This introduces a number of new classifications which will make it
easier to generate functions to assemble the new instructions of
ISA 3.1, and potentially earlier versions.
No code generation changes should occur as a result of these. These
allow finer control over how an opcode is matched to an optab entry.
Literal values are now classified based on the smallest number of bits
needed to encode, and matching rules will accept a literal if it
can be zero/sign extended to fit a larger literal class.
Likewise, support classifying even register numbers for GPR, VSX, and
FPR instructions. Some instructions require and even/odd register pair,
and these are usually represented by specifying the even register, and
similarly encoded.
Likewise, add a unit test for the argument classifier function (aclass).
This caught an off-by-one bug in aclass which is also fixed.
Updates #44549
Change-Id: Ia03013aea8b56c4d59b7c3812cdd67ddb3b720b9
Reviewed-on: https://go-review.googlesource.com/c/go/+/350152 Reviewed-by: Cherry Mui <cherryyz@google.com> Reviewed-by: Lynn Boger <laboger@linux.vnet.ibm.com>
Run-TryBot: Cherry Mui <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Dan Scales [Wed, 20 Oct 2021 20:56:15 +0000 (13:56 -0700)]
cmd/compile: fix inlining of labeled for loops
There is already a mechanism using inlgen to rename labels insided
inlined functions so that they are unique and don't clash with loops in
the outer function. This is used for OLABEL and OGOTO. Now that we are
doing inlining of OFOR loops, we need to do this translation for OBREAK,
OCONTINUE, and OFOR. I also added the translation for ORANGE loops, in
anticipation of a CL that will allow inlining of ORANGE for loops.
Fixes #49100
Change-Id: I2ccddc3350370825c386965f4a1e4bc54d3c369b
Reviewed-on: https://go-review.googlesource.com/c/go/+/357649
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Trust: Dan Scales <danscales@google.com>
Michael Anthony Knyszek [Thu, 8 Apr 2021 22:01:13 +0000 (22:01 +0000)]
runtime: detangle gcPaceScavenger from the pacer
Currently gcPaceScavenger is called by gcControllerState.commit, but it
manipulates global state which precludes testing. This change detangles
the two.
Change-Id: I10d8ebdf426d99ba49d2f2cb4fb64891e9fd6091
Reviewed-on: https://go-review.googlesource.com/c/go/+/309272 Reviewed-by: Michael Pratt <mpratt@google.com>
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Michael Anthony Knyszek [Thu, 8 Apr 2021 21:07:02 +0000 (21:07 +0000)]
runtime: formalize and fix gcPercent synchronization
Currently gcController.gcPercent is read non-atomically by
gcControllerState.revise and gcTrigger.test, but these users may
execute concurrently with an update to gcPercent.
Although revise's results are best-effort, reading it directly in this
way is, generally speaking, unsafe.
This change makes gcPercent atomically updated for concurrent readers
and documents the complete synchronization semantics.
Because gcPercent otherwise only updated with the heap lock held or the
world stopped, all other reads can remain unsynchronized.
For #44167.
Change-Id: If09af103aae84a1e133e2d4fed8ab888d4b8f457
Reviewed-on: https://go-review.googlesource.com/c/go/+/308690
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Michael Pratt <mpratt@google.com>
Robert Findley [Thu, 21 Oct 2021 17:23:29 +0000 (13:23 -0400)]
go/types, types2: always return an underlying type from optype
Optype should never return a defined type.
Change-Id: I37b29e0c958e127e75e834e71d6392ea80827773
Reviewed-on: https://go-review.googlesource.com/c/go/+/357694
Trust: Robert Findley <rfindley@google.com>
Run-TryBot: Robert Findley <rfindley@google.com> Reviewed-by: Robert Griesemer <gri@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Amelia Downs [Mon, 18 Oct 2021 16:36:07 +0000 (12:36 -0400)]
net/url: add testable examples for Values funcs
Change-Id: Id71f3d8d7c1ef7910d5d9497167dc677f2f0a2ef
Reviewed-on: https://go-review.googlesource.com/c/go/+/356535
Trust: Damien Neil <dneil@google.com>
Trust: Cherry Mui <cherryyz@google.com>
Run-TryBot: Damien Neil <dneil@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Damien Neil <dneil@google.com>
Amelia Downs [Mon, 18 Oct 2021 17:06:58 +0000 (13:06 -0400)]
net/url: add testable examples for url pkg funcs
Change-Id: I61011b75128478aa50308d84f4cba23b3e241b3f
Reviewed-on: https://go-review.googlesource.com/c/go/+/356536
Trust: Damien Neil <dneil@google.com>
Trust: Cherry Mui <cherryyz@google.com>
Run-TryBot: Damien Neil <dneil@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Damien Neil <dneil@google.com>
Archana R [Tue, 19 Oct 2021 09:11:46 +0000 (04:11 -0500)]
internal/bytealg: fix Separator length check for Index/ppc64le
Modified condition in the ASM implementation of indexbody to
determine if separator length crosses 16 bytes to BGT from BGE
to avoid incorrectly crossing a page.
