Joel Sing [Wed, 11 Nov 2020 17:30:15 +0000 (04:30 +1100)]
cmd/link/internal/ld: dedup shared libraries on openbsd
When linking internally on OpenBSD, dedup libraries treating versioned
and unversioned libraries as equivalents. Versioned libraries are preferred
and are retained over unversioned libraries.
This avoids the situation where the use of cgo results in a DT_NEEDED for a
versioned library (for example, libc.so.96.1), while a dynamic import
specifies an unversioned library (for example, libc.so). Without deduplication
this would result in two DT_NEEDED entries, causing a failure when ld.so
attempts to load the Go binrary.
Updates #36435
Fixes #39257
Change-Id: I4a4942f259dece01d97bb51df9e13d67c9f94d34
Reviewed-on: https://go-review.googlesource.com/c/go/+/249978
Trust: Joel Sing <joel@sing.id.au>
Run-TryBot: Joel Sing <joel@sing.id.au>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Cherry Zhang <cherryyz@google.com>
Cherry Zhang [Sat, 14 Nov 2020 02:08:26 +0000 (21:08 -0500)]
runtime: declare arg size/map for race version of sync/atomic functions
The argument size and map are used in stack scanning if those
functions are deferred. Declare the right argument size and map
so they can be scanned correctly.
A typo caused the validation rule to check against -WL,-O... which is
not a regular flag because the L should be lowercase as in the other
rules. This caused valid linker flags to be rejected and people had to
work around this by filtering their default flags that include, e.g.,
-Wl,-O1 for a simple link optimization.
Fix the typo that wrongly rejected -Wl,-O... but allowed a non-existing
-WL,-O flag.
Change-Id: Ia3bf730f16f5ad98a39d7f17159de17b44075462
GitHub-Last-Rev: 2ec7f2a2b92c9c76e707a71f6c5273aa1d512006
GitHub-Pull-Request: golang/go#42631
Reviewed-on: https://go-review.googlesource.com/c/go/+/270278
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Jay Conrod <jayconrod@google.com> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Trust: Jay Conrod <jayconrod@google.com>
David Chase [Fri, 13 Nov 2020 21:54:48 +0000 (16:54 -0500)]
cmd/compile: fix load of interface{}-typed OpIData in expand_calls
In certain cases, the declkared type of an OpIData is interface{}.
This was not expected (since interface{} is a pair, right?) and
thus caused a crash. What is intended is that these be treated as
a byteptr, so do that instead (this is what happens in 1.15).
Fixes #42568.
Change-Id: Id7c9e5dc2cbb5d7c71c6748832491ea62b0b339f
Reviewed-on: https://go-review.googlesource.com/c/go/+/270057
Trust: David Chase <drchase@google.com>
Run-TryBot: David Chase <drchase@google.com> Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Ian Lance Taylor [Fri, 13 Nov 2020 19:05:37 +0000 (11:05 -0800)]
cmd/go: permit CGO_LDFLAGS to appear in //go:ldflag
Fixes #42565
Change-Id: If7cf39905d124dbd54dfac6a53ee38270498efed
Reviewed-on: https://go-review.googlesource.com/c/go/+/269818
Trust: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Jay Conrod <jayconrod@google.com>
Austin Clements [Wed, 11 Nov 2020 20:33:39 +0000 (15:33 -0500)]
cmd/cgo: fix initialization of empty argument types
CL 258938 changed the way C to Go calls work such that they now
construct a C struct on the C side for the arguments and space for the
results. Any pointers in the result space must be zeroed, so we just
zero the whole struct.
However, C makes it surprisingly hard to robustly zero any struct
type. We had used a "{0}" initializer, which works in the vast
majority of cases, but fails if the type is empty or effectively
empty.
This CL fixes this by changing how the cgo tool zero-initializes the
argument struct to be more robust.
Fixes #42495.
Change-Id: Id1749b9d751e59eb7a02a9d44fec0698a2bf63cd
Reviewed-on: https://go-review.googlesource.com/c/go/+/269337
Trust: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Cherry Zhang <cherryyz@google.com> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Daniel S Fava [Thu, 12 Nov 2020 16:25:47 +0000 (17:25 +0100)]
runtime: swap the order of raceacquire() and racerelease()
In chansend() and chanrecv() of chan.go, the order of calls to
raceacquire() and racerelease() was swapped, which meant that the
code was not following the memory model "by the letter of the law."
