]> Cypherpunks repositories - gostls13.git/log
gostls13.git
2 years agonet/http: regenerate h2_bundle.go
Dmitri Shuralyov [Fri, 18 Nov 2022 23:34:15 +0000 (18:34 -0500)]
net/http: regenerate h2_bundle.go

Done with:

go generate -run=bundle std

After CL 452096 updated the x/net version.

Change-Id: I1c1cd76d4ec9e14f45dc66c945c74e41ff689a30
Reviewed-on: https://go-review.googlesource.com/c/go/+/452195
Reviewed-by: Roland Shoemaker <roland@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Dmitri Shuralyov <dmitshur@golang.org>
Run-TryBot: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
2 years agocrypto/x509: implement SetFallbackRoots
Roland Shoemaker [Wed, 9 Nov 2022 22:04:10 +0000 (14:04 -0800)]
crypto/x509: implement SetFallbackRoots

Adds a method which allows users to set a fallback certificate pool for
usage during verification if the system certificate pool is empty.

Updates #43958

Change-Id: I279dd2f753743bce19790f2ae29f063c89c9359d
Reviewed-on: https://go-review.googlesource.com/c/go/+/449235
Run-TryBot: Roland Shoemaker <roland@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Roland Shoemaker <roland@golang.org>
Reviewed-by: Damien Neil <dneil@google.com>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
2 years agoall: update golang.org/x/crypto to 2c476679df9a
Filippo Valsorda [Fri, 18 Nov 2022 21:25:29 +0000 (22:25 +0100)]
all: update golang.org/x/crypto to 2c476679df9a

To pick up CL 451515.

This CL also updates x/net because x/crypto's dependency was bumped
while tagging v0.3.0.

Done by
        go get -d golang.org/x/crypto@2c476679df9a
        go mod tidy
        go mod vendor

Change-Id: I432a04586be3784b1027aa9b62d86c0df6d4a97e
Reviewed-on: https://go-review.googlesource.com/c/go/+/452096
Reviewed-by: Roland Shoemaker <roland@golang.org>
Auto-Submit: Filippo Valsorda <filippo@golang.org>
Reviewed-by: Damien Neil <dneil@google.com>
Run-TryBot: Filippo Valsorda <filippo@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/compile: package-annotate structs when error would be ambiguous
David Chase [Thu, 11 Aug 2022 13:58:23 +0000 (09:58 -0400)]
cmd/compile: package-annotate structs when error would be ambiguous

Before emitting a "wanted Foo but got Bar" message for an interface
type match failure, check that Foo and Bar are different.  If they
are not, add package paths to first unexported struct field seen,
because that is the cause (a cause, there could be more than one).

Replicated in go/types.

Added tests to go/types and cmd/compile/internal/types2

Fixes #54258.

Change-Id: Ifc2b2067d62fe2138996972cdf3b6cb7ca0ed456
Reviewed-on: https://go-review.googlesource.com/c/go/+/422914
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: David Chase <drchase@google.com>
Reviewed-by: Robert Griesemer <gri@google.com>
2 years agocmd/go: don't report non-go files in CompiledGoFiles
Michael Matloob [Fri, 18 Nov 2022 19:54:14 +0000 (14:54 -0500)]
cmd/go: don't report non-go files in CompiledGoFiles

We save non-go files in the cached srcfiles file because we want the
non-go files for vet, but we shouldn't report them in CompiledGoFiles.
Filter them out before adding them to CompiledGoFiles.

Fixes #28749

Change-Id: I889d4bbf8c4ec1348584a62ef5e4f8b3f05e97da
Reviewed-on: https://go-review.googlesource.com/c/go/+/451285
Run-TryBot: Michael Matloob <matloob@golang.org>
Reviewed-by: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
2 years agocmd/go/internal/script: check lack of error for non-waiting cmds
Michael Matloob [Fri, 18 Nov 2022 19:33:23 +0000 (14:33 -0500)]
cmd/go/internal/script: check lack of error for non-waiting cmds

In the script engine, if a command does not return a Wait function and
it succeeds, we won't call checkStatus. That means that commands that
don't have a wait function, have a "!" indicating that they are
supposed to fail, and then succeed will spuriously not fail the script
engine test even they were supposed to fail but didn't.

Change-Id: Ic88c3cdd628064d48f14a8a4a2e97cded48890fb
Reviewed-on: https://go-review.googlesource.com/c/go/+/451284
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Michael Matloob <matloob@golang.org>

2 years agodoc/go1.20: add release notes for net package
Damien Neil [Fri, 18 Nov 2022 18:57:46 +0000 (10:57 -0800)]
doc/go1.20: add release notes for net package

For #50101
For #51152
For #53482
For #55301
For #56515

Change-Id: I11edeb4be0a7f80fb72fd7680a3407d081f83b8b
Reviewed-on: https://go-review.googlesource.com/c/go/+/451420
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Damien Neil <dneil@google.com>
Run-TryBot: Damien Neil <dneil@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
2 years agonet: fix typo in ControlContext parameter names
Damien Neil [Fri, 18 Nov 2022 18:44:55 +0000 (10:44 -0800)]
net: fix typo in ControlContext parameter names

Change-Id: I35fcfb2d8cafadca36cffeebe0858973895946d7
Reviewed-on: https://go-review.googlesource.com/c/go/+/451419
Reviewed-by: Ian Lance Taylor <iant@google.com>
Auto-Submit: Damien Neil <dneil@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Damien Neil <dneil@google.com>

2 years agocrypto/tls: add CertificateVerificationError to tls handshake
Gabor Tanz [Fri, 18 Nov 2022 07:59:03 +0000 (07:59 +0000)]
crypto/tls: add CertificateVerificationError to tls handshake

Fixes #48152

Change-Id: I503f088edeb5574fd5eb5905bff7c3c23b2bc8fc
GitHub-Last-Rev: 2b0e982f3f6bca33062b0bbd64ed1804801e2c13
GitHub-Pull-Request: golang/go#56686
Reviewed-on: https://go-review.googlesource.com/c/go/+/449336
Run-TryBot: Roland Shoemaker <roland@golang.org>
Auto-Submit: Roland Shoemaker <roland@golang.org>
Reviewed-by: Damien Neil <dneil@google.com>
Reviewed-by: Julie Qiu <julieqiu@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
2 years agocmd/go: replace 'directory .' with 'current directory' in some errors
Michael Matloob [Wed, 16 Nov 2022 18:28:32 +0000 (13:28 -0500)]
cmd/go: replace 'directory .' with  'current directory' in some errors

To make the error clearer

Fixes #56697

Change-Id: Idfb5e8704d1bfc64bd0a09d5b553086d9ba5ac33
Reviewed-on: https://go-review.googlesource.com/c/go/+/451295
Reviewed-by: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Bryan Mills <bcmills@google.com>
Reviewed-by: Joedian Reid <joedian@golang.org>
2 years agoall: add missing periods in comments
cui fliter [Fri, 11 Nov 2022 11:22:35 +0000 (19:22 +0800)]
all: add missing periods in comments

Change-Id: I69065f8adf101fdb28682c55997f503013a50e29
Reviewed-on: https://go-review.googlesource.com/c/go/+/449757
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Joedian Reid <joedian@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Joedian Reid <joedian@golang.org>
Run-TryBot: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
2 years agoruntime,cmd/link: increase stack guard space when building with -race
Keith Randall [Wed, 16 Nov 2022 20:56:40 +0000 (12:56 -0800)]
runtime,cmd/link: increase stack guard space when building with -race

More stuff to do = more stack needed. Bump up the guard space when
building with the race detector.

Fixes #54291

Change-Id: I701bc8800507921bed568047d35b8f49c26e7df7
Reviewed-on: https://go-review.googlesource.com/c/go/+/451217
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
2 years agoruntime: optimise memmove on riscv64
Joel Sing [Sat, 27 Aug 2022 19:35:31 +0000 (05:35 +1000)]
runtime: optimise memmove on riscv64

Implement a more optimised memmove on riscv64, where up to 64 bytes are moved
per loop after achieving alignment. In the unaligned case, memory is moved at
up to 8 bytes per loop.

This also avoids doing unaligned loads and stores, which results in kernel
traps and a significant performance penality.

Fixes #48248.