Also fixed IndexString to invoke indexbodyp9 when on the POWER9
platform
Change-Id: I0602a797cc75287990eea1972e9e473744f6f5a9
Reviewed-on: https://go-review.googlesource.com/c/go/+/356849
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Lynn Boger <laboger@linux.vnet.ibm.com>
Trust: Keith Randall <khr@golang.org>
Lynn Boger [Thu, 14 Oct 2021 17:55:43 +0000 (12:55 -0500)]
cmd/compile/internal/ssa/gen: eliminate unnecessary neg and xori on PPC64
This adds a few rules to PPC64 to eliminate some instructions:
- when an isel is used to generate a boolean value based on a
condition and followed by an xori to flip the result, it can
instead flip the operands in the isel and avoid the xori.
= when a neg follows a sub the operands to the sub can be
swapped and the neg avoided.
There are several opportunities in reflect.DeepEqual to omit
xori which improves some of its benchmarks by as much as
5%
Michael Anthony Knyszek [Mon, 18 Oct 2021 23:14:20 +0000 (23:14 +0000)]
runtime: retype mheap.reclaimCredit as atomic.Uintptr
[git-generate]
cd src/runtime
mv export_test.go export.go
GOROOT=$(dirname $(dirname $PWD)) rf '
add mheap.reclaimCredit \
// reclaimCredit is spare credit for extra pages swept. Since \
// the page reclaimer works in large chunks, it may reclaim \
// more than requested. Any spare pages released go to this \
// credit pool. \
reclaimCredit_ atomic.Uintptr
ex {
import "runtime/internal/atomic"
Michael Anthony Knyszek [Mon, 18 Oct 2021 23:12:16 +0000 (23:12 +0000)]
runtime: retype mheap.reclaimIndex as atomic.Uint64
[git-generate]
cd src/runtime
mv export_test.go export.go
GOROOT=$(dirname $(dirname $PWD)) rf '
add mheap.reclaimIndex \
// reclaimIndex is the page index in allArenas of next page to \
// reclaim. Specifically, it refers to page (i % \
// pagesPerArena) of arena allArenas[i / pagesPerArena]. \
// \
// If this is >= 1<<63, the page reclaimer is done scanning \
// the page marks. \
reclaimIndex_ atomic.Uint64
ex {
import "runtime/internal/atomic"
Michael Anthony Knyszek [Mon, 18 Oct 2021 23:10:43 +0000 (23:10 +0000)]
runtime: retype mheap.pagesSweptBasis as atomic.Uint64
[git-generate]
cd src/runtime
mv export_test.go export.go
GOROOT=$(dirname $(dirname $PWD)) rf '
add mheap.pagesSweptBasis pagesSweptBasis_ atomic.Uint64 // pagesSwept to use as the origin of the sweep ratio
ex {
import "runtime/internal/atomic"
Michael Anthony Knyszek [Fri, 15 Oct 2021 19:22:10 +0000 (19:22 +0000)]
runtime: retype mheap.pagesInUse as atomic.Uint64
[git-generate]
cd src/runtime
mv export_test.go export.go
GOROOT=$(dirname $(dirname $PWD)) rf '
add mheap.pagesInUse \
// Proportional sweep \
// \
// These parameters represent a linear function from gcController.heapLive \
// to page sweep count. The proportional sweep system works to \
// stay in the black by keeping the current page sweep count \
// above this line at the current gcController.heapLive. \
// \
// The line has slope sweepPagesPerByte and passes through a \
// basis point at (sweepHeapLiveBasis, pagesSweptBasis). At \
// any given time, the system is at (gcController.heapLive, \
// pagesSwept) in this space. \
// \
// It is important that the line pass through a point we \
// control rather than simply starting at a 0,0 origin \
// because that lets us adjust sweep pacing at any time while \
// accounting for current progress. If we could only adjust \
// the slope, it would create a discontinuity in debt if any \
// progress has already been made. \
pagesInUse_ atomic.Uint64 // pages of spans in stats mSpanInUse
ex {
import "runtime/internal/atomic"
Robert Findley [Wed, 20 Oct 2021 18:54:13 +0000 (14:54 -0400)]
go/internal/gcimporter: avoid setting unnecessary lines in fakeFileSet
This is a clean port of CL 357291 from x/tools.
For #46586
Change-Id: Ib22087ae7fe8477d368acd230b263b83cdad4d36
Reviewed-on: https://go-review.googlesource.com/c/go/+/357429
Trust: Robert Findley <rfindley@google.com>
Run-TryBot: Robert Findley <rfindley@google.com> Reviewed-by: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Cuong Manh Le [Mon, 18 Oct 2021 07:31:48 +0000 (14:31 +0700)]
cmd/compile: fix crawling of embeddable types during inline
In CL 327872, there's a fix for crawling of embeddable types directly
reached by the user, so all of its methods need to be re-exported. But
we missed the cased when an un-exported type may be reachable by
embedding in exported type. Example:
type t struct {}
func (t) M() {}
func F() interface{} { return struct{ t }{} }
We generate the wrapper for "struct{ t }".M, and when inlining call to
"struct{ t }".M makes "t.M" reachable.
It works well, and only be revealed in CL 327871, when we changed
methodWrapper to always call inline.InlineCalls, thus causes the crash
in #49016, which involve dot type in inlined function.
