
Change-Id: I980d7bdb19abfd0b6f58a232876861b88dee1e47
Reviewed-on: https://go-review.googlesource.com/c/go/+/275034
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>
Trust: Matthew Dempsky <mdempsky@google.com>
OEMPTY is an empty *statement*, but it confusingly
gets handled as an expression in a few places.
More confusingly, OEMPTY often has an init list,
making it not empty at all. Replace uses and analysis
of OEMPTY with OBLOCK instead.
Passes buildall w/ toolstash -cmp.
Change-Id: I8d4fcef151e4f441fa19b1b96da5272d778131d6
Reviewed-on: https://go-review.googlesource.com/c/go/+/274594
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
These are fairly rote implementations of structs appropriate to
each Op (or group of Ops).
The names of these are unknown except to ir.NodAt for now.
A later, automated change will introduce direct use of the types
throughout package gc.
(This CL is expressions; the previous one was statements.)
This is the last of the Ops that were previously handled by the
generic node struct, so that struct and its methods can be
and are deleted in this CL.
Passes buildall w/ toolstash -cmp.
Change-Id: I1703f35f24dcd3f7c5782a278e53c3fe04e87c37
Reviewed-on: https://go-review.googlesource.com/c/go/+/274109
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Using statement nodes restricts the set of valid SetOp operations,
because you can't SetOp across representation. Rewrite various
code to avoid crossing those as-yet-unintroduced boundaries.
In particular, code like
x, y := v.(T)
x, y := f()
x, y := m[k]
x, y := <-c
starts out with Op = OAS2, and then it turns into a specific Op
OAS2DOTTYPE, OAS2FUNC, OAS2MAPR, OAS2RECV, and then
later in walk is lowered to an OAS2 again.
In the middle, the specific forms move the right-hand side from
n.Rlist().First() to n.Right(), and then the conversion to OAS2 moves
it back. This is unnecessary and makes it hard for these all to
share an underlying Node implementation.
This CL changes these specific forms to leave the right-hand side
in n.Rlist().First().
Similarly, OSELRECV2 is really just a temporary form of OAS2.
This CL changes it to use same fields too.
Finally, this CL fixes the printing of OAS2 nodes in ir/fmt.go,
which formerly printed n.Right() instead of n.Rlist().
This results in a (correct!) update to cmd/compile/internal/logopt's
expected output: ~R0 = <N> becomes ~R0 = &y.b.
Passes buildall w/ toolstash -cmp.
Change-Id: I164aa2e17dc55bfb292024de53d7d250192ad64a
Reviewed-on: https://go-review.googlesource.com/c/go/+/274105
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
The type syntax is reused to stand in for the actual type once typechecked,
to avoid updating all the possible references to the original type syntax.
So all these implementations allow changing their Op from the raw syntax
like OTMAP to the finished form OTYPE, even though obviously the
representation does not change.
Passes buildall w/ toolstash -cmp.
Change-Id: I4acca1a5b35fa2f48ee08e8f1e5a330a004c284b
Reviewed-on: https://go-review.googlesource.com/c/go/+/274103
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Now that we have specific types for ONAME and ODCLFUNC nodes
(*Name and *Func), use them throughout the compiler to be more
precise about what data is being operated on.
This is a somewhat large CL, but once you start applying the types
in a few places, you end up needing to apply them to many other
places to keep everything type-checking. A lot of code also melts
away as types are added.
Passes buildall w/ toolstash -cmp.
Change-Id: I21dd9b945d701c470332bac5394fca744a5b232d
Reviewed-on: https://go-review.googlesource.com/c/go/+/274097
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Before this CL, an ODCLFUNC Node was represented by both
a node struct and a Func struct (and a Name for the ONAME,
which isn't changing here). Now Func can be repurposed as
the ODCLFUNC implementation, replacing the two structs
totaling 280+144 = 424 bytes (64-bit) with a single 320-byte struct.