Similar for bufrecv and bufsend in select.go
The memory model says:
- A send happens before the corresponding receive completes, and
- the kth receive on a channel with capacity C happens before the
k+C send on that channel completes.
The operative word here is "completes." For example, a sender obtains
happens-before information on completion of the send-operation, which
means, after the sender has deposited its message onto the channel.
Similarly for receives.
If the order of raceacquire() and racerelease() is incorrect, the race
detector may fail to report some race conditions.
The fix is minimal from the point of view of Go. The fix does, however,
rely on a new function added to TSan:
https://reviews.llvm.org/D76322
This commit only affects execution when race detection is enabled.
Added two tests into `runtime/race/output_test.go`:
- `chanmm` tests for the issue addressed by this patch
- `mutex` is a test for inverted semaphores, which must not be broken
by this (or any other) patch
Fixes #37355
Change-Id: I5e886879ead2bd456a4b7dd1d17253641b767f63
Reviewed-on: https://go-review.googlesource.com/c/go/+/220419
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Dmitri Shuralyov <dmitshur@golang.org> Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
Joel Sing [Mon, 2 Nov 2020 15:11:51 +0000 (02:11 +1100)]
syscall: use correct type for TIOCSPGRP/TIOCGPGRP
These ioctls take a pid_t (generally a C integer aka int32) and not an int64 - we
currently get away with this on little endian 64 bit platforms, since the bytes
fall into the correct place, however this breaks on big endian 64 bit platforms
(like openbsd/mips64).
Update #40995
Change-Id: I622a0543fd562d97f76a7376a84fd2641e6d6a24
Reviewed-on: https://go-review.googlesource.com/c/go/+/267605
Trust: Joel Sing <joel@sing.id.au>
Run-TryBot: Joel Sing <joel@sing.id.au>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Michael Matloob [Fri, 30 Oct 2020 20:47:58 +0000 (16:47 -0400)]
cmd/go: use overlaid path contents in build cache
When caching actions, use the overlaid file contents, because those
are the ones actually used to produce the outputs.
For #39958
Change-Id: Ia1f85b2fcf1f26e3b5be82f4d35c2726b134a36b
Reviewed-on: https://go-review.googlesource.com/c/go/+/266720
Trust: Michael Matloob <matloob@golang.org>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Bryan C. Mills <bcmills@google.com> Reviewed-by: Jay Conrod <jayconrod@google.com>
Michael Matloob [Thu, 29 Oct 2020 23:28:07 +0000 (19:28 -0400)]
cmd/go: pass in overlaid paths for .s files
This change adds support for adding overlays on assembly files.
For #39958
Change-Id: I1a328656199cc836f48e16de1ffd944fdd07fb39
Reviewed-on: https://go-review.googlesource.com/c/go/+/266417
Trust: Michael Matloob <matloob@golang.org>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Bryan C. Mills <bcmills@google.com> Reviewed-by: Jay Conrod <jayconrod@google.com>
Michael Matloob [Thu, 29 Oct 2020 23:05:56 +0000 (19:05 -0400)]
cmd/go: pass in overlaid file paths to C compiler
This change moves the code in work.(*Builder).cgo that, when there is
an overlay, copies non-Go files to objdir into work.(*Builder).Build,
and creates an overlay structure mapping from the nominal file paths
into the copies in objdir. That's propagated through to
work.(*Builder).ccompile, which will use it to pass in the path to the
overlaid contents in objdir when calling the compiler.
This allows for overlays of C/C++/Fortran files.
For #39958
Change-Id: I9a2e3d3ba6afdf7ce19be1dbf4eee34805cdc05f
Reviewed-on: https://go-review.googlesource.com/c/go/+/266376
Trust: Michael Matloob <matloob@golang.org>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Jay Conrod <jayconrod@google.com> Reviewed-by: Bryan C. Mills <bcmills@google.com>
Ian Lance Taylor [Fri, 6 Nov 2020 17:38:38 +0000 (09:38 -0800)]
cmd/go, cmd/cgo: don't let bogus symbol set cgo_ldflag
A hand-edited object file can have a symbol name that uses newline and
other normally invalid characters. The cgo tool will generate Go files
containing symbol names, unquoted. That can permit those symbol names
to inject Go code into a cgo-generated file. If that Go code uses the
//go:cgo_ldflag pragma, it can cause the C linker to run arbitrary
code when building a package. If you build an imported package we
permit arbitrary code at run time, but we don't want to permit it at
package build time. This CL prevents this in two ways.