Fixes #49016
Change-Id: If174fa5575132da5cf60e4bd052f7011c4e76c5d
Reviewed-on: https://go-review.googlesource.com/c/go/+/356254
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>
Than McIntosh [Wed, 20 Oct 2021 14:43:43 +0000 (10:43 -0400)]
cmd/internal/goobj: minor dead code cleanup
Remove a bit of dead code from the Go object file reader (io.ReaderAt
no longer needed in goobj.Reader).
Change-Id: I04150d37fb90b59c9dbe930878d4dd21cdcd7ca7
Reviewed-on: https://go-review.googlesource.com/c/go/+/357309
Trust: Than McIntosh <thanm@google.com>
Run-TryBot: Than McIntosh <thanm@google.com> Reviewed-by: Cherry Mui <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Cuong Manh Le [Wed, 20 Oct 2021 02:55:26 +0000 (09:55 +0700)]
cmd/compile/internal/types2: print assignment operation for invalid operation errors
When invoking check.binary for assignment operation, the expression will
be nil, thus for printing the assignment operation error message, we
need to reconstruct the statement from lhs, op, rhs.
Fixes #48472
Change-Id: Ie38c3dd8069b47e508968d6e43cedcf7536559ef
Reviewed-on: https://go-review.googlesource.com/c/go/+/357229
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: Robert Griesemer <gri@golang.org>
Keith Randall [Mon, 18 Oct 2021 17:59:29 +0000 (10:59 -0700)]
cmd/compile: allow importing and exporting of ODYNAMICTYPE
Change-Id: I2fca7a801c85ed93c002c23bfcb0cf9593f1bdf4
Reviewed-on: https://go-review.googlesource.com/c/go/+/356571
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Dan Scales <danscales@google.com>
Keith Randall [Mon, 18 Oct 2021 19:19:07 +0000 (12:19 -0700)]
cmd/compile: substitute "". prefix with package in more import locations
The "" stand-in for the current package needs to be substituted
in more places when importing, because of generics.
""..dict.conv4["".MyString]
when imported in main and then exported, this becomes
a..dict.conv4["".MyString]
and then the linker makes that into
a..dict.conv4[main.MyString]
Which isn't correct. We need to replace on import not just
function names, but also globals, which this CL does.
Change-Id: Ia04a23b5ffd60aeeaba72c807f69261105670f8e
Reviewed-on: https://go-review.googlesource.com/c/go/+/356570
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Dan Scales <danscales@google.com>
Robert Griesemer [Tue, 19 Oct 2021 20:44:01 +0000 (13:44 -0700)]
go/types, types2: always accept type parameters when running Manual test
This makes it easier to run tests on existing generic code that is
not using the `.go2` ending currently used by type checker tests.
For #49074.
Change-Id: I1501463c55dbe1d709918786e1a43b6d1bf1629a
Reviewed-on: https://go-review.googlesource.com/c/go/+/357050
Trust: Robert Griesemer <gri@golang.org>
Run-TryBot: Robert Griesemer <gri@golang.org> Reviewed-by: Robert Findley <rfindley@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Cuong Manh Le [Tue, 19 Oct 2021 17:09:09 +0000 (00:09 +0700)]
reflect: restore Value.Pointer implementation
CL 350691 added Value.UnsafePointer and make Value.Pointer call it
internally. It has a downside that Value.Pointer can now eligible to be
inlined, thus making un-intentional side effect, like the test in
fixedbugs/issue15329.go becomes flaky.
This CL restore Value.Pointer original implementation, pre CL 350691,
with the deprecation TODO removed.
Fixes #49067
Change-Id: I735af182f8e729294333ca906ffc062f477cfc99
Reviewed-on: https://go-review.googlesource.com/c/go/+/356949
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: Keith Randall <khr@golang.org>
Michael Matloob [Fri, 17 Sep 2021 23:38:33 +0000 (19:38 -0400)]
cmd/go: add GOWORK to go env command
GOWORK will be set to the go.work file's path, if in workspace mode
or will be empty otherwise.
For #45713
Fixes #48589
Change-Id: I163ffaf274e0a41469c1f3b8514d6f90e20423b0
Reviewed-on: https://go-review.googlesource.com/c/go/+/355689
Trust: Michael Matloob <matloob@golang.org>
Trust: Daniel Martí <mvdan@mvdan.cc>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Bryan C. Mills <bcmills@google.com>
Michael Matloob [Fri, 17 Sep 2021 23:38:33 +0000 (19:38 -0400)]
cmd/go: support replaces in the go.work file
Add support for replace directives in the go.work file. If there are
conflicting replaces in go.mod files, suggest that users add an
overriding replace in the go.work file.
Add HighestReplaced to MainModules so that it accounts for the
replacements in the go.work file.
(Reviewers: I'm not totally sure that HighestReplace is computed
correctly. Could you take a closer look at that?)