Using the *Func as the node also gives us a clear, typed answer to
“which node should we use to represent functions?”
The next CL will clean up uses. This CL is just the trivial
change in representation.
Passes buildall w/ toolstash -cmp.
Change-Id: Ie6d670da91d6eb8d67a85f8f83630b9586dc7443
Reviewed-on: https://go-review.googlesource.com/c/go/+/274096
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Before this CL, an ONAME Node was represented by three structs
linked together: a node, a Name, and a Param. Previous CLs removed
OLABEL and OPACK from the set of nodes that knew about Name.
Now Name can be repurposed to *be* the ONAME Node implementation,
replacing three linked structs totaling 152+64+88 = 304 bytes (64-bit)
with a single 232-byte struct.
Many expressions in the code become simpler as well, without having
to use .Param. and sometimes even .Name().
(For a node n where n.Name() != nil, n.Name() == n.(*Name) now.)
Passes buildall w/ toolstash -cmp.
Change-Id: Ie719f1285c05623b9fd2faaa059e5b360a64b3be
Reviewed-on: https://go-review.googlesource.com/c/go/+/274094
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
These are only needed for a few opcodes, and we can avoid
wasting storage in every implementation by using the extension
interface pattern with a helper function for access.
Of course, in the current codebase, there is only one Node
implementation (*node) and it has these methods, so there
is no danger of a functional change in this particular CL.
Passes buildall w/ toolstash -cmp.
Change-Id: I440c6c232f1fe7b56b852a00dc530f8f49a6b12d
Reviewed-on: https://go-review.googlesource.com/c/go/+/274089
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
The plan is to introduce a Node interface that replaces the old *Node pointer-to-struct.
The previous CL defined an interface INode modeling a *Node.
This CL:
- Changes all references outside internal/ir to use INode,
along with many references inside internal/ir as well.
- Renames Node to node.
- Renames INode to Node
So now ir.Node is an interface implemented by *ir.node, which is otherwise inaccessible,
and the code outside package ir is now (clearly) using only the interface.
The usual rule is never to redefine an existing name with a new meaning,
so that old code that hasn't been updated gets a "unknown name" error
instead of more mysterious errors or silent misbehavior. That rule would
caution against replacing Node-the-struct with Node-the-interface,
as in this CL, because code that says *Node would now be using a pointer
to an interface. But this CL is being landed at the same time as another that
moves Node from gc to ir. So the net effect is to replace *gc.Node with ir.Node,
which does follow the rule: any lingering references to gc.Node will be told
it's gone, not silently start using pointers to interfaces. So the rule is followed
by the CL sequence, just not this specific CL.
Overall, the loss of inlining caused by using interfaces cuts the compiler speed
by about 6%, a not insignificant amount. However, as we convert the representation
to concrete structs that are not the giant Node over the next weeks, that speed
should come back as more of the compiler starts operating directly on concrete types
and the memory taken up by the graph of Nodes drops due to the more precise
structs. Honestly, I was expecting worse.