In cgo, reject invalid symbols that contain non-printable or space
characters, or that contain anything that looks like a Go comment.
In the go tool, double check all //go:cgo_ldflag directives in
generated code, to make sure they follow the existing LDFLAG restrictions.
Thanks to Imre Rad / https://www.linkedin.com/in/imre-rad-2358749b for
reporting this.
Fixes CVE-2020-28367
Change-Id: Ia1ad8f3791ea79612690fa7d26ac451d0f6df7c1
Reviewed-on: https://team-review.git.corp.google.com/c/golang/go-private/+/895832 Reviewed-by: Than McIntosh <thanm@google.com> Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-on: https://go-review.googlesource.com/c/go/+/269658
Trust: Katie Hockman <katie@golang.org>
Trust: Roland Shoemaker <roland@golang.org>
Run-TryBot: Katie Hockman <katie@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Roland Shoemaker <roland@golang.org>
Katie Hockman [Tue, 10 Nov 2020 20:54:12 +0000 (15:54 -0500)]
math/big: fix shift for recursive division
The previous s value could cause a crash
for certain inputs.
Will check in tests and documentation improvements later.
Thanks to the Go Ethereum team and the OSS-Fuzz project for reporting this.
Thanks to Rémy Oudompheng and Robert Griesemer for their help
developing and validating the fix.
Fixes CVE-2020-28362
Change-Id: Ibbf455c4436bcdb07c84a34fa6551fb3422356d3
Reviewed-on: https://team-review.git.corp.google.com/c/golang/go-private/+/899974 Reviewed-by: Roland Shoemaker <bracewell@google.com> Reviewed-by: Filippo Valsorda <valsorda@google.com>
Reviewed-on: https://go-review.googlesource.com/c/go/+/269657
Trust: Katie Hockman <katie@golang.org>
Trust: Roland Shoemaker <roland@golang.org>
Run-TryBot: Katie Hockman <katie@golang.org> Reviewed-by: Roland Shoemaker <roland@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Alessandro Arzilli [Tue, 10 Nov 2020 16:02:27 +0000 (17:02 +0100)]
cmd/compile: do not emit an extra debug_line entry for the end of seq addr
Uses DW_LNS_advance_pc directly, instead of calling putpclcdelta
because the latter will create a new debug_line entry for the end of
sequence address.
Fixes #42484
Change-Id: Ib6355605cac101b9bf37a3b4961ab0cee678a839
Reviewed-on: https://go-review.googlesource.com/c/go/+/268937
Trust: Than McIntosh <thanm@google.com>
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Than McIntosh <thanm@google.com> Reviewed-by: Cherry Zhang <cherryyz@google.com>
Federico Guerinoni [Sun, 8 Nov 2020 14:35:35 +0000 (15:35 +0100)]
bytes: add example for (*Buffer).Bytes
Change-Id: I49ac604530fff7928fa15de07563418b104da5e4
Reviewed-on: https://go-review.googlesource.com/c/go/+/268260
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Trust: Giovanni Bajo <rasky@develer.com>
Andrew G. Morgan [Tue, 10 Nov 2020 03:28:04 +0000 (19:28 -0800)]
syscall: improve TestSetuidEtc() /proc/ parsing against races
TestSetuidEtc() was failing sporadically on linux-ppc64. From the
three https://build.golang.org/ logs, it looked like the logged
errors could be associated with threads dying, but proc reads
were, in some way, racing with their demise.
Exploring ways to increase thread demise, revealed that races
of this type can happen on non-ppc64 systems, and that
os.IsNotExist(err) was not a sufficient error condition test
for a thread's status file disappearing. This change includes a
fix for that to.
The actual issue on linux-ppc64 appears to be tied to PID reaping
and reuse latency on whatever the build test environment is for
linux-ppc64-buildlet. I suspect this can happen on any linux
system, however, especially where the container has a limited PID
range.
The fix for this, limited to the test (the runtime syscall support
is unchanged), is to confirm that the Pid for the interrogated
thread's /proc/<TID>/status file confirms that it is still
associated with the test-process' PID.
linux-ppc64-buildlet:
go/bin/go test syscall -run=TestSetuidEtc -count=10000
ok syscall 104.285s
Fixes #42462
Change-Id: I55c84ab8361003570a405fa52ffec4949bf91113
Reviewed-on: https://go-review.googlesource.com/c/go/+/268717 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Tobias Klauser <tobias.klauser@gmail.com>
Bryan Boreham [Wed, 11 Nov 2020 13:59:59 +0000 (13:59 +0000)]
testing: reduce memory allocation in Helper
Store the PC instead of the string name of the function, and defer
that conversion until we need it.