For #45713
Change-Id: I1d789219ca1dd065ba009ce5d38db9a1fc38ba83
Reviewed-on: https://go-review.googlesource.com/c/go/+/352812
Trust: Michael Matloob <matloob@golang.org>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Bryan C. Mills <bcmills@google.com>
Matt T. Proud [Tue, 19 Oct 2021 08:09:06 +0000 (10:09 +0200)]
errors: mention Is methods should not call Unwrap
errors.Is internally unwraps the error until the error matches the
target. Because of this, a user-authored Is method on an error type
need not call errors.Unwrap on itself or the target, because that would
make the unwrapping operation O(N^2). It is a subtle detail to remind
authors for resource efficiency reasons.
Change-Id: Ic1ba59a5bdbfe2c7cb51a2cba2537ab6de4a13ff
Reviewed-on: https://go-review.googlesource.com/c/go/+/356789 Reviewed-by: Jean de Klerk <deklerk@google.com> Reviewed-by: Damien Neil <dneil@google.com>
Trust: Jean de Klerk <deklerk@google.com>
Trust: Damien Neil <dneil@google.com>
Run-TryBot: Jean de Klerk <deklerk@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Carlo Alberto Ferraris [Thu, 14 Oct 2021 07:50:41 +0000 (16:50 +0900)]
sync: avoid a dynamic check in WaitGroup on 64-bit architectures
uint64 is guaranteed by the compiler to be aligned on 64-bit archs.
By using uint64+uint32 instead of [3]uint32 we can make use of the
guaranteed alignment to avoid the run-time alignment check.
Change-Id: I314340f2ed8a47d8b9c15f8a3b07e41f252f4831
Reviewed-on: https://go-review.googlesource.com/c/go/+/189837 Reviewed-by: Cherry Mui <cherryyz@google.com> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Cherry Mui <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Katie Hockman [Fri, 15 Oct 2021 15:16:54 +0000 (11:16 -0400)]
testing: don't allow f.Log/Logf or f.Skipped inside f.Fuzz
This change also does some refactors around how
we prevent many (*F) methods from being called
inside (*F).Fuzz. Previously, there was a lot of
comment/code duplication, which was going to be
difficult to maintain and brittle. The refactor
lessens this duplication.
Previously, the methods Log, Logf, Failed, Name and
Skipped were the only (*common) methods that were
allowed to be called inside (*F).Fuzz. After this
change, Failed and Name are still allowed, but
Log, Logf, and Skipped are not (t.Log, t.Logf, or
t.Skipped should be used instead).
Fixes #48988
Change-Id: I4066247d551ea1908e8a2ca2889509fc68e3bb44
Reviewed-on: https://go-review.googlesource.com/c/go/+/356151
Trust: Katie Hockman <katie@golang.org>
Run-TryBot: Katie Hockman <katie@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Bryan C. Mills <bcmills@google.com> Reviewed-by: Jay Conrod <jayconrod@google.com>
Robert Griesemer [Mon, 18 Oct 2021 21:56:08 +0000 (14:56 -0700)]
cmd/compile, types2: better error message for invalid type assertion
This CL addresses the 2nd part of the issue below.
- For types2, now use the same error messages as the compiler in this case.
- Make the mechanism for reporting clarifying error messages handle the case
where we don't have additional position information.
- Provide context information (type assertion vs type switch).
Fixes #49005.
Change-Id: I4eeaf4f0c3f2f8735b63993778f58d713fef21ee
Reviewed-on: https://go-review.googlesource.com/c/go/+/356512
Trust: Robert Griesemer <gri@golang.org> Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com> Reviewed-by: Robert Findley <rfindley@google.com>
CL 322350 changed how to take address of assembly functions, using
abi.FuncPCABI0 intrinsic. But we forgot to update the code in
Value.UnsafePointer (was Value.Pointer) to reflect that change.
This CL fixes that bug, and also add a test to make sure the code
pointer is in sync.
Change-Id: I05ae7df31c706583a0f374d8af027066528f5ceb
Reviewed-on: https://go-review.googlesource.com/c/go/+/356809
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: Cherry Mui <cherryyz@google.com>
wdvxdr [Tue, 19 Oct 2021 11:39:21 +0000 (19:39 +0800)]
cmd/compile: use MOVBE instruction for GOAMD64>=v3
In CL 354670, I copied some existing rules for convenience but forgot
to update the last rule which broke `GOAMD64=v3 ./make.bat`
Revive CL 354670
Change-Id: Ic1e2047c603f0122482a4b293ce1ef74d806c019
Reviewed-on: https://go-review.googlesource.com/c/go/+/356810 Reviewed-by: Daniel Martí <mvdan@mvdan.cc> Reviewed-by: Keith Randall <khr@golang.org>
Trust: Daniel Martí <mvdan@mvdan.cc>
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Go Bot <gobot@golang.org>
Joel Sing [Fri, 15 Oct 2021 16:59:41 +0000 (03:59 +1100)]
cmd/internal/obj/riscv: fix trampoline calls from large functions
On riscv64, the JAL instruction is only capable of reaching +/-1MB. In the case where
a single function and its trampolines exceeds this size, it is possible that the JAL
is unable to reach the trampoline, which is laid down after the function text. In the
case of large functions, switch back to using a AUIPC+JALR pairs rather than using
trampolines.