name                               old speed      new speed        delta
Memmove/1-4                        31.3MB/s _ 0%    26.6MB/s _ 0%    -14.95%  (p=0.000 n=3+3)
Memmove/2-4                        50.6MB/s _ 1%    42.6MB/s _ 0%    -15.75%  (p=0.000 n=3+3)
Memmove/3-4                        64.5MB/s _ 1%    53.4MB/s _ 2%    -17.11%  (p=0.001 n=3+3)
Memmove/4-4                        74.9MB/s _ 0%    99.2MB/s _ 0%    +32.55%  (p=0.000 n=3+3)
Memmove/5-4                        82.3MB/s _ 0%    99.0MB/s _ 1%    +20.29%  (p=0.000 n=3+3)
Memmove/6-4                        88.2MB/s _ 0%   102.3MB/s _ 1%    +15.87%  (p=0.000 n=3+3)
Memmove/7-4                        93.4MB/s _ 0%   102.0MB/s _ 0%     +9.18%  (p=0.000 n=3+3)
Memmove/8-4                         188MB/s _ 3%     188MB/s _ 6%       ~     (p=0.964 n=3+3)
Memmove/9-4                         182MB/s _ 6%     163MB/s _ 1%       ~     (p=0.069 n=3+3)
Memmove/10-4                        177MB/s _ 0%     149MB/s _ 4%    -15.93%  (p=0.012 n=3+3)
Memmove/11-4                        171MB/s _ 6%     148MB/s _ 0%    -13.65%  (p=0.045 n=3+3)
Memmove/12-4                        166MB/s _ 5%     209MB/s _ 0%    +26.12%  (p=0.009 n=3+3)
Memmove/13-4                        170MB/s _ 1%     188MB/s _ 4%    +10.76%  (p=0.039 n=3+3)
Memmove/14-4                        158MB/s _ 0%     185MB/s _ 0%    +17.13%  (p=0.000 n=3+3)
Memmove/15-4                        166MB/s _ 0%     175MB/s _ 0%     +5.38%  (p=0.000 n=3+3)
Memmove/16-4                        320MB/s _ 6%     343MB/s _ 0%       ~     (p=0.149 n=3+3)
Memmove/32-4                        493MB/s _ 5%     628MB/s _ 1%    +27.51%  (p=0.008 n=3+3)
Memmove/64-4                        706MB/s _ 0%    1132MB/s _ 0%    +60.32%  (p=0.000 n=3+3)
Memmove/128-4                       837MB/s _ 1%    1623MB/s _ 1%    +93.96%  (p=0.000 n=3+3)
Memmove/256-4                       960MB/s _ 0%    2070MB/s _ 6%   +115.68%  (p=0.003 n=3+3)
Memmove/512-4                      1.04GB/s _ 0%    2.55GB/s _ 0%   +146.05%  (p=0.000 n=3+3)
Memmove/1024-4                     1.08GB/s _ 0%    2.76GB/s _ 0%   +155.62%  (p=0.000 n=3+3)
Memmove/2048-4                     1.10GB/s _ 0%    2.90GB/s _ 1%   +164.31%  (p=0.000 n=3+3)
Memmove/4096-4                     1.11GB/s _ 0%    2.98GB/s _ 0%   +169.77%  (p=0.000 n=3+3)
MemmoveOverlap/32-4                 443MB/s _ 0%     500MB/s _ 0%    +12.81%  (p=0.000 n=3+3)
MemmoveOverlap/64-4                 635MB/s _ 0%     908MB/s _ 0%    +42.92%  (p=0.000 n=3+3)
MemmoveOverlap/128-4                789MB/s _ 0%    1423MB/s _ 0%    +80.28%  (p=0.000 n=3+3)
MemmoveOverlap/256-4                925MB/s _ 0%    1941MB/s _ 0%   +109.86%  (p=0.000 n=3+3)
MemmoveOverlap/512-4               1.01GB/s _ 2%    2.37GB/s _ 0%   +134.86%  (p=0.000 n=3+3)
MemmoveOverlap/1024-4              1.06GB/s _ 0%    2.68GB/s _ 1%   +151.67%  (p=0.000 n=3+3)
MemmoveOverlap/2048-4              1.09GB/s _ 0%    2.89GB/s _ 0%   +164.82%  (p=0.000 n=3+3)
MemmoveOverlap/4096-4              1.11GB/s _ 0%    3.01GB/s _ 0%   +171.30%  (p=0.000 n=3+3)
MemmoveUnalignedDst/1-4            24.1MB/s _ 1%    21.3MB/s _ 0%    -11.76%  (p=0.000 n=3+3)
MemmoveUnalignedDst/2-4            41.6MB/s _ 1%    35.9MB/s _ 0%    -13.72%  (p=0.000 n=3+3)
MemmoveUnalignedDst/3-4            54.0MB/s _ 0%    45.5MB/s _ 2%    -15.76%  (p=0.004 n=3+3)
MemmoveUnalignedDst/4-4            63.9MB/s _ 1%    81.6MB/s _ 0%    +27.70%  (p=0.000 n=3+3)
MemmoveUnalignedDst/5-4            69.4MB/s _ 6%    84.8MB/s _ 0%    +22.08%  (p=0.015 n=3+3)
MemmoveUnalignedDst/6-4            77.8MB/s _ 2%    89.0MB/s _ 0%    +14.53%  (p=0.004 n=3+3)
MemmoveUnalignedDst/7-4            83.0MB/s _ 0%    90.7MB/s _ 1%     +9.30%  (p=0.000 n=3+3)
MemmoveUnalignedDst/8-4            6.97MB/s _ 2%  127.73MB/s _ 0%  +1732.57%  (p=0.000 n=3+3)
MemmoveUnalignedDst/9-4            7.81MB/s _ 1%  125.41MB/s _ 0%  +1506.45%  (p=0.000 n=3+3)
MemmoveUnalignedDst/10-4           8.59MB/s _ 2%  123.52MB/s _ 0%  +1337.43%  (p=0.000 n=3+3)
MemmoveUnalignedDst/11-4           9.23MB/s _ 6%  119.81MB/s _ 4%  +1197.55%  (p=0.000 n=3+3)
MemmoveUnalignedDst/12-4           10.3MB/s _ 0%   155.9MB/s _ 7%  +1416.08%  (p=0.001 n=3+3)
MemmoveUnalignedDst/13-4           10.9MB/s _ 3%   155.1MB/s _ 0%  +1321.26%  (p=0.000 n=3+3)
MemmoveUnalignedDst/14-4           11.4MB/s _ 5%   151.0MB/s _ 0%  +1229.37%  (p=0.000 n=3+3)
MemmoveUnalignedDst/15-4           12.6MB/s _ 0%   147.0MB/s _ 0%  +1066.39%  (p=0.000 n=3+3)
MemmoveUnalignedDst/16-4           7.17MB/s _ 0%  184.33MB/s _ 5%  +2470.90%  (p=0.001 n=3+3)
MemmoveUnalignedDst/32-4           7.26MB/s _ 0%  252.00MB/s _ 2%  +3371.12%  (p=0.000 n=3+3)
MemmoveUnalignedDst/64-4           7.25MB/s _ 2%  306.37MB/s _ 1%  +4125.75%  (p=0.000 n=3+3)
MemmoveUnalignedDst/128-4          7.32MB/s _ 1%  338.03MB/s _ 1%  +4517.85%  (p=0.000 n=3+3)
MemmoveUnalignedDst/256-4          7.31MB/s _ 0%  361.06MB/s _ 0%  +4841.47%  (p=0.000 n=3+3)
MemmoveUnalignedDst/512-4          7.35MB/s _ 0%  373.55MB/s _ 0%  +4982.36%  (p=0.000 n=3+3)
MemmoveUnalignedDst/1024-4         7.33MB/s _ 0%  379.00MB/s _ 2%  +5068.18%  (p=0.000 n=3+3)
MemmoveUnalignedDst/2048-4         7.31MB/s _ 2%  383.05MB/s _ 0%  +5142.47%  (p=0.000 n=3+3)
MemmoveUnalignedDst/4096-4         7.35MB/s _ 1%  385.97MB/s _ 1%  +5151.25%  (p=0.000 n=3+3)
MemmoveUnalignedDstOverlap/32-4    9.43MB/s _ 0%  233.72MB/s _ 0%  +2377.56%  (p=0.000 n=3+3)
MemmoveUnalignedDstOverlap/64-4    8.13MB/s _ 3%  288.77MB/s _ 0%  +3451.91%  (p=0.000 n=3+3)
MemmoveUnalignedDstOverlap/128-4   7.77MB/s _ 0%  326.62MB/s _ 3%  +4103.65%  (p=0.000 n=3+3)
MemmoveUnalignedDstOverlap/256-4   7.28MB/s _ 6%  357.24MB/s _ 0%  +4804.85%  (p=0.000 n=3+3)
MemmoveUnalignedDstOverlap/512-4   7.44MB/s _ 0%  363.63MB/s _ 7%  +4787.54%  (p=0.001 n=3+3)
MemmoveUnalignedDstOverlap/1024-4  7.37MB/s _ 0%  383.17MB/s _ 0%  +5101.40%  (p=0.000 n=3+3)
MemmoveUnalignedDstOverlap/2048-4  7.29MB/s _ 2%  387.69MB/s _ 0%  +5215.68%  (p=0.000 n=3+3)
MemmoveUnalignedDstOverlap/4096-4  7.18MB/s _ 5%  389.22MB/s _ 0%  +5320.84%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/1-4            24.2MB/s _ 0%    21.4MB/s _ 1%    -11.70%  (p=0.001 n=3+3)
MemmoveUnalignedSrc/2-4            41.7MB/s _ 0%    36.0MB/s _ 0%    -13.71%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/3-4            52.1MB/s _ 6%    46.4MB/s _ 1%       ~     (p=0.074 n=3+3)
MemmoveUnalignedSrc/4-4            60.4MB/s _ 0%    76.4MB/s _ 0%    +26.39%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/5-4            71.2MB/s _ 1%    84.7MB/s _ 0%    +18.90%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/6-4            77.7MB/s _ 0%    88.7MB/s _ 0%    +14.06%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/7-4            82.9MB/s _ 1%    90.7MB/s _ 1%     +9.42%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/8-4            74.6MB/s _ 0%   120.6MB/s _ 0%    +61.62%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/9-4            78.7MB/s _ 1%   123.9MB/s _ 1%    +57.42%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/10-4           82.1MB/s _ 0%   121.7MB/s _ 0%    +48.21%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/11-4           83.7MB/s _ 5%   122.0MB/s _ 0%    +45.79%  (p=0.003 n=3+3)
MemmoveUnalignedSrc/12-4           88.6MB/s _ 0%   160.8MB/s _ 0%    +81.56%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/13-4           91.0MB/s _ 0%   155.0MB/s _ 0%    +70.29%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/14-4           92.0MB/s _ 2%   151.0MB/s _ 0%    +64.09%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/15-4           12.6MB/s _ 0%   146.6MB/s _ 0%  +1063.32%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/16-4           13.3MB/s _ 0%   188.8MB/s _ 2%  +1319.02%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/32-4           9.44MB/s _ 0%  254.24MB/s _ 1%  +2594.21%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/64-4           8.27MB/s _ 0%  302.33MB/s _ 2%  +3555.78%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/128-4          7.73MB/s _ 3%  338.82MB/s _ 0%  +4281.29%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/256-4          7.58MB/s _ 0%  362.19MB/s _ 0%  +4678.23%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/512-4          7.44MB/s _ 1%  374.49MB/s _ 0%  +4933.51%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/1024-4         7.30MB/s _ 2%  379.74MB/s _ 0%  +5099.54%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/2048-4         7.34MB/s _ 2%  385.50MB/s _ 0%  +5154.38%  (p=0.000 n=3+3)
MemmoveUnalignedSrc/4096-4         7.35MB/s _ 1%  383.64MB/s _ 0%  +5119.59%  (p=0.000 n=3+3)
MemmoveUnalignedSrcOverlap/32-4    7.22MB/s _ 0%  254.94MB/s _ 0%  +3432.66%  (p=0.000 n=3+3)
MemmoveUnalignedSrcOverlap/64-4    7.29MB/s _ 1%  296.99MB/s _ 5%  +3973.89%  (p=0.001 n=3+3)
MemmoveUnalignedSrcOverlap/128-4   7.32MB/s _ 1%  336.73MB/s _ 1%  +4500.09%  (p=0.000 n=3+3)
MemmoveUnalignedSrcOverlap/256-4   7.30MB/s _ 1%  361.41MB/s _ 0%  +4850.82%  (p=0.000 n=3+3)
MemmoveUnalignedSrcOverlap/512-4   7.34MB/s _ 0%  374.92MB/s _ 0%  +5007.90%  (p=0.000 n=3+3)
MemmoveUnalignedSrcOverlap/1024-4  7.34MB/s _ 0%  380.15MB/s _ 0%  +5079.16%  (p=0.000 n=3+3)
MemmoveUnalignedSrcOverlap/2048-4  7.36MB/s _ 0%  383.78MB/s _ 0%  +5116.76%  (p=0.000 n=3+3)
MemmoveUnalignedSrcOverlap/4096-4  7.35MB/s _ 0%  386.32MB/s _ 0%  +5156.05%  (p=0.000 n=3+3)

Change-Id: Ibc13230af7b1e205ed95a6470e2cf64ff4251405
Reviewed-on: https://go-review.googlesource.com/c/go/+/426256
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Meng Zhuo <mzh@golangcn.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Reviewed-by: Joedian Reid <joedian@golang.org>
Run-TryBot: Joel Sing <joel@sing.id.au>

2 years agoio/fs: clean up test helper functions
Tobias Klauser [Thu, 17 Nov 2022 16:38:20 +0000 (17:38 +0100)]
io/fs: clean up test helper functions

Inline the only use of checkMarks which also allows to drop the
always-true report argument. This also ensures the correct line gets
reported in case of an error.

Also remove the unused markTree function and drop the unused testing.T
argument from makeTree.

Change-Id: I4033d3e5ecd929d08ce03c563aa99444e102d931
Reviewed-on: https://go-review.googlesource.com/c/go/+/451615
Reviewed-by: Joedian Reid <joedian@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Auto-Submit: Tobias Klauser <tobias.klauser@gmail.com>
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/link/internal/ppc64: fix trampoline reuse distance calculation
Paul E. Murphy [Wed, 16 Nov 2022 20:53:39 +0000 (14:53 -0600)]
cmd/link/internal/ppc64: fix trampoline reuse distance calculation

If a compatible trampoline has been inserted by a previously laid
function in the same section, and is known to be sufficiently close,
it can be reused.

When testing if the trampoline can be reused, the addend of the direct
call should be ignored. It is already encoded in the trampoline. If the
addend is non-zero, and the target sufficiently far away, and just
beyond direct call reach, this may cause the trampoline to be
incorrectly reused.