% benchstat bench.old bench.new
name old time/op new time/op delta
Template 168ms ± 4% 182ms ± 2% +8.34% (p=0.000 n=9+9)
Unicode 72.2ms ±10% 82.5ms ± 6% +14.38% (p=0.000 n=9+9)
GoTypes 563ms ± 8% 598ms ± 2% +6.14% (p=0.006 n=9+9)
Compiler 2.89s ± 4% 3.04s ± 2% +5.37% (p=0.000 n=10+9)
SSA 6.45s ± 4% 7.25s ± 5% +12.41% (p=0.000 n=9+10)
Flate 105ms ± 2% 115ms ± 1% +9.66% (p=0.000 n=10+8)
GoParser 144ms ±10% 152ms ± 2% +5.79% (p=0.011 n=9+8)
Reflect 345ms ± 9% 370ms ± 4% +7.28% (p=0.001 n=10+9)
Tar 149ms ± 9% 161ms ± 5% +8.05% (p=0.001 n=10+9)
XML 190ms ± 3% 209ms ± 2% +9.54% (p=0.000 n=9+8)
LinkCompiler 327ms ± 2% 325ms ± 2% ~ (p=0.382 n=8+8)
ExternalLinkCompiler 1.77s ± 4% 1.73s ± 6% ~ (p=0.113 n=9+10)
LinkWithoutDebugCompiler 214ms ± 4% 211ms ± 2% ~ (p=0.360 n=10+8)
StdCmd 14.8s ± 3% 15.9s ± 1% +6.98% (p=0.000 n=10+9)
[Geo mean] 480ms 510ms +6.31%
name old user-time/op new user-time/op delta
Template 223ms ± 3% 237ms ± 3% +6.16% (p=0.000 n=9+10)
Unicode 103ms ± 6% 113ms ± 3% +9.53% (p=0.000 n=9+9)
GoTypes 758ms ± 8% 800ms ± 2% +5.55% (p=0.003 n=10+9)
Compiler 3.95s ± 2% 4.12s ± 2% +4.34% (p=0.000 n=10+9)
SSA 9.43s ± 1% 9.74s ± 4% +3.25% (p=0.000 n=8+10)
Flate 132ms ± 2% 141ms ± 2% +6.89% (p=0.000 n=9+9)
GoParser 177ms ± 9% 183ms ± 4% ~ (p=0.050 n=9+9)
Reflect 467ms ±10% 495ms ± 7% +6.17% (p=0.029 n=10+10)
Tar 183ms ± 9% 197ms ± 5% +7.92% (p=0.001 n=10+10)
XML 249ms ± 5% 268ms ± 4% +7.82% (p=0.000 n=10+9)
LinkCompiler 544ms ± 5% 544ms ± 6% ~ (p=0.863 n=9+9)
ExternalLinkCompiler 1.79s ± 4% 1.75s ± 6% ~ (p=0.075 n=10+10)
LinkWithoutDebugCompiler 248ms ± 6% 246ms ± 2% ~ (p=0.965 n=10+8)
[Geo mean] 483ms 504ms +4.41%
[git-generate]
cd src/cmd/compile/internal/ir
: # We need to do the conversion in multiple steps, so we introduce
: # a temporary type alias that will start out meaning the pointer-to-struct
: # and then change to mean the interface.
rf '
mv Node OldNode
add node.go \
type Node = *OldNode
'
: # It should work to do this ex in ir, but it misses test files, due to a bug in rf.
: # Run the command in gc to handle gc's tests, and then again in ssa for ssa's tests.
cd ../gc
rf '
ex . ../arm ../riscv64 ../arm64 ../mips64 ../ppc64 ../mips ../wasm {
import "cmd/compile/internal/ir"
*ir.OldNode -> ir.Node
}
'
cd ../ssa
rf '
ex {
import "cmd/compile/internal/ir"
*ir.OldNode -> ir.Node
}
'
: # Back in ir, finish conversion clumsily with sed,
: # because type checking and circular aliases do not mix.
cd ../ir
sed -i '' '
/type Node = \*OldNode/d
s/\*OldNode/Node/g
s/^func (n Node)/func (n *OldNode)/
s/OldNode/node/g
s/type INode interface/type Node interface/
s/var _ INode = (Node)(nil)/var _ Node = (*node)(nil)/
' *.go
gofmt -w *.go
sed -i '' '
s/{Func{}, 136, 248}/{Func{}, 152, 280}/
s/{Name{}, 32, 56}/{Name{}, 44, 80}/
s/{Param{}, 24, 48}/{Param{}, 44, 88}/
s/{node{}, 76, 128}/{node{}, 88, 152}/
' sizeof_test.go
cd ../ssa
sed -i '' '
s/{LocalSlot{}, 28, 40}/{LocalSlot{}, 32, 48}/
' sizeof_test.go
cd ../gc
sed -i '' 's/\*ir.Node/ir.Node/' mkbuiltin.go
cd ../../../..
go install std cmd
cd cmd/compile
go test -u || go test -u
Change-Id: I196bbe3b648e4701662e4a2bada40bf155e2a553
Reviewed-on: https://go-review.googlesource.com/c/go/+/272935
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Define the interface for an IR node.