Fixes #48791
Change-Id: I119cf3bc20ce4933a9b7ab41a8e514437c6addb9
Reviewed-on: https://go-review.googlesource.com/c/go/+/356250
Trust: Joel Sing <joel@sing.id.au> Reviewed-by: Cherry Mui <cherryyz@google.com>
Change-Id: I83e5f6105748a0a9238322a4f7ec4b0bbf61a263
Reviewed-on: https://go-review.googlesource.com/c/go/+/348394 Reviewed-by: Daniel Martí <mvdan@mvdan.cc> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Trust: Daniel Martí <mvdan@mvdan.cc>
Trust: Cherry Mui <cherryyz@google.com>
Joel Sing [Fri, 8 Oct 2021 09:28:58 +0000 (20:28 +1100)]
cmd/internal/obj/riscv: fix and enable test in short mode
The branch test only takes a few seconds so enable it in short mode. Also fix a
typo that currently prevents the code from compiling.
Change-Id: I25fc646b002a66c042c61fb1f56d5efa8e47903c
Reviewed-on: https://go-review.googlesource.com/c/go/+/356249
Trust: Joel Sing <joel@sing.id.au> Reviewed-by: Cherry Mui <cherryyz@google.com>
Run-TryBot: Cherry Mui <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Robert Findley [Fri, 15 Oct 2021 20:39:39 +0000 (16:39 -0400)]
go/types: ensure named types are expanded after type-checking
Rather than using Checker.later in newNamed, add a Checker.defTypes
field to track named types that have been created during type-checking,
and use this to expand named types as a final phase in type checking.
We have encountered several bugs related to infinite recursion while
expanding named types, because (I would argue) we have two conflicting
requirements in the type checker: ensuring that we eventually collapse
underlying chains, and yet allowing lazy substitution of the underlying
type in instances. The former is necessary for correctness, and to
ensure that we detect cycles during the type-checking pass. The latter
is necessary to allow infinitely expanding patterns of instances through
underlying or method definitions.
I believe this CL reconciles these conflicting requirements, by creating
a boundary between types that are encountered in the source during
type checking, and instances that are created by recursive evaluation.
At the end of the type checking pass, Checker.defTypes should contain
all possible origin types for instantiation. Once we compute the true
underlying for these origin types, any remaining instances that are
unresolved are guaranteed to have an origin with a valid underlying.
Therefore, we can return from the type-checking pass without calling
under() for these remaining instances.
Fixes #48703
Fixes #48974
Change-Id: I1474f514e2ab71c1ad4c3704fe32bfba11d59394
Reviewed-on: https://go-review.googlesource.com/c/go/+/356490
Trust: Robert Findley <rfindley@google.com>
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Robert Griesemer <gri@golang.org>
Robert Findley [Fri, 15 Oct 2021 14:52:55 +0000 (10:52 -0400)]
go/types: add support for inferring type instances
Add constraint type inference for type instances, to be consistent with
inference of function values.
Fixes #47990
Change-Id: Ib99b5215cb2da5c10badc4de7e9e60ca0e48489f
Reviewed-on: https://go-review.googlesource.com/c/go/+/356489
Trust: Robert Findley <rfindley@google.com>
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Robert Griesemer <gri@golang.org>
Damien Neil [Fri, 6 Aug 2021 20:23:13 +0000 (13:23 -0700)]
bufio: use underlying ReadFrom even when data is buffered
When (*bufio.Writer).ReadFrom is called with a partially filled buffer,
fill out and flush the buffer and then call the underlying writer's
ReadFrom method if present.
Fixes #44815.
Change-Id: I15b3ef0746d0d60fd62041189a9b9df11254dd29
Reviewed-on: https://go-review.googlesource.com/c/go/+/340530
Trust: Damien Neil <dneil@google.com>
Run-TryBot: Damien Neil <dneil@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Lynn Boger [Thu, 14 Oct 2021 15:48:08 +0000 (10:48 -0500)]
cmd/compile/internal/ssagen: set BitLen32 as intrinsic on PPC64
It was noticed through some other investigation that BitLen32
was not generating the best code and found that it wasn't recognized
as an intrinsic. This corrects that and enables the test for PPC64.
Change-Id: Iab496a8830c8552f507b7292649b1b660f3848b5
Reviewed-on: https://go-review.googlesource.com/c/go/+/355872
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Lynn Boger <laboger@linux.vnet.ibm.com> Reviewed-by: Cherry Mui <cherryyz@google.com>
Keith Randall [Mon, 18 Oct 2021 17:26:18 +0000 (10:26 -0700)]
cmd/compile: allow importing and exporting of ODYANMICDOTTYPE[2]
Fixes #49027
Change-Id: I4520b5c754027bfffbc5cd92c9c27002b248c99a
Reviewed-on: https://go-review.googlesource.com/c/go/+/356569
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Keith Randall <khr@golang.org> Reviewed-by: Dan Scales <danscales@google.com>
Allowing eliminates a class of possible misuse of unsafe.Pointer, and
allow callers to migrate from Value.Addr and Value.Pointer, thus they
can be now deprecated.
Fixes #40592
Change-Id: I798e507c748922cac5cc1c1971c1b2cc7095a068
Reviewed-on: https://go-review.googlesource.com/c/go/+/350691
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: Keith Randall <khr@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
reflect: correct documentation of Value.UnsafeAddr
The doc mentions that "UnsafeAddr returns a _pointer_ to v's data", but
it returns a uintptr instead, which don't have pointer semantic.