Helper is still relatively expensive in CPU time (few hundred ns),
but memory allocation is now constant for a test rather than linear in
the number of times Helper is called.
benchstat:
name old time/op new time/op delta
TBHelper-4 1.30µs ±27% 0.53µs ± 1% -59.03% (p=0.008 n=5+5)
name old alloc/op new alloc/op delta
TBHelper-4 216B ± 0% 0B -100.00% (p=0.008 n=5+5)
name old allocs/op new allocs/op delta
TBHelper-4 2.00 ± 0% 0.00 -100.00% (p=0.008 n=5+5)
Change-Id: I6565feb491513815e1058637d086b0374fa94e19
GitHub-Last-Rev: c2329cf225dab6de7309af3583daa011bafb9a62
GitHub-Pull-Request: golang/go#38834
Reviewed-on: https://go-review.googlesource.com/c/go/+/231717
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org> Reviewed-by: Emmanuel Odeke <emmanuel@orijtech.com>
Trust: Emmanuel Odeke <emmanuel@orijtech.com>
Michael Matloob [Tue, 10 Nov 2020 18:59:48 +0000 (13:59 -0500)]
cmd/go: don't copy cgo files to objdir when overlay is present
This cl is a roll-forward of golang.org/cl/265758, which was rolled back
in golang.org/cl/268900. The changes made are removing cgofiles
from the list of files that are copied to objdir (because the cgofiles
themselves aren't actually provided to the compiler) and fixing test
cases to properly provide the overlay flag and to allow for paths with
backslashes (as in Windows).
The previous cl (golang.org/cl/262618) copied non-overlaid cgo files
to objdir, mostly to get around the issue that otherwise cgo-generated
files were written out with the wrong names (they'd get the base path
of the overlay file containing the replaced contents, instead of the
base path of the path whose contents are being replaced). So that CL
it would copy the files to objdir with the base path of the file
being replaced to circumvent that.
This CL changes cmd/go and cmd/cgo so that instead of copying
files, it passes the actual path of the file on disk either of
the original file (if it is not overlaid) or its replacement
file (if it is) as well as a flag --path_rewrite, newly added to
cmd/cgo, that specifies the actual original file path that corresponds
to the replaced files.
Updates #39958
Change-Id: Ia45b022f9d27cfce0f9ec6da5f3a9f53654c67b8
Reviewed-on: https://go-review.googlesource.com/c/go/+/269017
Trust: Michael Matloob <matloob@golang.org>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Bryan C. Mills <bcmills@google.com> Reviewed-by: Jay Conrod <jayconrod@google.com>
Russ Cox [Mon, 9 Nov 2020 14:25:05 +0000 (09:25 -0500)]
io/fs: fix reference to WalkFunc
The comment explains differences between WalkDirFunc and WalkFunc,
but when this code moved out of path/filepath, we forgot to change
the reference to be filepath.WalkFunc. Fix that.
(The text should not be deleted, because path/filepath does not
contain this type - WalkDirFunc - nor this text anymore.)
Pointed out by Carl Johnson on CL 243916 post-submit.
Dmitri Shuralyov [Sat, 31 Oct 2020 04:28:53 +0000 (00:28 -0400)]
all: update vendored dependencies for Go 1.16 release
The Go 1.16 code freeze has recently started. This is a time to update
all golang.org/x/... module versions that contribute packages to the
std and cmd modules in the standard library to latest master versions.
Those versions have already gone through code review, and now they
will undergo additional testing during the upcoming freeze period.
If new issues in these dependencies are discovered, we have the freeze
period to address them. By the end of the freeze period, we will have
confidence that the Go 1.16 release and the dependency versions it has
selected are robust.