This was observed on go1.17.13 and openshift-installer commit f3c53b382
building in release mode with the following error:

github.com/aliyun/alibaba-cloud-sdk-go/services/cms.(*Client).DescribeMonitoringAgentAccessKeyWithChan.func1: direct call too far: runtime.duffzero+1f0-tramp0-1 -2000078

Fixes #56775

Change-Id: I54af957302506d4e3cd5d3121542c83fe980e912
Reviewed-on: https://go-review.googlesource.com/c/go/+/451415
Reviewed-by: Cherry Mui <cherryyz@google.com>
Run-TryBot: Paul Murphy <murp@ibm.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Lynn Boger <laboger@linux.vnet.ibm.com>
Reviewed-by: Than McIntosh <thanm@google.com>
2 years agocmd/compile/internal/base, cmd/internal/bio: use syscall.Mmap on aix
Tobias Klauser [Mon, 14 Nov 2022 09:00:57 +0000 (10:00 +0100)]
cmd/compile/internal/base, cmd/internal/bio: use syscall.Mmap on aix

Change-Id: Ic28612952eb9abf14425f0bb14043b10f6050d94
Reviewed-on: https://go-review.googlesource.com/c/go/+/450195
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
Reviewed-by: Joedian Reid <joedian@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Tobias Klauser <tobias.klauser@gmail.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
2 years agocmd/compile: fix wrong optimization for eliding Not in Phi
Wayne Zuo [Thu, 17 Nov 2022 10:10:00 +0000 (18:10 +0800)]
cmd/compile: fix wrong optimization for eliding Not in Phi

The previous rule may move the phi value into a wrong block.
This CL make it only rewrite the phi value not the If block,
so that the phi value will stay in old block.

Fixes #56777

Change-Id: I9479a5c7f28529786968413d35b82a16181bb1f1
Reviewed-on: https://go-review.googlesource.com/c/go/+/451496
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Wayne Zuo <wdvxdr@golangcn.org>
Reviewed-by: Keith Randall <khr@google.com>
Reviewed-by: David Chase <drchase@google.com>
2 years agocmd/internal/obj/arm64: tidy literal pool
eric fang [Mon, 8 Aug 2022 07:42:43 +0000 (07:42 +0000)]
cmd/internal/obj/arm64: tidy literal pool

This CL cleans up the literal pool implementation and inserts an UNDEF
instruction before the literal pool if the last instruction of the
function is not an unconditional jump instruction, RET or ERET
instruction.

Change-Id: Ifecb9e3372478362dde246c1bc9bc8d527a469d5
Reviewed-on: https://go-review.googlesource.com/c/go/+/424134
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Joedian Reid <joedian@golang.org>
Run-TryBot: Eric Fang <eric.fang@arm.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agoruntime: add page tracer
Michael Knyszek [Wed, 19 Oct 2022 18:51:15 +0000 (14:51 -0400)]
runtime: add page tracer

This change adds a new GODEBUG flag called pagetrace that writes a
low-overhead trace of how pages of memory are managed by the Go runtime.

The page tracer is kept behind a GOEXPERIMENT flag due to a potential
security risk for setuid binaries.

Change-Id: I6f4a2447d02693c25214400846a5d2832ad6e5c0
Reviewed-on: https://go-review.googlesource.com/c/go/+/444157
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: David Chase <drchase@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/internal/obj/arm64: mark branch instructions in optab
eric fang [Tue, 9 Aug 2022 06:10:16 +0000 (06:10 +0000)]
cmd/internal/obj/arm64: mark branch instructions in optab

Currently, we judge whether we need to fix up the branch instruction
based on Optab.type_ field, but the type_ field in optab may change.
This CL marks the branch instruction in optab, and checks whether to
do fixing up according to the mark. Depending on the constant parameter
range of the branch instruction, there are two labels, BRANCH14BITS,
BRANCH19BITS. For the 26-bit branch, linker will handle it.

Besides this CL removes the unnecessary alignment of the DWORD
instruction. Because the ISA doesn't require it and no 64-bit load
assume it. The only effect is that there is some performance penalty
for loading from DWORDs if the 8-byte DWORD instruction crosses the
cache line, but this is very rare.

Change-Id: I993902b3fb5ad8e081dd6c441e86bcf581031835
Reviewed-on: https://go-review.googlesource.com/c/go/+/424135
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Eric Fang <eric.fang@arm.com>
Reviewed-by: Joedian Reid <joedian@golang.org>
2 years agogo/types, types2: ensure signatures are instantiated if all type args
Robert Findley [Thu, 17 Nov 2022 01:58:58 +0000 (20:58 -0500)]
go/types, types2: ensure signatures are instantiated if all type args
are provided

Improve the accuracy of recorded types and instances for function calls,
by instantiating their signature before checking arguments if all type
arguments are provided. This avoids a problem where fully instantiated
function signatures are are not recorded as such following an error
checking their arguments.

Fixes golang/go#51803

Change-Id: Iec4cbd219a2cd19bb1bcf2a5c4019f556e4304b1
Reviewed-on: https://go-review.googlesource.com/c/go/+/451436
Reviewed-by: Robert Griesemer <gri@google.com>
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agogo/types, types2: replace some Errorf calls with Error calls (cleanup)
Robert Griesemer [Fri, 18 Nov 2022 00:09:15 +0000 (16:09 -0800)]
go/types, types2: replace some Errorf calls with Error calls (cleanup)

Change-Id: I9b6759a82b8009b323132c78cb7d78c2c35652bd
Reviewed-on: https://go-review.googlesource.com/c/go/+/451815
Reviewed-by: Robert Griesemer <gri@google.com>
Reviewed-by: Robert Findley <rfindley@google.com>
Auto-Submit: Robert Griesemer <gri@google.com>
Run-TryBot: Robert Griesemer <gri@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agogo/types, types2: replace (internal) writePackage with packagePrefix
Robert Griesemer [Thu, 17 Nov 2022 22:25:50 +0000 (14:25 -0800)]
go/types, types2: replace (internal) writePackage with packagePrefix

This makes it easier to use the package string prefix in some cases
(cleanup).

Change-Id: I0ae74bf8770999110e7d6e49eac4e42962e78596
Reviewed-on: https://go-review.googlesource.com/c/go/+/451795
Auto-Submit: Robert Griesemer <gri@google.com>
Run-TryBot: Robert Griesemer <gri@google.com>
Reviewed-by: Robert Griesemer <gri@google.com>
Reviewed-by: David Chase <drchase@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocrypto/ecdsa: improve benchmarks
Filippo Valsorda [Mon, 14 Nov 2022 17:43:43 +0000 (18:43 +0100)]
crypto/ecdsa: improve benchmarks

While at it, drop P-224 benchmarks. Nobody cares about P-224.

Change-Id: I31db6fedde6026deff36de963690275dacf5fda1
Reviewed-on: https://go-review.googlesource.com/c/go/+/451196
Auto-Submit: Filippo Valsorda <filippo@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: Joedian Reid <joedian@golang.org>
Run-TryBot: Filippo Valsorda <filippo@golang.org>

2 years agoruntime: fix conflict between lfstack and checkptr
Keith Randall [Wed, 9 Nov 2022 01:48:48 +0000 (17:48 -0800)]
runtime: fix conflict between lfstack and checkptr

lfstack does very unsafe things. In particular, it will not
work with nodes that live on the heap. In normal use by the runtime,
that is the case (it is only used for gc work bufs). But the lfstack
test does use heap objects. It goes through some hoops to prevent
premature deallocation, but those hoops are not enough to convince
-d=checkptr that everything is ok.

Instead, allocate the test objects outside the heap, like the runtime
does for all of its lfstack usage. Remove the lifetime workaround
from the test.

Reported in https://groups.google.com/g/golang-nuts/c/psjrUV2ZKyI

Change-Id: If611105eab6c823a4d6c105938ce145ed731781d
Reviewed-on: https://go-review.googlesource.com/c/go/+/448899
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Keith Randall <khr@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Keith Randall <khr@golang.org>

2 years agonet: return errNoSuchHost when no entry found in /etc/hosts and order is hostLookupFiles
Mateusz Poliwczak [Thu, 17 Nov 2022 08:49:32 +0000 (08:49 +0000)]
net: return errNoSuchHost when no entry found in /etc/hosts and order is hostLookupFiles

When /etc/nsswitch.conf lists: "hosts: files" then LookupHost returns two nils when no entry inside /etc/hosts is found.

Change-Id: I96d68a079dfe009655c84cf0e697ce19a5bb6698
GitHub-Last-Rev: 894f066bbcc7c975f1975bd0d1dcb5726f590bc5
GitHub-Pull-Request: golang/go#56747
Reviewed-on: https://go-review.googlesource.com/c/go/+/450875
Run-TryBot: Ian Lance Taylor <iant@google.com>
Reviewed-by: Joedian Reid <joedian@golang.org>
Auto-Submit: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
2 years agoruntime: work around Apple libc bugs to make exec stop hanging
Russ Cox [Thu, 17 Nov 2022 19:24:06 +0000 (14:24 -0500)]
runtime: work around Apple libc bugs to make exec stop hanging

For a while now, we've had intermittent reports about problems with
os/exec on macOS, but no clear way to reproduce them. Recent changes
in the os/exec package test seem to have aligned the stars just right,
at least on my two x86 and ARM MacBook Pro laptops, to make the
package test hang with roughly 50% probability. When it does hang, the
stacks I see in the hung process match the ones reported for the
Go-based hangs in #33565. (They do not match the ones reported in the
so-called C reproducer in that issue, but I think that reproducer is
actually reproducing a different race, between fork and exit.)

The stacks obtained from the hung child processes are in
libSystem_atfork_child, which is supposed to reinitialize various
parts of the C library in the new process.

One common stack dies in _notify_fork_child calling _notify_globals
(inlined) calling _os_alloc_once, because _os_alloc_once detects that
the once lock is held by the parent process and then calls
_os_once_gate_corruption_abort. The allocation is setting up the
globals for the notification subsystem. See the source code at [1].
To work around this, we can allocate the globals earlier in the Go
program's lifetime, before any execs are involved, by calling any
notify routine that is exported, calls _notify_globals, and doesn't do
anything too expensive otherwise. notify_is_valid_token(0) fits the bill.

The other common stack dies in xpc_atfork_child calling
_objc_msgSend_uncached which ends up in
WAITING_FOR_ANOTHER_THREAD_TO_FINISH_CALLING_+initialize. Of course,
whatever thread the child is waiting for is in the parent process and
is not going to finish anything in the child process. There is no
public source code for these routines, so it is unclear exactly what
the problem is. However, xpc_atfork_child turns out to be exported
(for use by libSystem_atfork_child, which is in a different library,
so xpc_atfork_child is unlikely to be unexported any time soon).
It also stands to reason that since xpc_atfork_child is called at the
start of any forked child process, it can't be too harmful to call at
the start of an ordinary Go process. And whatever caches it needs for
a non-deadlocking fast path during exec empirically do get initialized
by calling it at startup.

This CL introduces a function osinit_hack, called at osinit time,
which calls notify_is_valid_token(0) and xpc_atfork_child().
Doing so makes the os/exec test pass reliably on both my laptops -
I can run it successfully hundreds of times in a row when my previous
record was twice in a row.

Fixes #33565.
Fixes #56784.

[1] https://opensource.apple.com/source/Libnotify/Libnotify-241/notify_client.c.auto.html

Change-Id: I16a14a800893c40244678203532a3e8d6214b6bd
Reviewed-on: https://go-review.googlesource.com/c/go/+/451735
Run-TryBot: Russ Cox <rsc@golang.org>
Auto-Submit: Russ Cox <rsc@golang.org>
Reviewed-by: Cherry Mui <cherryyz@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/compile/internal/pgo: count only the last two frames as a call edge
Cherry Mui [Tue, 15 Nov 2022 18:37:42 +0000 (13:37 -0500)]
cmd/compile/internal/pgo: count only the last two frames as a call edge

Currently for every CPU profile sample, we apply its weight to all
call edges of the entire call stack. Frames higher up the stack
are unlikely to be repeated calls (e.g. runtime.main calling
main.main). So adding weights to call edges higher up the stack
may be not reflecting the actual call edge weights in the program.
This CL changes it to add weights to only the edge between the
last two frames.

Without a branch profile (e.g. LBR records) this is not perfect,
but seems more reasonable.