Change-Id: I557d5597cbc485356ca803eb496a99d6db8c63ba
Reviewed-on: https://go-review.googlesource.com/c/go/+/350690
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: Ian Lance Taylor <iant@golang.org> Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Change-Id: Ifda0d48d54665cef45d46d3aad974062633142c4
Reviewed-on: https://go-review.googlesource.com/c/go/+/354670
Run-TryBot: Alberto Donizetti <alb.donizetti@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
Trust: Matthew Dempsky <mdempsky@google.com>
wdvxdr [Wed, 13 Oct 2021 11:44:38 +0000 (19:44 +0800)]
cmd/compile: allow inlining labeled for-statement and switch-statement
After CL 349012 and CL 350911, we can fully handle these
labeled statements, so we can allow them when inlining.
Updates #14768
Change-Id: I0ab3fd3f8d7436b49b1aedd946516b33c63f5747
Reviewed-on: https://go-review.googlesource.com/c/go/+/355497
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Matthew Dempsky <mdempsky@google.com> Reviewed-by: Dan Scales <danscales@google.com> Reviewed-by: David Chase <drchase@google.com>
Trust: Dan Scales <danscales@google.com>
Robert Griesemer [Sun, 17 Oct 2021 03:30:12 +0000 (20:30 -0700)]
cmd/compile, types2: avoid confusing follow-on error in invalid type assertion
This CL avoids a useless follow-on error (that gets reported before the
actual error due to source position). This addresses the first part of
the issue below.
Thanks to @cuonglm for the suggestion for the fix.
For #49005.
Change-Id: Ifdd83072a05c32e115dc58a0233868a64f336f3f
Reviewed-on: https://go-review.googlesource.com/c/go/+/356449
Trust: Robert Griesemer <gri@golang.org>
Run-TryBot: Robert Griesemer <gri@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Robert Griesemer [Sat, 16 Oct 2021 21:54:42 +0000 (14:54 -0700)]
go/types, types2: add test case for missing return
The respective issue was fixed in types2 with CL 356189;
and the problem didn't exist in go/types. This CL simply
adds the test case to the type checkers as well.
For #49003.
Change-Id: Ib50ee8bb0ad21f2916f2b79d4f77593302899a3e
Reviewed-on: https://go-review.googlesource.com/c/go/+/356411
Trust: Robert Griesemer <gri@golang.org>
Run-TryBot: Robert Griesemer <gri@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Richard Musiol [Fri, 3 Sep 2021 19:58:57 +0000 (21:58 +0200)]
misc/wasm: expect environment to provide polyfills
The list of environments to support with wasm_exec.js was becoming too
large to maintain. With this change, wasm_exec.js expects that the
environment provides all necessary polyfills.
The standardized "globalThis" is used for accessing the environment.
wasm_exec.js now only provides stub fallbacks for globalThis.fs and
globalThis.process.
All code specific to Node.js is now in a separate file.
Change-Id: I076febbd94d4d7845260faad972f450f74a7b983
Reviewed-on: https://go-review.googlesource.com/c/go/+/347353
Trust: Richard Musiol <neelance@gmail.com> Reviewed-by: Cherry Mui <cherryyz@google.com>
Dan Scales [Fri, 15 Oct 2021 19:31:39 +0000 (12:31 -0700)]
cmd/compile: cleanup code in getInstInfo to use switch statement
Simple cleanup: convert a bunch of if's to a switch statement in
getInstInfo. Also, use a few extra variables to avoid repeated node
conversions (such as n.(*ir.CallExpr))
Change-Id: I7a2a4efb569415256a8bc9350fb100bd2d8cfb39
Reviewed-on: https://go-review.googlesource.com/c/go/+/356311
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Dan Scales [Fri, 15 Oct 2021 18:54:25 +0000 (11:54 -0700)]
cmd/compile: convert to using a map in getInstInfo, rather than SetImplicit()
SetImplicit() has an explicit meaning and really shouldn't be used in
this way - its use is left over from early prototype of the dictionary
code. Convert from using SetImplicit to just using a map during
traversal.
Change-Id: I3d257c101a859f000e159d7ced307d1b7cf990d4
Reviewed-on: https://go-review.googlesource.com/c/go/+/356310
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Michael Pratt [Thu, 14 Oct 2021 22:18:49 +0000 (18:18 -0400)]
runtime: consistently access pollDesc r/w Gs with atomics
Both netpollblock and netpollunblock read gpp using a non-atomic load.
When consuming a ready event, netpollblock clears gpp using a non-atomic
store, thus skipping a barrier.
Thus on systems with weak memory ordering, a sequence like so this is
possible:
i.e., without a happens-before edge between (2) and (3), netpollunblock
may read the stale value of gpp.
Switch these access to use atomic loads and stores in order to create
these edges.
For ease of future maintainance, I've simply changed rg and wg to always
be accessed atomically, though I don't believe pollOpen or pollClose
require atomics today.