The dependency module versions that are selected in this commit are:
This change was created with a program from CL 256357 patch set 3
(which updates golang.org/x modules only) and the latest bundle tool,
but replacing golang.org/x/net version with a slightly older commit
golang/net@28c70e62bb1d140c3f2579fb7bb5095134d9cb1e due to #42498:
$ updatestd -goroot=$HOME/gotip -branch=master
> go version
go version devel +ecc3f5112e Thu Nov 5 23:21:33 2020 +0000 darwin/amd64
> go env GOROOT
/Users/dmitshur/gotip
> go version -m /Users/dmitshur/go/bin/bundle
/Users/dmitshur/go/bin/bundle: go1.15.4
path golang.org/x/tools/cmd/bundle
mod golang.org/x/tools v0.0.0-20201110201400-7099162a900a h1:5E6TPwSBG74zT8xSrVc8W59K4ch4NFobVTnh2BYzHyU=
dep golang.org/x/mod v0.3.0 h1:RM4zey1++hCTbCVQfnWeKs9/IEsaBLA8vTkd0WVtmH4=
dep golang.org/x/xerrors v0.0.0-20200804184101-5ec99f83aff1 h1:go1bK/D/BFZV2I8cIQd1NKEZ+0owSTG1fDTci4IqFcE=
Daniel Martí [Tue, 27 Oct 2020 18:12:03 +0000 (18:12 +0000)]
cmd/go: introduce the GOVERSION env variable
This is an extra variable available via 'go env', but not read from the
user's environment. It corresponds to the same string that
runtime.Version returns, assuming a program is built by the same version
of Go.
It's similar to the output of 'go version', but without the "go version"
prefix nor the "$GOOS/$GOARCH" suffix.
The main use case here is tools, which often use 'go env' to query basic
information about the installed Go tree. Its version was one missing
piece of information, which required an extra call to 'go version'
before this change.
Fixes #41116.
Change-Id: I5c9d8c2ba856c816c9f4c462ba73c907b3441445
Reviewed-on: https://go-review.googlesource.com/c/go/+/265637
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Jay Conrod <jayconrod@google.com> Reviewed-by: Russ Cox <rsc@golang.org>
Trust: Jay Conrod <jayconrod@google.com>
Trust: Daniel Martí <mvdan@mvdan.cc>
Michael Anthony Knyszek [Tue, 10 Nov 2020 15:35:43 +0000 (15:35 +0000)]
runtime: add lock rank partial order edge pollDesc -> spanSetSpine
This change adds a missing partial order edge. This edge captures of
the case of `wakep` getting called in `wakeNetPoller` which may then
allocate.
Fixes #42461.
Change-Id: Ie67d868e9cd24ed3cc94381dbf8a691dd13f068d
Reviewed-on: https://go-review.googlesource.com/c/go/+/268858
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Michael Pratt <mpratt@google.com>
Michael Anthony Knyszek [Tue, 10 Nov 2020 15:32:59 +0000 (15:32 +0000)]
runtime: add lock rank partial order edge sweep -> mspanSpecial
This change adds a missing partial order edge. The edge captures the
case where the background sweeper handles some specials (i.e. finalizers
or memory profile sampling) and is otherwise correct.
Fixes #42472.
Change-Id: Ic45f6cc1635fd3d6bc6c91ff6f64d436088cef33
Reviewed-on: https://go-review.googlesource.com/c/go/+/268857
Trust: Michael Knyszek <mknyszek@google.com>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Dan Scales <danscales@google.com>
Michael Matloob [Tue, 27 Oct 2020 21:51:58 +0000 (17:51 -0400)]
cmd/go: don't copy cgo files to objdir when overlay is present
The previous cl (golang.org/cl/262618) copied non-overlaid cgo files
to objdir, mostly to get around the issue that otherwise cgo-generated
files were written out with the wrong names (they'd get the base path
of the overlay file containing the replaced contents, instead of the
base path of the path whose contents are being replaced). So that CL
it would copy the files to objdir with the base path of the file
being replaced to circumvent that.
This CL changes cmd/go and cmd/cgo so that instead of copying
files, it passes the actual path of the file on disk either of
the original file (if it is not overlaid) or its replacement
file (if it is) as well as a flag --path_rewrite, newly added to
cmd/cgo, that specifies the actual original file path that corresponds
to the replaced files.
Updates #39958
Change-Id: Ic4aae5ef77fe405011fcdce7f6c162488d13daa2
Reviewed-on: https://go-review.googlesource.com/c/go/+/265758
Trust: Michael Matloob <matloob@golang.org>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Jay Conrod <jayconrod@google.com> Reviewed-by: Bryan C. Mills <bcmills@google.com>
SYNC is supposed to correspond to 'fast-BCR-serialization' which is
encoded as 'bcr 14,0'. In CL 197178 I accidentally modified the
encoding to 'bcr 7,0' which is a no-op. This CL reverses that change.
Fixes #42479.