The next CL will shuffle the names and leave us with ir.Node being the interface.
Change-Id: Ifc40f7846d522cf99efa6b4e558bebb6db5218f9
Reviewed-on: https://go-review.googlesource.com/c/go/+/272934
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
If we want to break up package gc at all, we will need to move
the compiler IR it defines into a separate package that can be
imported by packages that gc itself imports. This CL does that.
It also removes the TINT8 etc aliases so that all code is clear
about which package things are coming from.
This CL is automatically generated by the script below.
See the comments in the script for details about the changes.
[git-generate]
cd src/cmd/compile/internal/gc
rf '
# These names were never fully qualified
# when the types package was added.
# Do it now, to avoid confusion about where they live.
inline -rm \
Txxx \
TINT8 \
TUINT8 \
TINT16 \
TUINT16 \
TINT32 \
TUINT32 \
TINT64 \
TUINT64 \
TINT \
TUINT \
TUINTPTR \
TCOMPLEX64 \
TCOMPLEX128 \
TFLOAT32 \
TFLOAT64 \
TBOOL \
TPTR \
TFUNC \
TSLICE \
TARRAY \
TSTRUCT \
TCHAN \
TMAP \
TINTER \
TFORW \
TANY \
TSTRING \
TUNSAFEPTR \
TIDEAL \
TNIL \
TBLANK \
TFUNCARGS \
TCHANARGS \
NTYPE \
BADWIDTH
# esc.go and escape.go do not need to be split.
# Append esc.go onto the end of escape.go.
mv esc.go escape.go
# Pull out the type format installation from func Main,
# so it can be carried into package ir.
mv Main:/Sconv.=/-0,/TypeLinkSym/-1 InstallTypeFormats
# Names that need to be exported for use by code left in gc.
mv Isconst IsConst
mv asNode AsNode
mv asNodes AsNodes
mv asTypesNode AsTypesNode
mv basicnames BasicTypeNames
mv builtinpkg BuiltinPkg
mv consttype ConstType
mv dumplist DumpList
mv fdumplist FDumpList
mv fmtMode FmtMode
mv goopnames OpNames
mv inspect Inspect
mv inspectList InspectList
mv localpkg LocalPkg
mv nblank BlankNode
mv numImport NumImport
mv opprec OpPrec
mv origSym OrigSym
mv stmtwithinit StmtWithInit
mv dump DumpAny
mv fdump FDumpAny
mv nod Nod
mv nodl NodAt
mv newname NewName
mv newnamel NewNameAt
mv assertRepresents AssertValidTypeForConst
mv represents ValidTypeForConst
mv nodlit NewLiteral
# Types and fields that need to be exported for use by gc.
mv nowritebarrierrecCallSym SymAndPos
mv SymAndPos.lineno SymAndPos.Pos
mv SymAndPos.target SymAndPos.Sym
mv Func.lsym Func.LSym
mv Func.setWBPos Func.SetWBPos
mv Func.numReturns Func.NumReturns
mv Func.numDefers Func.NumDefers
mv Func.nwbrCalls Func.NWBRCalls
# initLSym is an algorithm left behind in gc,
# not an operation on Func itself.
mv Func.initLSym initLSym
mv nodeQueue NodeQueue
mv NodeQueue.empty NodeQueue.Empty
mv NodeQueue.popLeft NodeQueue.PopLeft
mv NodeQueue.pushRight NodeQueue.PushRight
# Many methods on Node are actually algorithms that
# would apply to any node implementation.
# Those become plain functions.
mv Node.funcname FuncName