Fixes #48925
Change-Id: I903ea667eea320277610b4f969129935731520c3
Reviewed-on: https://go-review.googlesource.com/c/go/+/355952
Trust: Michael Pratt <mpratt@google.com>
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Michael Knyszek <mknyszek@google.com> Reviewed-by: David Chase <drchase@google.com>
Jay Conrod [Fri, 15 Oct 2021 18:57:58 +0000 (11:57 -0700)]
cmd/go: allow nested VCS repositories when preparing build stamp
The go command no longer reports an error when invoked in a repository
nested inside another. This check is still used by 'go get' in GOPATH
mode when locating a repository, but it's not needed when preparing
the build stamp.
Fixes #49004
Change-Id: I4ed4dcc04174d2d42da8651d47e52ab1d7d66e35
Reviewed-on: https://go-review.googlesource.com/c/go/+/356309
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com> Reviewed-by: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Keith Randall [Fri, 15 Oct 2021 15:06:58 +0000 (08:06 -0700)]
cmd/compile: make for loops with range statements not terminating
Fixes #49003
Change-Id: If09c6f028dce5440b1be238612653ffdd626113a
Reviewed-on: https://go-review.googlesource.com/c/go/+/356189
Trust: Keith Randall <khr@golang.org> Reviewed-by: roger peppe <rogpeppe@gmail.com> Reviewed-by: Matthew Dempsky <mdempsky@google.com> Reviewed-by: Robert Griesemer <gri@golang.org> Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Katie Hockman [Wed, 13 Oct 2021 20:49:27 +0000 (16:49 -0400)]
internal/fuzz: fix bugs with minimization
This pulls in some code and tests from CL 353355.
This change makes some refactors for when we read
to and write from memory during minimization.
That fixes a bug when minimizing interesting inputs.
Now, if an error occurs while minimizing an interesting
input, that value will continue to be minimized as a
crash, and returned to the user.
This change also allows minimization of a crash that
occurred during the warmup phase. We don't want to
minimize failures in the seed corpus, but if an entry
in the cache causes a new failure, then there's no
compelling reason why we shouldn't try to minimize it.
Fixes #48731
Change-Id: I7262cecd8ea7ae6fdf932f3a36db55fb062a1f2a
Reviewed-on: https://go-review.googlesource.com/c/go/+/355691
Trust: Katie Hockman <katie@golang.org>
Run-TryBot: Katie Hockman <katie@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Roland Shoemaker <roland@golang.org>
Keith Randall [Wed, 15 Sep 2021 16:56:09 +0000 (09:56 -0700)]
reflect: make Elem panic on bad notinheap pointers
This CL fixes the subtle issue that Elem can promote a
not-in-heap pointer, which could be any bit pattern, into an
unsafe.Pointer, which the garbage collector can see. If that
resulting value is bad, it can crash the GC.
Make sure that we don't introduce bad pointers that way. We can
make Elem() panic, because any such bad pointers are in the Go heap,
and not-in-heap pointers are not allowed to point into the Go heap.
Update #48399
Change-Id: Ieaf35a611b16b4dfb5e907e229ed4a2aed30e18c
Reviewed-on: https://go-review.googlesource.com/c/go/+/350153
Trust: Keith Randall <khr@golang.org>
Trust: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Robert Griesemer [Thu, 14 Oct 2021 01:50:06 +0000 (18:50 -0700)]
cmd/compile/internal/types2: add debugging support for delayed actions
Add a simple mechanism to provide formatted descriptions for
delayed actions. The comment strings are printed when tracing
is enabled and the delayed action is executed. This results
in more easily decipherable tracing output.
Requires debug mode in order to minimize the overhead during normal
execution. Use the mechanism in a few places to show typical use.
Also cleaned up a few unrelated comments.
Change-Id: Ic273c380c3963341500396ec62b694d143c25de2
Reviewed-on: https://go-review.googlesource.com/c/go/+/355871
Trust: Robert Griesemer <gri@golang.org>
Run-TryBot: Robert Griesemer <gri@golang.org> Reviewed-by: Robert Findley <rfindley@google.com>
Dan Scales [Sun, 19 Sep 2021 16:13:47 +0000 (09:13 -0700)]
cmd/compile: support new fully-inst types referenced during inlining
Modify the phase for creating needed function/method instantiations and
modifying functions to use those instantiations, so that the phase is
self-contained and can be called again after inlining. This is to deal
with the issue that inlining may reveal new fully-instantiated types
whose methods must be instantiated.
With this change, we have an extra phase for instantiation after
inlining, to take care of the new fully-instantiated types that have
shown up during inlining. We call inline.InlineCalls() for any new
instantiated functions that are created.
Change-Id: I4ddf0b1907e5f1f7d45891db7876455a99381133
Reviewed-on: https://go-review.googlesource.com/c/go/+/352870
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
Trust: Alexander Rakoczy <alex@golang.org>
Jay Conrod [Thu, 14 Oct 2021 23:47:47 +0000 (16:47 -0700)]
cmd/go: don't stamp build or vcs info for GOROOT binaries
Fixes a test failure in cmd/go TestScript/mod_outside.
make.bash (cmd/dist) builds everything with -gcflags=all= -ldflags=all=
by default. If those no-op flags aren't used, all GOROOT binaries
appear stale.