Change-Id: I9918d93d720f5e12acc3014cde20d2d32cc87ee5
Reviewed-on: https://go-review.googlesource.com/c/go/+/268797
Run-TryBot: Michael Munday <mike.munday@ibm.com>
Trust: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Cherry Zhang <cherryyz@google.com>
A major rework of the kernel implementation occurred in 5.3. Areas
of the API that weren't clearly defined were clarified and the API
bounds are much more strictly checked than on earlier kernels.
Applications should target the behaviour and requirements of 5.3
kernels.
Rather than attempting to detect the file system for source and
destination files (which means two additional statfs syscalls) and skip
copy_file_range in case of known defects (e.g. CIFS -> CIFS), just
assume copy_file_range to be broken on kernels < 5.3.
Fixes #42400
Change-Id: I3a531296182c1d6e341772cc9d2be5bf83e52575
Reviewed-on: https://go-review.googlesource.com/c/go/+/268338
Trust: Tobias Klauser <tobias.klauser@gmail.com>
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Bryan C. Mills [Tue, 10 Nov 2020 02:27:35 +0000 (21:27 -0500)]
cmd/go: update test_race_install expected output for CL 266368
test_race_install checks that 'go test -i -race …' does not rebuild
already installed packages, by also passing '-v' and verifying that no
package names are printed to stderr.
CL 266368 added a deprecation message for the '-i' flag that caused
the stderr output to be non-empty, although it still does not print
any package names.
Updates #41696
Change-Id: I13e10e49b7c33139be9b13f24cb393c9f58fd85d
Reviewed-on: https://go-review.googlesource.com/c/go/+/268581
Trust: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Roland Shoemaker [Fri, 16 Oct 2020 01:32:20 +0000 (18:32 -0700)]
crypto/tls: de-prioritize AES-GCM ciphers when lacking hardware support
When either the server or client are lacking hardware support for
AES-GCM ciphers, indicated by the server lacking the relevant
instructions and by the client not putting AES-GCM ciphers at the top
of its preference list, reorder the preference list to de-prioritize
AES-GCM based ciphers when they are adjacent to other AEAD ciphers.
Also updates a number of recorded openssl TLS tests which previously
only specified TLS 1.2 cipher preferences (using -cipher), but not
TLS 1.3 cipher preferences (using -ciphersuites), to specify both
preferences, making these tests more predictable.
Fixes #41181.
Change-Id: Ied896c96c095481e755aaff9ff0746fb4cb9568e
Reviewed-on: https://go-review.googlesource.com/c/go/+/262857
Run-TryBot: Roland Shoemaker <roland@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Filippo Valsorda <filippo@golang.org>
Trust: Roland Shoemaker <roland@golang.org>
Trust: Katie Hockman <katie@golang.org>
Reason for revert: Some edge cases not properly covered due to changes within runtime traceback generation since 2017, that need to be examined. This change landed very late in the Go1.16 cycle.
Change-Id: I8cf6f46ea0ef6161d878e79943e6c7cdac94bccf
Reviewed-on: https://go-review.googlesource.com/c/go/+/268577
Trust: Emmanuel Odeke <emmanuel@orijtech.com>
Run-TryBot: Emmanuel Odeke <emmanuel@orijtech.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
Filippo Valsorda [Mon, 9 Nov 2020 13:54:55 +0000 (14:54 +0100)]
crypto/x509: drop the cgo implementation of root_darwin_amd64.go
This code was preserved just to do side-by-side testing while
transitioning to the Go implementation. There haven't been mismatch
issues, so drop the cgo code, which was making it hard to improve the Go
code without diverging.
Filippo Valsorda [Wed, 24 Jun 2020 21:01:00 +0000 (17:01 -0400)]
crypto/tls: ensure the server picked an advertised ALPN protocol
This is a SHALL in RFC 7301, Section 3.2.
Also some more cleanup after NPN, which worked the other way around
(with the possibility that the client could pick a protocol the server
did not suggest).