mv Node.isBlank IsBlank
mv Node.isGoConst isGoConst
mv Node.isNil IsNil
mv Node.isParamHeapCopy isParamHeapCopy
mv Node.isParamStackCopy isParamStackCopy
mv Node.isSimpleName isSimpleName
mv Node.mayBeShared MayBeShared
mv Node.pkgFuncName PkgFuncName
mv Node.backingArrayPtrLen backingArrayPtrLen
mv Node.isterminating isTermNode
mv Node.labeledControl labeledControl
mv Nodes.isterminating isTermNodes
mv Nodes.sigerr fmtSignature
mv Node.MethodName methodExprName
mv Node.MethodFunc methodExprFunc
mv Node.IsMethod IsMethod
# Every node will need to implement RawCopy;
# Copy and SepCopy algorithms will use it.
mv Node.rawcopy Node.RawCopy
mv Node.copy Copy
mv Node.sepcopy SepCopy
# Extract Node.Format method body into func FmtNode,
# but leave method wrapper behind.
mv Node.Format:0,$ FmtNode
# Formatting helpers that will apply to all node implementations.
mv Node.Line Line
mv Node.exprfmt exprFmt
mv Node.jconv jconvFmt
mv Node.modeString modeString
mv Node.nconv nconvFmt
mv Node.nodedump nodeDumpFmt
mv Node.nodefmt nodeFmt
mv Node.stmtfmt stmtFmt
# Constant support needed for code moving to ir.
mv okforconst OKForConst
mv vconv FmtConst
mv int64Val Int64Val
mv float64Val Float64Val
mv Node.ValueInterface ConstValue
# Organize code into files.
mv LocalPkg BuiltinPkg ir.go
mv NumImport InstallTypeFormats Line fmt.go
mv syntax.go Nod NodAt NewNameAt Class Pxxx PragmaFlag Nointerface SymAndPos \
AsNode AsTypesNode BlankNode OrigSym \
Node.SliceBounds Node.SetSliceBounds Op.IsSlice3 \
IsConst Node.Int64Val Node.CanInt64 Node.Uint64Val Node.BoolVal Node.StringVal \
Node.RawCopy SepCopy Copy \
IsNil IsBlank IsMethod \
Node.Typ Node.StorageClass node.go
mv ConstType ConstValue Int64Val Float64Val AssertValidTypeForConst ValidTypeForConst NewLiteral idealType OKForConst val.go
# Move files to new ir package.
mv bitset.go class_string.go dump.go fmt.go \
ir.go node.go op_string.go val.go \
sizeof_test.go cmd/compile/internal/ir
'
: # fix mkbuiltin.go to generate the changes made to builtin.go during rf
sed -i '' '
s/\[T/[types.T/g
s/\*Node/*ir.Node/g
/internal\/types/c \
fmt.Fprintln(&b, `import (`) \
fmt.Fprintln(&b, ` "cmd/compile/internal/ir"`) \
fmt.Fprintln(&b, ` "cmd/compile/internal/types"`) \
fmt.Fprintln(&b, `)`)
' mkbuiltin.go
gofmt -w mkbuiltin.go
: # update cmd/dist to add internal/ir
cd ../../../dist
sed -i '' '/compile.internal.gc/a\
"cmd/compile/internal/ir",
' buildtool.go
gofmt -w buildtool.go
: # update cmd/compile TestFormats
cd ../..
go install std cmd
cd cmd/compile
go test -u || go test # first one updates but fails; second passes
Change-Id: I5f7caf6b20629b51970279e81231a3574d5b51db
Reviewed-on: https://go-review.googlesource.com/c/go/+/273008
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
2020-11-25 16:53:33 +00:00
Renamed from src/cmd/compile/internal/gc/fmt.go (Browse further)