Change-Id: I0aee75cc608a152adad41c51120b661a6c542283
Reviewed-on: https://go-review.googlesource.com/c/go/+/450915
Run-TryBot: Cherry Mui <cherryyz@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
2 years agocmd/compile: simplify PGO hot caller/callee computation
Cherry Mui [Tue, 15 Nov 2022 18:32:32 +0000 (13:32 -0500)]
cmd/compile: simplify PGO hot caller/callee computation

Currently, we use CDF to compute a weight threshold and then use
the weight threshold to determine whether a call site is hot. As
when we compute the CDF we already have a list of hot call sites
that make up the given percentage of the CDF, just use that list.

Also, when computing the CDF threshold, include the very last node
that makes it to go over the threshold. (I.e. if the CDF threshold
is 50% and one hot node takes 60% of weight, we should include that
node instead of excluding it. In practice it rarely matters,
probably only for testing and micro-benchmarks.)

Change-Id: I535ae9cd6b679609e247c3d0d9ee572c1a1187cc
Reviewed-on: https://go-review.googlesource.com/c/go/+/450737
Reviewed-by: Austin Clements <austin@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Cherry Mui <cherryyz@google.com>

2 years agocmd/compile: fix broken IR for iface -> eface
Cuong Manh Le [Wed, 16 Nov 2022 18:02:26 +0000 (01:02 +0700)]
cmd/compile: fix broken IR for iface -> eface

For implementing interface to empty interface conversion, the compiler
generate code like:

var res *uint8
res = itab
if res != nil {
res = res.type
}

However, itab has type *uintptr, so the assignment is broken. The
problem is not shown up, until CL 450215, which call typecheck on this
broken assignment.

To fix this, just cast itab to *uint8 when doing the conversion.

Fixes #56768

Change-Id: Id42792d18e7f382578b40854d46eecd49673792c
Reviewed-on: https://go-review.googlesource.com/c/go/+/451256
Reviewed-by: Keith Randall <khr@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Auto-Submit: Cuong Manh Le <cuong.manhle.vn@gmail.com>

2 years agosync/atomic: hint users of old API to use new type-based API instead
Keith Randall [Wed, 9 Nov 2022 23:28:44 +0000 (15:28 -0800)]
sync/atomic: hint users of old API to use new type-based API instead

Fixes #56495

Change-Id: Ib2f39273da68e3056688306aa0d5e274b5507bf4
Reviewed-on: https://go-review.googlesource.com/c/go/+/449237
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
Reviewed-by: Sean Liao <sean@liao.dev>
Reviewed-by: Ian Lance Taylor <iant@google.com>
2 years agonet/http: deflake TestResponseControllerSetFutureWriteDeadline
Damien Neil [Thu, 17 Nov 2022 19:09:11 +0000 (11:09 -0800)]
net/http: deflake TestResponseControllerSetFutureWriteDeadline

Don't set the server's write deadline until after the client has
read the response headers, avoiding test failures if the deadline
expires before or while writing headers.

Fixes #56807.

Change-Id: I5f80c108b360d030132a13661774a30fac453856
Reviewed-on: https://go-review.googlesource.com/c/go/+/451715
Auto-Submit: Damien Neil <dneil@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Damien Neil <dneil@google.com>
Reviewed-by: Bryan Mills <bcmills@google.com>
2 years agocmd/compile: teach regalloc about temporary registers
Keith Randall [Tue, 5 Apr 2022 22:07:29 +0000 (15:07 -0700)]
cmd/compile: teach regalloc about temporary registers

Temporary registers are sometimes needed for an architecture backend
which needs to use several machine instructions to implement a single
SSA instruction.

Mark such instructions so that regalloc can reserve the temporary register
for it. That way we don't have to reserve a fixed register like we do now.

Convert the temp-register-using instructions on amd64 to use this
new mechanism. Other archs can follow as needed.

Change-Id: I1d0c8588afdad5cd18b4398eb5a0f755be5dead7
Reviewed-on: https://go-review.googlesource.com/c/go/+/398556
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: Cherry Mui <cherryyz@google.com>
Reviewed-by: David Chase <drchase@google.com>
2 years agocmd/compile: fix static init for inlined calls
Cuong Manh Le [Thu, 17 Nov 2022 12:21:45 +0000 (19:21 +0700)]
cmd/compile: fix static init for inlined calls

CL 450136 made the compiler to be able to handle simple inlined calls in
staticinit. However, it's missed a condition when checking substituting
arg for param. If there's any non-trivial closures, it has captured one
of the param, so the substitution could not happen.

Fixes #56778

Change-Id: I427c9134e333e2f9af136c1a124da4d37d326f10
Reviewed-on: https://go-review.googlesource.com/c/go/+/451555
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Auto-Submit: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
Reviewed-by: David Chase <drchase@google.com>
2 years agopath/filepath: detect Windows CONIN$ and CONOUT$ paths in IsLocal
Damien Neil [Thu, 17 Nov 2022 18:06:50 +0000 (10:06 -0800)]
path/filepath: detect Windows CONIN$ and CONOUT$ paths in IsLocal

CreateFile creates a handle to the console input or screen buffer
when opening a file named CONIN$ or CONOUT$:

https://learn.microsoft.com/en-us/windows/win32/api/fileapi/nf-fileapi-createfilea#consoles

Detect these paths as non-local.

For #56219.

Change-Id: Ib09e76a110d6ec09aef8038074b9bcbae09d00d7
Reviewed-on: https://go-review.googlesource.com/c/go/+/451657
Run-TryBot: Damien Neil <dneil@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Bryan Mills <bcmills@google.com>
2 years agocrypto/x509: reduce boring test key size
Roland Shoemaker [Thu, 17 Nov 2022 16:51:03 +0000 (08:51 -0800)]
crypto/x509: reduce boring test key size

Generating 8192 bit keys times out on builders relatively frequently. We
just need something that isn't a boringAllowCert allowed key size so we
can test that a non-boringAllowCert signed intermediate works, so just
use 512 instead since it'll be significantly faster.

Fixes #56798

Change-Id: I416e0d8c3aa11ff44e9870755efa95c74d1013f1
Reviewed-on: https://go-review.googlesource.com/c/go/+/451656
Auto-Submit: Roland Shoemaker <roland@golang.org>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Run-TryBot: Roland Shoemaker <roland@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Bryan Mills <bcmills@google.com>
2 years agogo/types, types2: implement type checking of "clear" built-in
Robert Griesemer [Fri, 4 Nov 2022 22:12:32 +0000 (15:12 -0700)]
go/types, types2: implement type checking of "clear" built-in

Will become available with Go 1.21.

Recognizing the `clear` built-in early is not causing any problems:
if existing code defines a `clear`, that will be used as before. If
code doesn't define `clear` the error message will make it clear
that with 1.21 the function will be available. It's still possible
to define a local `clear` and get rid of the error; but more likely
the name choice should be avoided going forward, so this provides a
useful early "heads-up".

For #56351.

Change-Id: I3d0fb1eb3508fbc78d7514b6238eac89610158c9
Reviewed-on: https://go-review.googlesource.com/c/go/+/448076
Run-TryBot: Robert Griesemer <gri@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Robert Findley <rfindley@google.com>
Reviewed-by: Robert Griesemer <gri@google.com>
Auto-Submit: Robert Griesemer <gri@google.com>

2 years agogo/parser: allow trailing commas in embedded instantiated types
cia-rana [Mon, 14 Nov 2022 08:01:03 +0000 (17:01 +0900)]
go/parser: allow trailing commas in embedded instantiated types

go/parser can correctly parse interfaces that instantiate and embed
generic interfaces, but not structs. This is because in the case of
structs, it does not expect RBRACK as a token trailing COMMA in the type
argument, even though it is allowed by the spec.

For example, go/parser produces an error for the type declaration below:

type A struct {
    B[byte, []byte,]
}

Fixes #56748

Change-Id: Ibb2addd6cf9b381d8470a6d20eedb93f13f93cd6
Reviewed-on: https://go-review.googlesource.com/c/go/+/450175
Run-TryBot: Robert Griesemer <gri@google.com>
Reviewed-by: Robert Findley <rfindley@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Robert Griesemer <gri@google.com>
Reviewed-by: Robert Griesemer <gri@google.com>
2 years agoos: use testenv.Command instead of exec.Command in tests
cui fliter [Wed, 16 Nov 2022 14:45:54 +0000 (22:45 +0800)]
os: use testenv.Command instead of exec.Command in tests

testenv.Command sets a default timeout based on the test's deadline
and sends SIGQUIT (where supported) in case of a hang.

Change-Id: I32ea9ca11c30d8af3d5490f2db1674314962cc80
Reviewed-on: https://go-review.googlesource.com/c/go/+/451195
Reviewed-by: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Joedian Reid <joedian@golang.org>
Auto-Submit: Bryan Mills <bcmills@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>

2 years agointernal/profile: remove unused encodeStringOpt
Tobias Klauser [Thu, 17 Nov 2022 12:27:57 +0000 (13:27 +0100)]
internal/profile: remove unused encodeStringOpt

Change-Id: Icaee8064c5ab61cc2ad28c2d1d06d70f845754fa
Reviewed-on: https://go-review.googlesource.com/c/go/+/451535
Reviewed-by: Joedian Reid <joedian@golang.org>
Reviewed-by: Cherry Mui <cherryyz@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>

2 years agotest: remove optimizationOff
Cuong Manh Le [Wed, 16 Nov 2022 16:11:07 +0000 (23:11 +0700)]
test: remove optimizationOff

Cl 426334 removed its only usage, and now we have gcflags_noopt.

Change-Id: I3b33a8c868669deea00bf6dfcf8d81981504e293
Reviewed-on: https://go-review.googlesource.com/c/go/+/451255
Reviewed-by: Joedian Reid <joedian@golang.org>
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Auto-Submit: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agotime: avoid creating a parse error from the next chunk of the value
Zeke Lu [Thu, 17 Nov 2022 12:19:49 +0000 (12:19 +0000)]
time: avoid creating a parse error from the next chunk of the value

When it reports a parse error, it uses the "value" variable as the
value element of the parse error. Previously, in some of the cases,
the "value" variable is always updated to the next chunk of the value
to be parsed (even if an earlier chunk is invalid). The reported
parse error is confusing in this case.

This CL addresses this issue by holding the original value, and when
it fails to parse the time, use it to create the parse error.

Fixes #56730.

Change-Id: I445b1d8a1b910208d0608b2186881746adb550e0
GitHub-Last-Rev: 67b1102b5e9b345beb2ddcc529a8e608e5afc865
GitHub-Pull-Request: golang/go#56754
Reviewed-on: https://go-review.googlesource.com/c/go/+/450936
Reviewed-by: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Joedian Reid <joedian@golang.org>
Auto-Submit: Bryan Mills <bcmills@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>

2 years agocmd/compile: handle integer conversions in static init inliner
Russ Cox [Wed, 16 Nov 2022 21:04:08 +0000 (16:04 -0500)]
cmd/compile: handle integer conversions in static init inliner

Given code like

func itou(i int) uint { return uint(i) }
var x = itou(-1)

the static inliner from CL 450136 was rewriting the code to

var x = uint(-1)

which is not valid Go code. Fix this by converting the
constants appropriately during inlining.

Fixes golang.org/x/image/vector test.

Change-Id: I13448df8504c6a70525b1cdc36e2c947e22cdd33
Reviewed-on: https://go-review.googlesource.com/c/go/+/451376
Auto-Submit: Russ Cox <rsc@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
Run-TryBot: Russ Cox <rsc@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agogo/types: mention FakeImportC in Package.Import doc comment
Alan Donovan [Tue, 18 Oct 2022 12:25:51 +0000 (08:25 -0400)]
go/types: mention FakeImportC in Package.Import doc comment

Change-Id: If4a82893f855d86de3311247424b1a3740c1f212
Reviewed-on: https://go-review.googlesource.com/c/go/+/443615
Reviewed-by: Robert Griesemer <gri@google.com>
Run-TryBot: Alan Donovan <adonovan@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/dist: skip non-race tests
Austin Clements [Thu, 17 Nov 2022 01:58:38 +0000 (20:58 -0500)]
cmd/dist: skip non-race tests

In -race mode, the dist test command only registers the std, race,
osusergo, and amd64ios tests before returning early from
(*tester).registerTests. Prior to CL 450018, the osusergo and amd64ios
tests weren't even affected by -race mode, so it seems their inclusion
was unintentional. CL 450018 lifted the logic to run tests in race
mode, which means these tests went from running without -race to
running with -race. Unfortunately, amd64ios is not compatible with
-race, so it is now failing on the darwin-amd64-race builder.