Change-Id: I83cc43ca1b3c686dfece8315436441c077065d82
Reviewed-on: https://go-review.googlesource.com/c/go/+/239748
Run-TryBot: Filippo Valsorda <filippo@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Filippo Valsorda <filippo@golang.org>
Trust: Roland Shoemaker <roland@golang.org> Reviewed-by: Roland Shoemaker <roland@golang.org>
Michael Matloob [Tue, 3 Nov 2020 18:12:19 +0000 (13:12 -0500)]
cmd/go: support cgo files in overlays
This is a roll-forward of golang.org/cl/267197, which was reverted in
golang.org/cl/267357. It makes the following changes in addition to
the ones in the next paragraph: It avoids outputting trimpath
arguments for an overlay unless the overlay affects the package being
compiled (to avoid hitting windows command line argument limits), and
it fixes processing of regexps in the script test framework to treat
the first *non flag* argument to grep, stdout, and stderr as a regexp,
not just the first argument.
golang.org/cl/267917 was a roll-forward of golang.org/cl/262618, which
was reverted in golang.org/cl/267037. The only differences between
this CL and the original were the three calls to fflush from the C
files in build_overlay.txt, to guarantee that the string we were
expecting was
actually written out.
The CL requires rewriting the paths of the files passed to the cgo
tool toolchain to use the overlaid paths instead of the disk paths of
files. Because the directories of the overlaid paths don't exist in
general, the cgo tool have been updated to run in base.Cwd instead of
the package directory.
For #39958
Change-Id: I1bd96db257564bcfd95b3502aeca14d04bd28618
Reviewed-on: https://go-review.googlesource.com/c/go/+/267797
Trust: Michael Matloob <matloob@golang.org>
Trust: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Bryan C. Mills <bcmills@google.com>
Filippo Valsorda [Mon, 31 Aug 2020 21:09:57 +0000 (17:09 -0400)]
crypto/tls: drop macFunction abstraction
Since we dropped SSLv3, there is only one MAC scheme, and it doesn't
need any state beyond a keyed HMAC, so we can replace the macFunction
with the hash.Hash it wraps.
Pointed out by mtp@.
Change-Id: I5545be0e6ccb34a3055fad7f6cb5f628ff748e9f
Reviewed-on: https://go-review.googlesource.com/c/go/+/251859
Run-TryBot: Filippo Valsorda <filippo@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Roland Shoemaker <roland@golang.org>
Trust: Roland Shoemaker <roland@golang.org>
Trust: Filippo Valsorda <filippo@golang.org>
Johan Brandhorst [Sat, 1 Aug 2020 11:18:31 +0000 (12:18 +0100)]
crypto/tls: add HandshakeContext method to Conn
Adds the (*tls.Conn).HandshakeContext method. This allows
us to pass the context provided down the call stack to
eventually reach the tls.ClientHelloInfo and
tls.CertificateRequestInfo structs.
These contexts are exposed to the user as read-only via Context()
methods.
This allows users of (*tls.Config).GetCertificate and
(*tls.Config).GetClientCertificate to use the context for
request scoped parameters and cancellation.
Replace uses of (*tls.Conn).Handshake with (*tls.Conn).HandshakeContext
where appropriate, to propagate existing contexts.
Fixes #32406
Change-Id: I33c228904fe82dcf57683b63627497d3eb841ff2
Reviewed-on: https://go-review.googlesource.com/c/go/+/246338
Run-TryBot: Filippo Valsorda <filippo@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Roland Shoemaker <roland@golang.org> Reviewed-by: Filippo Valsorda <filippo@golang.org>
Koen [Tue, 13 Oct 2020 23:05:43 +0000 (01:05 +0200)]
crypto/x509: return additional chains from Verify on Windows
Previously windows only returned the certificate-chain with the highest quality.
This change makes it so chains with a potentially lower quality
originating from other root certificates are also returned by verify.
Tests in verify_test flagged with systemLax are now allowed to pass if the system returns additional chains
Roland Shoemaker [Sun, 10 May 2020 18:23:58 +0000 (11:23 -0700)]
crypto/x509: add additional convenience fields to CertificateRequest
Adds the following additional convenience fields to CertificateRequest:
* KeyUsage
* ExtKeyUsage
* UnknownExtKeyUsage
* IsCA
* MaxPathLen
* BasicConstraintsValid
* MaxPathLenZero
* SubjectKeyId
* PolicyIdentifier
These fields are parsed during ParseCertificateRequest and marshalled
during CreateCertificateRequest. The parsing/marshalling code is
factored out of parseCertificate and buildExtensions (which is renamed
buildCertExtensions). This has the side effect of making these methods
somewhat easier to read.
Documentation for the fields is copied from Certificate.