It's likely safe to omit those flags in cmd/dist if they're
empty. Checking out a new commit in GOROOT would always cause
staleness since the VCS info would change.
For #37475
Change-Id: Ic9aa0f3b7318e05fbb2f7d2c008ad07a4c61952f
Reviewed-on: https://go-review.googlesource.com/c/go/+/356014
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Bryan C. Mills <bcmills@google.com>
Cuong Manh Le [Wed, 13 Oct 2021 10:11:16 +0000 (17:11 +0700)]
cmd/compile: fix irgen reports wrong error message for misuse of //go:embed
Fixes #48230
Change-Id: Ic6490e065e7e79793faa0d0201dc94f5fcea694a
Reviewed-on: https://go-review.googlesource.com/c/go/+/355529
Trust: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Trust: Dan Scales <danscales@google.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> Reviewed-by: Dan Scales <danscales@google.com>
Robert Findley [Thu, 14 Oct 2021 15:34:00 +0000 (11:34 -0400)]
cmd/go: produce a better error for generic test functions
Test functions with type parameters cannot be invoked without type
arguments, so cmd/go would previously fail with a type-checking error in
the test harness.
Avoid this by producing an error explaining that test functions cannot
have type parameters.
Fixes #48953
Change-Id: I4e39c9b7a06c964fad5f12d74d262aa090c25d79
Reviewed-on: https://go-review.googlesource.com/c/go/+/355889
Trust: Robert Findley <rfindley@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com> Reviewed-by: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Jay Conrod [Tue, 12 Oct 2021 22:08:54 +0000 (15:08 -0700)]
cmd/go: stamp tags and flags in build info
Toolchain flags (like -gcflags), build tags (including race and msan),
and cgo variables (including CGO_ENABLED, CGO_CPPFLAGS and others) are
now stamped into binaries.
For #37475
Change-Id: I9023e682c0618f91805434946c6bc937536b69bb
Reviewed-on: https://go-review.googlesource.com/c/go/+/355493
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Bryan C. Mills <bcmills@google.com>
Jay Conrod [Fri, 1 Oct 2021 18:21:49 +0000 (11:21 -0700)]
cmd/go: stamp VCS revision and uncommitted status into binaries
When the go command builds a binary, it will now stamp the current
revision from the local Git or Mercurial repository, and it will also
stamp whether there are uncommitted edited or untracked files. Only
Git and Mercurial are supported for now.
If no repository is found containing the current working directory
(where the go command was started), or if either the main package
directory or the containing module's root directory is outside the
repository, no VCS information will be stamped. If the VCS tool is
missing or returns an error, that error is reported on the main
package (hinting that -buildvcs may be disabled).
This change introduces the -buildvcs flag, which is enabled by
default. When disabled, VCS information won't be stamped when it would
be otherwise.
Stamped information may be read using 'go version -m file' or
debug.ReadBuildInfo.
For #37475
Change-Id: I4e7d3159e1c270d85869ad99f10502e546e7582d
Reviewed-on: https://go-review.googlesource.com/c/go/+/353930
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Bryan C. Mills <bcmills@google.com>
Jay Conrod [Thu, 30 Sep 2021 17:25:49 +0000 (10:25 -0700)]
debug/buildinfo: new package with Read and ReadFile
These functions provide access to module information stamped into Go
binaries. In the future, they'll provide access to other information
(like VCS info).
These functions are added in a new package instead of runtime/debug
since they use binary parsing packages like debug/elf, which would
make runtime/debug an unacceptably heavy dependency. The types in
runtime/debug are still used; debug/buildinfo uses them via type
aliases.
This information is already available for the running binary through
debug.ReadBuildInfo and for other binaries with 'go version -m', but
until now, there hasn't been a way to get it for other binaries
without installing cmd/go.
This change copies most of the code in cmd/go/internal/version. A
later CL will migrate 'go version -m' to use this package.
For #37475
Fixes #39301
Change-Id: I0fbe0896e04f12ef81c6d79fb61b20daede86159
Reviewed-on: https://go-review.googlesource.com/c/go/+/353887
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com> Reviewed-by: Bryan C. Mills <bcmills@google.com>
Jay Conrod [Thu, 30 Sep 2021 21:37:30 +0000 (14:37 -0700)]
cmd/go: move module build info formatting into runtime/debug
Previously, modload.PackageBuildInfo returned a string containing
information about modules used to build an executable. This string is
embedded in the binary and can be read with debug.ReadBuildInfo or
'go version -m'.
With this change, debug.BuildInfo now has a MarshalText method that
returns a string in the same format as modload.PackageBuildInfo.
Package.load now calls Package.setBuildInfo, which constructs a
debug.BuildInfo, formats it with MarshalText, then sets
Package.Internal.BuildInfo. This is equivalent to what
modload.PackageBuildInfo did.
modload.PackageBuildInfo is deleted, since it's no longer used.
For #37475
Change-Id: I5875a98cb64737637fec2a450ab2ffa7f1805707
Reviewed-on: https://go-review.googlesource.com/c/go/+/353886
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com> Reviewed-by: Bryan C. Mills <bcmills@google.com>