Fix this by omitting the osusergo and amd64ios tests from -race mode,
since it seems like they were really intended to be included anyway.

This should fix the darwin-amd64-race builder.

Updates #37486.

Change-Id: I554bb60bc729dbb6f1bc926f1ea329768b0d6d81
Reviewed-on: https://go-review.googlesource.com/c/go/+/451437
Reviewed-by: Bryan Mills <bcmills@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Austin Clements <austin@google.com>

2 years agocmd/fix: allow cgo commands in tests to fail if 'go build' is not supported
Bryan C. Mills [Thu, 17 Nov 2022 01:24:03 +0000 (20:24 -0500)]
cmd/fix: allow cgo commands in tests to fail if 'go build' is not supported

testenv.HasCgo reports whether the test binary may have been built
with cgo enabled, but having been built with cgo does not necessarily
imply that the test can invoke the cgo tool itself.

This should fix a test failure on the android builders introduced in
CL 450714.

Change-Id: I2eed4098736e1cb285ca20bc248b0ab3515f0dea
Reviewed-on: https://go-review.googlesource.com/c/go/+/451221
Auto-Submit: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>

2 years agopath/filepath: remove IsLocal tests dependent on Windows version
Damien Neil [Thu, 17 Nov 2022 01:33:03 +0000 (17:33 -0800)]
path/filepath: remove IsLocal tests dependent on Windows version

Don't test IsLocal("nul.txt"), since whether this name is reserved
or not is version-dependent.

Change-Id: Ifff3edc77454e052080e325871c08bbba49e692c
Reviewed-on: https://go-review.googlesource.com/c/go/+/451222
Reviewed-by: Austin Clements <austin@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Damien Neil <dneil@google.com>
Reviewed-by: Bryan Mills <bcmills@google.com>
2 years agocmd/go: disable cgo by default if DefaultCC is absolute and doesn't exist
Bryan C. Mills [Wed, 16 Nov 2022 22:27:06 +0000 (17:27 -0500)]
cmd/go: disable cgo by default if DefaultCC is absolute and doesn't exist

Also fix the autocgo test from CL 450739 when DefaultCC is absolute.

Change-Id: Ie282a42a1334660225e88680b63b18b7c1ecba2c
Reviewed-on: https://go-review.googlesource.com/c/go/+/451219
Auto-Submit: Bryan Mills <bcmills@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>
2 years agonet: change resolverConfig.dnsConfig to an atomic.Pointer
Ian Lance Taylor [Wed, 16 Nov 2022 22:19:47 +0000 (14:19 -0800)]
net: change resolverConfig.dnsConfig to an atomic.Pointer

We were using a RWMutex RLock around a single memory load,
which is not a good use of a RWMutex--it introduces extra work
for the RLock but contention around a single memory load is unlikely.
And, the tryUpdate method was not acquiring the mutex anyhow.

The new atomic.Pointer type is type-safe and easy to use correctly
for a simple use-case like this.

Change-Id: Ib3859c03414c44d2e897f6d15c92c8e4b5c81a11
Reviewed-on: https://go-review.googlesource.com/c/go/+/451416
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Damien Neil <dneil@google.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Run-TryBot: Ian Lance Taylor <iant@google.com>

2 years agoarchive/tar, archive/zip: return ErrInsecurePath for unsafe paths
Damien Neil [Thu, 22 Sep 2022 23:22:04 +0000 (16:22 -0700)]
archive/tar, archive/zip: return ErrInsecurePath for unsafe paths

Return a distinguishable error when reading an archive file
with a path that is:

- absolute
- escapes the current directory (../a)
- on Windows, a reserved name such as NUL

Users may ignore this error and proceed if they do not need name
sanitization or intend to perform it themselves.

Fixes #25849
Fixes #55356

Change-Id: Ieefa163f00384bc285ab329ea21a6561d39d8096
Reviewed-on: https://go-review.googlesource.com/c/go/+/449937
Reviewed-by: Joseph Tsai <joetsai@digital-static.net>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Damien Neil <dneil@google.com>
Auto-Submit: Damien Neil <dneil@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
2 years agopath/filepath: add IsLocal
Damien Neil [Thu, 10 Nov 2022 01:49:44 +0000 (17:49 -0800)]
path/filepath: add IsLocal

IsLocal reports whether a path lexically refers to a location
contained within the directory in which it is evaluated.
It identifies paths that are absolute, escape a directory
with ".." elements, and (on Windows) paths that reference
reserved device names.

For #56219.

Change-Id: I35edfa3ce77b40b8e66f1fc8e0ff73cfd06f2313
Reviewed-on: https://go-review.googlesource.com/c/go/+/449239
Run-TryBot: Damien Neil <dneil@google.com>
Reviewed-by: Joseph Tsai <joetsai@digital-static.net>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Joedian Reid <joedian@golang.org>
2 years agocmd/compile: elide unnecessary Not in Phi block controls
Jakub Ciolek [Wed, 9 Nov 2022 15:07:11 +0000 (16:07 +0100)]
cmd/compile: elide unnecessary Not in Phi block controls

For a BlockIf, we can change the order of the successors if all OpPhi args are an OpNot and this allows us to elide said OpNots.

When compiling Go itself, there were no hits for (If (Phi (Not x) (Not y) (Not z)) or any other longer patterns.

compilecmp:

errors
errors.As changed

reflect
reflect.Value.FieldByIndex changed
reflect.Value.Method changed
reflect.cvtI2I changed
reflect.Value.FieldByIndexErr changed
reflect.deepValueEqual.func1 496 -> 502  (+1.21%)
reflect.deepValueEqual changed

internal/fmtsort
internal/fmtsort.nilCompare 652 -> 648  (-0.61%)

database/sql
database/sql.convertAssignRows changed

encoding/json
encoding/json.interfaceEncoder changed
encoding/json.(*decodeState).unmarshal 574 -> 571  (-0.52%)
encoding/json.mapEncoder.encode changed
encoding/json.ptrEncoder.encode changed
encoding/json.encodeByteSlice changed
encoding/json.addrMarshalerEncoder changed
encoding/json.addrTextMarshalerEncoder changed
encoding/json.(*decodeState).object changed
encoding/json.sliceEncoder.encode changed
encoding/json.indirect 1303 -> 1286  (-1.30%)

encoding/gob
encoding/gob.encOpFor.func3 changed
encoding/gob.encIndirect changed
encoding/gob.encOpFor.func5 changed
encoding/gob.(*Encoder).encodeInterface changed
encoding/gob.(*Decoder).decodeMap changed

encoding/xml
encoding/xml.(*printer).marshalStruct changed
encoding/xml.(*fieldInfo).value changed
encoding/xml.(*printer).marshalAttr changed
encoding/xml.indirect changed
encoding/xml.(*printer).marshalValue changed

encoding/asn1
encoding/asn1.UnmarshalWithParams 837 -> 845  (+0.96%)

text/template
text/template.indirectInterface changed
text/template.indirect changed
text/template.safeCall changed

net/http/httptrace
net/http/httptrace.(*ClientTrace).compose changed

cmd/fix
main.typefix.func2 changed

cmd/compile/internal/ir
cmd/compile/internal/ir.dumpNode changed

cmd/gofmt
main.match changed
main.subst changed

cmd/compile/internal/ssa
cmd/compile/internal/ssa.rewriteBlockgeneric 626 -> 1030  (+64.54%)

Change-Id: I645b3b3e37302a63e06b79ce74674882fb603ef3
Reviewed-on: https://go-review.googlesource.com/c/go/+/449055
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Joedian Reid <joedian@golang.org>
Run-TryBot: Jakub Ciolek <jakub@ciolek.dev>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
2 years agocmd/go: add PGO auto mode
Cherry Mui [Mon, 3 Oct 2022 23:07:39 +0000 (19:07 -0400)]
cmd/go: add PGO auto mode

Add "auto" mode for the -pgo build flag. When -pgo=auto is
specified, if there is a default.pgo file in the directory of the
main package, it will be selected and used for the build.

Currently it requires exactly one main package when -pgo=auto is
specified. (We'll support multiple main packages in the future.)

Also apply to other build-related subcommands, "go install", "go
run", "go test", and "go list".

For #55022.

Change-Id: Iab7974ab8932daf0e83506de505e044a8e412466
Reviewed-on: https://go-review.googlesource.com/c/go/+/438737
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Bryan Mills <bcmills@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>

2 years agocmd/fix: disallow cgo errors in tests
Bryan C. Mills [Tue, 15 Nov 2022 17:34:31 +0000 (12:34 -0500)]
cmd/fix: disallow cgo errors in tests

The 'cgo' command invoked by 'go fix' was not valid when built with
-trimpath, but the test was not failing because errors from the
command were being logged and ignored instead of causing tests to
fail. Changing the code and test not to ignore the errors revealed
that a number of existing tests were always, unconditionally
triggering cgo errors which were then ignored.

This change updates those tests to no longer produce cgo errors,
and to check their results when cgo is enabled.

For #51473.
Updates #51461.

Change-Id: Ib9d1ea93f26d30daa824d75ed634eaf530af086d
Reviewed-on: https://go-review.googlesource.com/c/go/+/450714
Reviewed-by: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Bryan Mills <bcmills@google.com>
Auto-Submit: Bryan Mills <bcmills@google.com>

2 years agocmd/go: automatically disable cgo on systems with no C compiler
Russ Cox [Wed, 16 Nov 2022 03:30:45 +0000 (22:30 -0500)]
cmd/go: automatically disable cgo on systems with no C compiler

The documentation for cgo has always said:

> The cgo tool is enabled by default for native builds
> on systems where it is expected to work.

Following the spirit of that rule, this CL disables cgo by default
on systems where $CC is unset and the default C compiler
(clang or gcc) is not found in $PATH.

This CL makes builds of Go code on systems with no C compiler
installed automatically fall back to non-cgo mode.
For example, if building a Go program using package net
in a stripped down Linux container, that build will now run
with cgo disabled, instead of attempting the build with cgo enabled
and only succeeding if the right pre-compiled .a files happen to
be loaded into the container.

This CL makes it safe to drop the pre-compiled .a files
from the Go distribution. Systems that don't have a C compiler
will simply disable cgo when building new .a files for that system.

In general keeping the pre-compiled .a files working in cgo mode
on systems without C compilers has had only mixed success due
to the precise build cache. Today we have had to disable various
checks in the precise build cache so that distributed .a files look
up-to-date even if the current machine's C compiler is a different
version than the one used when packaging the distribution.
Each time we improve precision we have a decent chance of
re-invalidating the files. This CL, combined with dropping the .a files
entirely, will let us re-enable those checks and ensure that the
.a files used in a build actually match the C compiler being used.

On macOS, the distributed .a files for cgo-dependent packages
have been stale (not actually used by the go command) since the
release of Go 1.14 in February 2020, due to CL 216304 setting
a CGO_CFLAGS environment variable that won't match the default
setting on users machines. (To keep the distributed .a files working,
that CL should have instead changed the default in the go command.)
The effect is that for the past six Go releases (!!!), the go command
has been unable to build basic programs like src/net/http/triv.go
on macOS without either disabling cgo or installing Xcode's C compiler.
This CL fixes that problem by disabling cgo when there's no C compiler.
Now it will once again be possible to build basic programs with just
a Go toolchain installed.

In the past, disabling cgo on macOS would have resulted in subpar
implementations of crypto/x509, net, and os/user, but as of CL 449316
those packages have all been updated to use libc calls directly,
so they now provide the same implementation whether or not cgo is enabled.
In the past, disabling cgo on macOS would also have made the
race detector unusable, but CL 451055 makes the race detector
work on macOS even when cgo is disabled.

On Windows, none of the standard library uses cgo today, so all
the necessary .a files can be rebuilt without a C toolchain,
and there is no loss of functionality in the standard library when
cgo is disabled. After this CL, the race detector won't work on
Windows without a C toolchain installed, but that turns out to be
true already: when linking race-enabled programs, even if the Go linker
does not invoke the host linker, it still attempts to read some of the
host C toolchain's .a files to resolve undefined references.

On Unix systems, disabling cgo when a C compiler is not present
will mean that builds get the pure Go net resolver, which is used
by default even in cgo builds when /etc/resolv.conf is simple enough.
It will also mean they get the pure os/user code, which reads
/etc/passwd and /etc/group instead of using shared libraries,
and therefore it may miss out on other sources of user information
such as LDAP. The race detector also will not work without a C compiler.
This would be dire except that nearly all Unix systems have a C compiler
installed by default, and on those that don't it is trivial to add one.
In particular, the vast majority of Go developers running on Linux
and other Unix systems will already have a C compiler and will be
unaffected.

Change-Id: I491e8a022fe3a64022e9dc593850d483a0d353fa
Reviewed-on: https://go-review.googlesource.com/c/go/+/450739
Reviewed-by: Bryan Mills <bcmills@google.com>
Auto-Submit: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agoruntime/race: do not use cgo on macOS
Russ Cox [Wed, 16 Nov 2022 01:36:10 +0000 (20:36 -0500)]
runtime/race: do not use cgo on macOS

The use of an empty import "C" to trigger cgo in runtime/race
serves two purposes:

1. Cause the runtime to use the C library to create system threads,
   because the race syso implementation expects things like
   thread-local storage to work correctly.

2. Derive the right set of //go:cgo_import_dynamic comments
   to pass to the Go linker, so that it doesn't diagnose them as
   undefined references.

On macOS, (1) is unnecessary because using the C library
(via DLL calls) is the only way the runtime ever creates threads.

We can accomplish (2) by writing those comments ourselves.

Having done that in this CL, cgo is no longer needed to run
the race detector on macOS, which means that having a
pre-compiled set of .a files is no longer necessary,
nor is having Xcode for use with cgo when rebuilding those .a files.

Change-Id: Iee24cc67900eb542141b32beaadafb2c94f5fe26
Reviewed-on: https://go-review.googlesource.com/c/go/+/451055
Run-TryBot: Russ Cox <rsc@golang.org>
Reviewed-by: Bryan Mills <bcmills@google.com>
Auto-Submit: Russ Cox <rsc@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Cherry Mui <cherryyz@google.com>
2 years agocmd/dist: simplify old code
Austin Clements [Sat, 12 Nov 2022 02:57:57 +0000 (21:57 -0500)]
cmd/dist: simplify old code

Now that all uses of "go test" have been converted over to the new
abstraction, we can delete the old helpers for building "go test"
commands and simplify some code that's only used by the new
abstraction now.

For #37486.

Change-Id: I770cd457e018160d694abcc0b6ac80f7dc2e8425
Reviewed-on: https://go-review.googlesource.com/c/go/+/450020
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Austin Clements <austin@google.com>

2 years agocmd/dist: convert host tests to use goTest
Austin Clements [Thu, 10 Nov 2022 18:54:25 +0000 (13:54 -0500)]
cmd/dist: convert host tests to use goTest

This adds support for host tests to goTest and registerTest and
modifies all uses of registerHostTest to use goTest and registerTest.

This eliminates the last case where go test command lines are
constructed by hand. Next we'll clean up all of the infrastructure
support for that.

I traced all exec calls from cmd/dist on linux/amd64 and this makes
only no-op changes (such as re-arranging the order of flags).

Preparation for #37486.

Change-Id: Icb7ec8efdac72bdb819ae24b2f585375d9d9d5b9
Reviewed-on: https://go-review.googlesource.com/c/go/+/450019
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
2 years agocmd/dist: convert most remaining tests to use goTest
Austin Clements [Fri, 4 Nov 2022 18:54:04 +0000 (14:54 -0400)]
cmd/dist: convert most remaining tests to use goTest

This converts most of the remaining manual "go test" command line
construction in cmd/dist to use the goTest abstraction and
registerTest.

At this point, the only remaining place that directly constructs go
test command lines is runHostTest.

This fixes a bug in the "nolibgcc:os/user" test. It was clearly
supposed to pass "-run=^Test[^CS]", but the logic to override the
"-run" flag for "nolibgcc:net" caused "nolibgcc:os/user" to pass
*both* "-run=^Test[^CS]" and "-run=". This was then rewritten into
just "-run=" by flattenCmdline, which caused all os/user tests to run,
and not actually skip the expensive tests as intended. (This is a
great example of why the new abstraction is much more robust than
command line construction.)

I traced all exec calls from cmd/dist on linux/amd64 and, other than
the fix to nolibgcc:os/user, this makes only no-op changes (such as
re-arranging the order of flags).

For #37486.

Change-Id: Ie8546bacc56640ea39f2804a87795c14a3fe4c7d
Reviewed-on: https://go-review.googlesource.com/c/go/+/450018
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
2 years agocmd/dist: restructure cgo_test
Austin Clements [Fri, 4 Nov 2022 18:49:37 +0000 (14:49 -0400)]
cmd/dist: restructure cgo_test

Currently, dist test has a single test called "cgo_test" that runs a
large number of different "go test"s.

This commit restructures cgo_test into several individual tests, each
of which runs a single "go test" that can be described by a goTest
object and registered with registerTest. Since this lets us raise the
abstraction level of constructing these tests and these tests are
mostly covering the Cartesian product of a small number of orthogonal
dimensions, we pull the common logic for constructing these tests into
a helper function.

For consistency, we now pass -tags=static to the static testtls and
nocgo tests, but this tag doesn't affect the build of these tests at
all. I traced all exec calls from cmd/dist on linux/amd64 and this is
the only non-trivial change.

For #37486.

Change-Id: I53c1efa1c38d785dc71968f05e8d7d636b553e96
Reviewed-on: https://go-review.googlesource.com/c/go/+/450017
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/dist: make registerTest take a goTest
Austin Clements [Thu, 13 Oct 2022 14:57:13 +0000 (10:57 -0400)]
cmd/dist: make registerTest take a goTest

The overall goal is to make registerTest the primary entry point for
adding dist tests and to convert nearly all dist tests to be
represented by a goTest, registered via registerTest. This will
centralize the logic for creating dist tests corresponding to go tool
tests.

I traced all exec calls from cmd/dist on linux/amd64 and this makes
only no-op changes (such as re-arranging the order of flags).

For #37486.

Change-Id: I4749e6f3666134d3259b54ee6055d76a4235c60c
Reviewed-on: https://go-review.googlesource.com/c/go/+/450016
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/dist: use goTest for manual go test invocations
Austin Clements [Wed, 2 Nov 2022 18:33:04 +0000 (14:33 -0400)]
cmd/dist: use goTest for manual go test invocations

This CL rewrites everywhere in dist that manually constructs an
exec.Cmd to run "go test" to use the goTest abstraction. All remaining
invocations of "go test" after this CL construct the command line
manually, but ultimately use addCmd to execute it.

I traced all exec calls from cmd/dist on linux/amd64 and this makes
only no-op changes (such as re-arranging the order of flags).

For #37486.

Change-Id: Idc7497e39bac04def7ddaf2010881c9623e76fd4
Reviewed-on: https://go-review.googlesource.com/c/go/+/450015
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Run-TryBot: Austin Clements <austin@google.com>

2 years agocmd/dist: introduce "go test" abstraction
Austin Clements [Wed, 2 Nov 2022 18:27:28 +0000 (14:27 -0400)]
cmd/dist: introduce "go test" abstraction

This introduces an abstraction for constructing and running "go test"
commands. Currently, dist test is basically a shell script written in
Go syntax: it mostly just invokes lots of subprocesses, almost all of
which are "go test" invocations, and it constructs those command lines
directly from strings all over the place.

This CL raises the level of abstraction of invoking go test. The
current level of abstraction is not serving us very well: it's
conveniently terse, but the actual logic for constructing a command
line is typically so spread out that it's difficult to predict what
command will actually run. For example, the `gotest` function
constructs the basic command, but many tests want to override at least
some of these flags, so flattenCmdLine has logic specific to `go test`
for eliminating duplicate flags that `go test` itself would reject. At
the same time, the logic for constructing many common flags is
conditional, leading to a bevy of helpers for constructing flags like
`-short` and `-timeout` and `-run` that are scattered throughout
test.go and very easy to forget to call.

This CL centralizes and flattens all of this knowledge into a new
`goTest` type. This type gives dist a single, unified point where we
can change anything about how it invokes "go test".

There's currently some "unnecessary" abstraction in the implementation
of the goTest type to separate "build" and "run" flags. This will
become important later when we convert host tests and to do separate
build and run steps.

The following CLs will convert dist test to use this type rather than
directly constructing "go test" command lines. Finally, we'll strip
out the scattered helper logic for building command lines.

For #37486.

Change-Id: I9f1633fe6c0921696419ce8127ed2ca7b7a4e01b
Reviewed-on: https://go-review.googlesource.com/c/go/+/448802
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agomath/rand: deprecate Seed
Russ Cox [Wed, 16 Nov 2022 17:04:32 +0000 (12:04 -0500)]
math/rand: deprecate Seed

Programs that call Seed and then expect a specific sequence
of results from the global random source (using functions such as Int)
can be broken when a dependency changes how much it consumes
from the global random source. To avoid such breakages, programs
that need a specific result sequence should use NewRand(NewSource(seed))
to obtain a random generator that other packages cannot access.

Fixes #56319.

Change-Id: Idac33991b719d2c71f109f51dacb3467a649e01e
Reviewed-on: https://go-review.googlesource.com/c/go/+/451375
Reviewed-by: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>

2 years agocmd/internal/obj/ppc64: add ISA 3.1B opcodes
Paul E. Murphy [Thu, 7 Jul 2022 19:26:31 +0000 (14:26 -0500)]
cmd/internal/obj/ppc64: add ISA 3.1B opcodes

A few new opcodes are added to support ROP mitigation on
Power10.

Change-Id: I13045aebc0b6fb09c64dc234ee5741318670d7ff
Reviewed-on: https://go-review.googlesource.com/c/go/+/425597
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Paul Murphy <murp@ibm.com>
Reviewed-by: Lynn Boger <laboger@linux.vnet.ibm.com>
Reviewed-by: Joedian Reid <joedian@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/go: make testterminal18153 a normal test
Austin Clements [Fri, 11 Nov 2022 16:51:28 +0000 (11:51 -0500)]
cmd/go: make testterminal18153 a normal test

Currently, cmd/go's testterminal18153 is implemented as a special test
that doesn't run as part of cmd/go's regular tests. Because the test
requires stdout and stderr to be a terminal, it is currently run
directly by "dist test" so it can inherit the terminal of all.bash.

This has a few problems. It's yet another special case in dist test.
dist test also has to be careful to not apply its own buffering to
this test, so it can't run in parallel and it limits dist test's own
scheduler. It doesn't run as part of regular "go test", which means it
usually only gets coverage from running all.bash. And since we have to
skip it if all.bash wasn't run at a terminal, I'm sure it often gets
skipped even when running all.bash.

Fix all of this by rewriting this test to create its own PTY and
re-exec "go test" to check that PTY passes through go test. This makes
the test self-contained, so it can be a regular cmd/go test, we can
drop it from dist test, and it's not sensitive to the environment of
all.bash.

Preparation for #37486.
Updates #18153.

Change-Id: I6493dbb0143348e299718f6e311ac8a63f5d69c5
Reviewed-on: https://go-review.googlesource.com/c/go/+/449503
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agointernal/testpty: move from os/signal/internal/pty
Austin Clements [Thu, 10 Nov 2022 21:02:48 +0000 (16:02 -0500)]
internal/testpty: move from os/signal/internal/pty

We're going to use this for another test, so make it more accessible.

Preparation for #37486.

Change-Id: If194cc4244c4b9e1b1f253759b813555b39ad67e
Reviewed-on: https://go-review.googlesource.com/c/go/+/449502
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Bryan Mills <bcmills@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/dist: eliminate "time" special case
Austin Clements [Mon, 26 Sep 2022 20:04:10 +0000 (16:04 -0400)]
cmd/dist: eliminate "time" special case

The registerTest function has a special case for commands that start
with "time", but we don't use this case anywhere. Delete this special
case and its support code.

Preparation for #37486.

Change-Id: Ica180417e7aa4e4fc260cb97467942bae972fdb6
Reviewed-on: https://go-review.googlesource.com/c/go/+/448801
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/dist: eliminate registerSeqTest
Austin Clements [Fri, 23 Sep 2022 21:23:54 +0000 (17:23 -0400)]
cmd/dist: eliminate registerSeqTest

This is unused, and eliminating it lets us simplify the whole
registerTest mechanism.

Preparation for #37486.

Change-Id: Ia6221e48192cd17775a5d662bdb389d67a9265bc
Reviewed-on: https://go-review.googlesource.com/c/go/+/448800
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/dist: add doc comments to some functions
Austin Clements [Thu, 13 Oct 2022 14:54:11 +0000 (10:54 -0400)]
cmd/dist: add doc comments to some functions

I keep having to reconstruct how dist test's command system works.
This CL documents the main functions to help with this.

Change-Id: Ic1102496d3c7f356e73f3fa2ea76498251f525d4
Reviewed-on: https://go-review.googlesource.com/c/go/+/448799
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
2 years agotesting: in TestRacyOutput, ensure that the Run calls complete before cleanup
Bryan C. Mills [Wed, 16 Nov 2022 15:17:17 +0000 (10:17 -0500)]
testing: in TestRacyOutput, ensure that the Run calls complete before cleanup

Also tweak the failure message added in CL 352349.

Updates #18741.
Updates #48515.

Change-Id: I46ed84c6f498d7a68414cc3dab3c1cd55da69aa9
Reviewed-on: https://go-review.googlesource.com/c/go/+/451215
Reviewed-by: Austin Clements <austin@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>
Reviewed-by: Changkun Ou <mail@changkun.de>
Auto-Submit: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/dist: add curiously missing GOOS/GOARCH combinations
Austin Clements [Fri, 4 Nov 2022 18:58:40 +0000 (14:58 -0400)]
cmd/dist: add curiously missing GOOS/GOARCH combinations

The cmd/dist cgo_test enumerates a large number of platforms in
various special cases. Some combinations are suspiciously absent. This
CL completes the combinations.

I've confirmed using trybots that the newly-enabled tests pass on
android/* (this is not surprising because the gohostos is never
"android" anyway), windows/arm64, linux/ppc64 (no cgo), linux/loong64
(except for one test, filed #56623), linux/mips*, netbsd/arm (except
for one test, filed #56629), and netbsd/arm64. The windows/arm builder
is out to lunch, so I'm assuming that works. Since netbsd/arm and
arm64 mostly passed these tests, I've also enabled them on netbsd/386
and netbsd/amd64, where they seem to work fine as well.

Preparation for #37486.

Change-Id: I04c3348e4f422d74d51e714647ca3db379e6e919
Reviewed-on: https://go-review.googlesource.com/c/go/+/448016
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Austin Clements <austin@google.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
2 years agonet: remove unused goLookupHost method
Mateusz Poliwczak [Wed, 16 Nov 2022 11:18:32 +0000 (11:18 +0000)]
net: remove unused goLookupHost method

Change-Id: I62b9d0dcbec647fdd3279d78b9999dd933ff0bbd
GitHub-Last-Rev: f83db66c935dd7a14128d1e87c9323539e9d2d16
GitHub-Pull-Request: golang/go#56758
Reviewed-on: https://go-review.googlesource.com/c/go/+/451135
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Run-TryBot: Ian Lance Taylor <iant@google.com>
Reviewed-by: Joedian Reid <joedian@golang.org>
Auto-Submit: Ian Lance Taylor <iant@google.com>

2 years agogo/types, types2: ensure invalid generic types are marked as invalid
Robert Findley [Thu, 10 Nov 2022 00:16:39 +0000 (19:16 -0500)]
go/types, types2: ensure invalid generic types are marked as invalid

When detecting invalid types, we may detect cycles through instances.
Ensure that the uninstantiated origin type is also marked invalid.

Fixes #56665

Change-Id: Id67653bcb072ac80161dea07d0ced566e61564a8
Reviewed-on: https://go-review.googlesource.com/c/go/+/449275
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@google.com>
2 years agotesting: reject calls to Run within Cleanup callbacks
Changkun Ou [Mon, 27 Sep 2021 09:37:56 +0000 (11:37 +0200)]
testing: reject calls to Run within Cleanup callbacks

Calling t.Run inside t.Cleanup can mess up the execution order of
registered Cleanup callbacks. Reject calls to Run within Cleanup
callbacks.

Fixes #48515

Change-Id: I61e4cb35253db1a8bbe3351d59055433030aa289
Reviewed-on: https://go-review.googlesource.com/c/go/+/352349
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Joedian Reid <joedian@golang.org>
Run-TryBot: Changkun Ou <mail@changkun.de>
Auto-Submit: Bryan Mills <bcmills@google.com>
Reviewed-by: Bryan Mills <bcmills@google.com>
2 years agointernal/goroot: report PkgfileMap error on invalid line
Tobias Klauser [Tue, 15 Nov 2022 11:26:11 +0000 (12:26 +0100)]
internal/goroot: report PkgfileMap error on invalid line

Ref. https://go-review.googlesource.com/c/go/+/442303/comment/7caca6eb_0ebe4d51/

Change-Id: Id351b9c25380f0959453bb84ed123d0e784e4866
Reviewed-on: https://go-review.googlesource.com/c/go/+/450595
Auto-Submit: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Joedian Reid <joedian@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>

2 years agocrypto/x509: add support for PKCS8/PKIX X25519 key encodings
Filippo Valsorda [Tue, 15 Nov 2022 17:43:56 +0000 (18:43 +0100)]
crypto/x509: add support for PKCS8/PKIX X25519 key encodings

This specifically doesn't add support for X25519 certificates.
Refactored parsePublicKey not to depend on the public PublicKeyAlgorithm
values, and ParseCertificate/ParseCertificateRequest to ignore keys that
don't have a PublicKeyAlgorithm even if parsePublicKey supports them.

Updates #56088

Change-Id: I2274deadfe9bb592e3547c0d4d48166de1006df0
Reviewed-on: https://go-review.googlesource.com/c/go/+/450815
Reviewed-by: Roland Shoemaker <roland@golang.org>
Run-TryBot: Filippo Valsorda <filippo@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Joedian Reid <joedian@golang.org>
Auto-Submit: Filippo Valsorda <filippo@golang.org>

2 years agocrypto/ecdh: move ECDH method to PrivateKey
Filippo Valsorda [Mon, 14 Nov 2022 11:13:46 +0000 (12:13 +0100)]
crypto/ecdh: move ECDH method to PrivateKey

Fixes #56052

Change-Id: Icacba0ed0f77519bca2140c8af68407af97f9734
Reviewed-on: https://go-review.googlesource.com/c/go/+/450335
Run-TryBot: Filippo Valsorda <filippo@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Joedian Reid <joedian@golang.org>
Auto-Submit: Filippo Valsorda <filippo@golang.org>

2 years agotest: fix noinit on noopt builder
Russ Cox [Wed, 16 Nov 2022 05:02:14 +0000 (00:02 -0500)]
test: fix noinit on noopt builder

Fix noopt build break from CL 450136 by not running test.

I can't reproduce the failure locally, but it's entirely reasonable
for this test to fail when optimizations are disabled, so just don't
run it when optimizations are disabled.

Change-Id: I882760fc7373ba0449379f81d295312a6be49be1
Reviewed-on: https://go-review.googlesource.com/c/go/+/450740
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: Michael Stapelberg <stapelberg@google.com>
Auto-Submit: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/compile: handle simple inlined calls in staticinit
Russ Cox [Sun, 13 Nov 2022 14:22:35 +0000 (09:22 -0500)]
cmd/compile: handle simple inlined calls in staticinit

Global variable initializers like

var myErr error = &myError{"msg"}

have been converted to statically initialized data
from the earliest days of Go: there is no init-time
execution or allocation for that line of code.

But if the expression is moved into an inlinable function,
the static initialization no longer happens.
That is, this code has always executed and allocated
at init time, even after we added inlining to the compiler,
which should in theory make this code equivalent to
the original:

func NewError(s string) error { return &myError{s} }
var myErr2 = NewError("msg")

This CL makes the static initialization rewriter understand
inlined functions consisting of a single return statement,
like in this example, so that myErr2 can be implemented as
statically initialized data too, just like myErr, with no init-time
execution or allocation.

A real example of code that benefits from this rewrite is
all globally declared errors created with errors.New, like

package io
var EOF = errors.New("EOF")

Package io no longer has to allocate and initialize EOF each
time a program starts.

Another example of code that benefits is any globally declared
godebug setting (using the API from CL 449504), like

package http
var http2server = godebug.New("http2server")

These are no longer allocated and initialized at program startup either.

The list of functions that are inlined into static initializers when
compiling std and cmd (along with how many times each occurs) is:

cmd/compile/internal/ssa.StringToAux (3)
cmd/compile/internal/walk.mkmapnames (4)
errors.New (360)
go/ast.NewIdent (1)
go/constant.MakeBool (4)
go/constant.MakeInt64 (3)
image.NewUniform (4)
image/color.ModelFunc (11)
internal/godebug.New (12)
vendor/golang.org/x/text/unicode/bidi.newBidiTrie (1)
vendor/golang.org/x/text/unicode/norm.newNfcTrie (1)
vendor/golang.org/x/text/unicode/norm.newNfkcTrie (1)

For the cmd/go binary, this CL cuts the number of init-time
allocations from about 1920 to about 1620 (a 15% reduction).

The total executable code footprint of init functions is reduced
by 24kB, from 137kB to 113kB (an 18% reduction).
The overall binary size is reduced by 45kB,
from 15.335MB to 15.290MB (a 0.3% reduction).
(The binary size savings is larger than the executable code savings
because every byte of executable code also requires corresponding
runtime tables for unwinding, source-line mapping, and so on.)

Also merge test/sinit_run.go, which had stopped testing anything
at all as of CL 161337 (Feb 2019) and initempty.go into a new test
noinit.go.

Fixes #30820.

Change-Id: I52f7275b1ac2a0a32e22c29f9095071c7b1fac20
Reviewed-on: https://go-review.googlesource.com/c/go/+/450136
Reviewed-by: Cherry Mui <cherryyz@google.com>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: Joedian Reid <joedian@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Auto-Submit: Russ Cox <rsc@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>

2 years agocmd/compile: do not emit a few more basic types from every compilation
Russ Cox [Sat, 12 Nov 2022 17:27:33 +0000 (12:27 -0500)]
cmd/compile: do not emit a few more basic types from every compilation

We already emit types for any and func(error) string in runtime.a
but unlike the other pre-emitted types, we don't then exclude them
from being emitted in other packages. Fix that.

Also add slices of non-func types that we already emit.

Saves 0.3% of .a files in std cmd deps, computed by adding sizes from:

ls -l $(go list -export -f '{{.Export}}' -deps std cmd

The effect is small and not worth doing on its own.
The real improvement is making “what to write always in runtime”
and “what not to write in other packages” more obviously aligned.

Change-Id: Ie5cb5fd7e5a3025d2776d9b4cece775fdf92d3b6
Reviewed-on: https://go-review.googlesource.com/c/go/+/450135
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Russ Cox <rsc@golang.org>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: Keith Randall <khr@google.com>
2 years agodoc/go1.20: document new sync.Map methods
Bryan C. Mills [Tue, 15 Nov 2022 22:07:00 +0000 (17:07 -0500)]
doc/go1.20: document new sync.Map methods

For #51972.

Change-Id: I86dcd8abc3b62e20b524541327af2cc891cb251d
Reviewed-on: https://go-review.googlesource.com/c/go/+/450797
Reviewed-by: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Bryan Mills <bcmills@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>

2 years agocmd/compile: be more careful about pointer incrementing in range loops
Keith Randall [Fri, 11 Nov 2022 23:30:10 +0000 (15:30 -0800)]
cmd/compile: be more careful about pointer incrementing in range loops

For range loops, we use a pointer to the backing store that gets
incremented on each iteration of the loop.

The problem with this scheme is that at the end of the last iteration,
we may briefly have a pointer that points past the end of the backing store
of the slice that is being iterated over. We cannot let the garbage collector
see that pointer.

To fix this problem, have the incremented pointer live briefly as
a uintptr instead of a normal pointer, so it doesn't keep anything
alive. Convert back to a normal pointer just after the loop condition
is checked, but before anything that requires a real pointer representation
(in practice, any call, which is what could cause a GC scan or stack copy).

Fixes #56699

Change-Id: Ia928d23f85a211565357603668bea4e5c534f989
Reviewed-on: https://go-review.googlesource.com/c/go/+/449995
Reviewed-by: David Chase <drchase@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
2 years agocrypto/rsa: return err when key too small to compute salt for RSA PSS
Roland Shoemaker [Tue, 15 Nov 2022 19:57:10 +0000 (11:57 -0800)]
crypto/rsa: return err when key too small to compute salt for RSA PSS

When PSSSaltLengthAuto is passed to SignPSS, and the key size is too
small to create a valid salt, return ErrMessageTooLong

Change-Id: I4e0d70bdd54fcd667eae10e0a70b4f540a4ebe93
Reviewed-on: https://go-review.googlesource.com/c/go/+/450796
Reviewed-by: Filippo Valsorda <filippo@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Bryan Mills <bcmills@google.com>
Run-TryBot: Roland Shoemaker <roland@golang.org>

2 years agocmd/addr2line: use the test binary as 'addr2line' instead of rebuilding it twice
Bryan C. Mills [Tue, 15 Nov 2022 16:06:03 +0000 (11:06 -0500)]
cmd/addr2line: use the test binary as 'addr2line' instead of rebuilding it twice

This not only reduces the latency of the test, but also respects
build flags like '-race' and '-cover' passed to the 'go test' command.

Change-Id: Ie9a3c03d32d7eea268ba6f8f8ac4000539434052
Reviewed-on: https://go-review.googlesource.com/c/go/+/450713
Auto-Submit: Bryan Mills <bcmills@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
2 years agocmd/addr2line: use testenv.Command instead of exec.Command in tests
Bryan C. Mills [Tue, 15 Nov 2022 14:48:39 +0000 (09:48 -0500)]
cmd/addr2line: use testenv.Command instead of exec.Command in tests

testenv.Command sets a default based on the test\'s deadline
and sends SIGQUIT (where supported) in case of a hang.

Change-Id: I6710d9866c16610310347d50be1c0916c2976e87
Reviewed-on: https://go-review.googlesource.com/c/go/+/450712
Reviewed-by: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Bryan Mills <bcmills@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>

2 years agocmd/vet: use testenv.Command instead of exec.Command in tests
Bryan C. Mills [Tue, 15 Nov 2022 15:53:41 +0000 (10:53 -0500)]
cmd/vet: use testenv.Command instead of exec.Command in tests

testenv.Command sets a default timeout based on the test's deadline
and sends SIGQUIT (where supported) in case of a hang.

Change-Id: I973be280cd0a7f495b491bfb0e04771389562e04
Reviewed-on: https://go-review.googlesource.com/c/go/+/450711
Reviewed-by: Ian Lance Taylor <iant@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Bryan Mills <bcmills@google.com>

2 years agocmd/vet: use the test binary as 'vet' instead of rebuilding it
Bryan C. Mills [Tue, 15 Nov 2022 15:51:11 +0000 (10:51 -0500)]
cmd/vet: use the test binary as 'vet' instead of rebuilding it

This not only reduces the latency of the test, but also respects
build flags like '-race' and '-cover' passed to the 'go test' command.

Change-Id: I4914870396f90097f45d00012e5601fe42655899
Reviewed-on: https://go-review.googlesource.com/c/go/+/450710
Run-TryBot: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Bryan Mills <bcmills@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
2 years agocmd/pprof: use testenv.Command instead of exec.Command in tests
Bryan C. Mills [Tue, 15 Nov 2022 15:47:46 +0000 (10:47 -0500)]
cmd/pprof: use testenv.Command instead of exec.Command in tests

testenv.Command sets a default timeout based on the test's deadline
and sends SIGQUIT (where supported) in case of a hang.

Change-Id: Iabd114dd23b85da524e7ea8415f2bbf2c54b380f
Reviewed-on: https://go-review.googlesource.com/c/go/+/450709
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Bryan Mills <bcmills@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Auto-Submit: Bryan Mills <bcmills@google.com>

2 years agocmd/pprof: use the test binary as 'pprof' instead of rebuilding it
Bryan C. Mills [Tue, 15 Nov 2022 15:47:00 +0000 (10:47 -0500)]
cmd/pprof: use the test binary as 'pprof' instead of rebuilding it

This not only reduces the latency of the test, but also respects
build flags like '-race' and '-cover' passed to the 'go test' command.

Change-Id: Icdf256420c4dce2da7a187513b7dd08393b76146
Reviewed-on: https://go-review.googlesource.com/c/go/+/450708
Reviewed-by: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Bryan Mills <bcmills@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>

2 years agocmd/pack: use testenv.Command instead of exec.Command in tests
Bryan C. Mills [Tue, 15 Nov 2022 15:45:00 +0000 (10:45 -0500)]
cmd/pack: use testenv.Command instead of exec.Command in tests

testenv.Command sets a default timeout based on the test's deadline
and sends SIGQUIT (where supported) in case of a hang.

Change-Id: Ia1dd1bacfa2c9378401bbce304ebf7ee3f159352
Reviewed-on: https://go-review.googlesource.com/c/go/+/450707
Run-TryBot: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Bryan Mills <bcmills@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
2 years agocmd/pack: use the test binary as 'pack' instead of rebuilding it
Bryan C. Mills [Tue, 15 Nov 2022 15:44:36 +0000 (10:44 -0500)]
cmd/pack: use the test binary as 'pack' instead of rebuilding it

This not only reduces the latency of the test, but also respects
build flags like '-race' and '-cover' passed to the 'go test' command.

Change-Id: Ifaf396465f54ad0a28977b6c0a4b87dbc359de37
Reviewed-on: https://go-review.googlesource.com/c/go/+/450706
Reviewed-by: Ian Lance Taylor <iant@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>
Auto-Submit: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agogo/internal/gcimporter: backport flattenImports optimization
Matthew Dempsky [Tue, 15 Nov 2022 19:08:03 +0000 (11:08 -0800)]
go/internal/gcimporter: backport flattenImports optimization

Backport of go.dev/cl/450755 from the x/tools importer.

Change-Id: Ia9a2cc038d589a4cde1b31f74484a5f88dabeef9
Reviewed-on: https://go-review.googlesource.com/c/go/+/450795
Auto-Submit: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Florian Zenker <floriank@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: David Chase <drchase@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/objdump: use testenv.Command instead of exec.Command in tests
Bryan C. Mills [Tue, 15 Nov 2022 15:38:07 +0000 (10:38 -0500)]
cmd/objdump: use testenv.Command instead of exec.Command in tests

testenv.Command sets a default timeout based on the test's deadline
and sends SIGQUIT (where supported) in case of a hang.

Change-Id: I1965ea453af6aa9eeae9669065deb0b372dc6caf
Reviewed-on: https://go-review.googlesource.com/c/go/+/450705
Run-TryBot: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Bryan Mills <bcmills@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
2 years agocmd/objdump: use the test binary as 'objdump' instead of rebuilding it
Bryan C. Mills [Tue, 15 Nov 2022 15:37:22 +0000 (10:37 -0500)]
cmd/objdump: use the test binary as 'objdump' instead of rebuilding it

This not only reduces the latency of the test, but also respects
build flags like '-race' and '-cover' passed to the 'go test' command.

Change-Id: Icd22062ab75964a74d011c81ea6c99be80bece18
Reviewed-on: https://go-review.googlesource.com/c/go/+/450704
Run-TryBot: Bryan Mills <bcmills@google.com>
Auto-Submit: Bryan Mills <bcmills@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/nm: use the test binary as 'nm' instead of rebuilding it
Bryan C. Mills [Tue, 15 Nov 2022 15:30:10 +0000 (10:30 -0500)]
cmd/nm: use the test binary as 'nm' instead of rebuilding it

This not only reduces the latency of the test, but also respects
build flags like '-race' and '-cover' passed to the 'go test' command.

Change-Id: Iffdc60d444a9ff1d4ff5e688bca1c2ef0dfa03c8
Reviewed-on: https://go-review.googlesource.com/c/go/+/450703
Run-TryBot: Bryan Mills <bcmills@google.com>
Auto-Submit: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
2 years agocmd/link: use testenv.Command instead of exec.Command in tests
Bryan C. Mills [Tue, 15 Nov 2022 15:21:21 +0000 (10:21 -0500)]
cmd/link: use testenv.Command instead of exec.Command in tests

testenv.Command sets a default timeout based on the test's deadline
and sends SIGQUIT (where supported) in case of a hang.

Change-Id: I25a67aa5793c6fa977ff6d4dc59ca1533bf58d41
Reviewed-on: https://go-review.googlesource.com/c/go/+/450702
Run-TryBot: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Auto-Submit: Bryan Mills <bcmills@google.com>

2 years agocmd/internal/goobj: use testenv.Command instead of exec.Command in tests
Bryan C. Mills [Tue, 15 Nov 2022 15:14:41 +0000 (10:14 -0500)]
cmd/internal/goobj: use testenv.Command instead of exec.Command in tests

testenv.Command sets a default timeout based on the test's deadline
and sends SIGQUIT (where supported) in case of a hang.

Change-Id: I464dc34b50f3360123aca9e8666df9799c15e457
Reviewed-on: https://go-review.googlesource.com/c/go/+/450701
Run-TryBot: Bryan Mills <bcmills@google.com>
Auto-Submit: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
2 years agocmd/internal/archive: use testenv.Command instead of exec.Command in tests
Bryan C. Mills [Tue, 15 Nov 2022 15:14:28 +0000 (10:14 -0500)]
cmd/internal/archive: use testenv.Command instead of exec.Command in tests

testenv.Command sets a default timeout based on the test's deadline
and sends SIGQUIT (where supported) in case of a hang.

Change-Id: I6f8391da50e74919ccc59cd43ccae99affab8798
Reviewed-on: https://go-review.googlesource.com/c/go/+/450700
Auto-Submit: Bryan Mills <bcmills@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>

2 years agocmd/internal/obj: use testenv.Command instead of exec.Command in tests
Bryan C. Mills [Tue, 15 Nov 2022 15:14:16 +0000 (10:14 -0500)]
cmd/internal/obj: use testenv.Command instead of exec.Command in tests

testenv.Command sets a default timeout based on the test's deadline
and sends SIGQUIT (where supported) in case of a hang.

Change-Id: Ica1a9985f9abb1935434367c9c8ba28fc50f331d
Reviewed-on: https://go-review.googlesource.com/c/go/+/450699
Auto-Submit: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Run-TryBot: Bryan Mills <bcmills@google.com>