Example CSR created with all of these fields parsed with openssl:
$ openssl req -in ~/test-csr.pem -noout -text
Certificate Request:
Data:
Version: 0 (0x0)
Subject:
Subject Public Key Info:
Public Key Algorithm: id-ecPublicKey
Public-Key: (256 bit)
pub:
04:a4:cb:64:35:8e:dd:8c:2b:a6:f1:aa:39:d1:be:
d0:b9:95:1e:59:19:82:76:28:d3:85:1b:c6:88:62:
e1:15:33:be:26:18:80:14:fe:f4:d4:91:66:4e:a4:
a4:47:bd:53:db:f7:2e:e3:31:ce:5f:86:cb:92:59:
93:bb:d0:7f:a2
ASN1 OID: prime256v1
NIST CURVE: P-256
Attributes:
Requested Extensions:
X509v3 Key Usage: critical
Certificate Sign
X509v3 Extended Key Usage:
Any Extended Key Usage, 1.2.3
X509v3 Basic Constraints: critical
CA:TRUE, pathlen:0
X509v3 Subject Key Identifier:
01:02:03
X509v3 Certificate Policies:
Policy: 1.2.3
Cherry Zhang [Sat, 7 Nov 2020 03:46:18 +0000 (22:46 -0500)]
cmd/compile: make sure linkname'd symbol is non-package
When a variable symbol is both imported (possibly through
inlining) and linkname'd, make sure its LSym is marked as
non-package for symbol indexing in the object file, so it is
resolved by name and dedup'd with the original definition.
Fixes #42401.
Change-Id: I8e90c0418c6f46a048945c5fdc06c022b77ed68d
Reviewed-on: https://go-review.googlesource.com/c/go/+/268178
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Than McIntosh <thanm@google.com> Reviewed-by: Jeremy Faller <jeremy@golang.org>
Russ Cox [Thu, 29 Oct 2020 14:57:38 +0000 (10:57 -0400)]
cmd/go: add GOVCS setting to control version control usage
The go command runs commands like git and hg to download modules.
In the past, we have had problems with security bugs in version
control systems becoming security bugs in “go get”.
The original modules draft design removed use of these commands
entirely, saying:
> We want to move away from invoking version control tools such as bzr,
> fossil, git, hg, and svn to download source code. These fragment the
> ecosystem: packages developed using Bazaar or Fossil, for example, are
> effectively unavailable to users who cannot or choose not to install
> these tools. The version control tools have also been a source of
> exciting security problems. It would be good to move them outside the
> security perimeter.
The removal of these commands was not possible in the end: being able
to fetch directly from Git repos is too important, especially for
closed source. But the security exposure has not gone away.
We remain vulnerable to problems in VCS systems, especially the less
scrutinized ones.
This change adds a GOVCS setting to let users control which version
control systems are allowed by default.
It also changes the default allowed version control systems to git and hg
for public code and any version control system for private code
(import path or module path matched by the GOPRIVATE setting).
See the changes in alldocs.go for detailed documentation.
See #41730 for proposal and discussion.
Fixes #41730.
[Replay of CL 266420. See changes from Patch Set 1 for updates to fix
a few long tests.]
Change-Id: I4fe93804548956c42aea985368b4571bdb220f48
Reviewed-on: https://go-review.googlesource.com/c/go/+/267888
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Dmitri Shuralyov [Sun, 18 Oct 2020 17:02:06 +0000 (17:02 +0000)]
cmd/go: add /v2 to another require example
This is the same change as in CL 144917, but applied to a nearby line.
For #28374.
Change-Id: I6e1693d3a14e2517d863d1052a06c1156fc1edd4
Reviewed-on: https://go-review.googlesource.com/c/go/+/263437
Trust: Dmitri Shuralyov <dmitshur@golang.org>
Run-TryBot: Dmitri Shuralyov <dmitshur@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Jay Conrod <jayconrod@google.com>
Alberto Donizetti [Sun, 8 Nov 2020 08:44:33 +0000 (09:44 +0100)]
test/codegen: go fmt
Fixes #42445
Change-Id: I9653ef094dba2a1ac2e3daaa98279d10df17a2a1
Reviewed-on: https://go-review.googlesource.com/c/go/+/268257
Trust: Alberto Donizetti <alb.donizetti@gmail.com>
Trust: Martin Möhrmann <moehrmann@google.com>
Run-TryBot: Alberto Donizetti <alb.donizetti@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Martin Möhrmann <moehrmann@google.com>
Austin Clements [Mon, 2 Nov 2020 13:48:23 +0000 (08:48 -0500)]
testing: increase benchmark output to four significant figures
Currently, the benchmark output from the testing package prints small
values with three significant figures. This means it can only
distinguish 1 part in 100, or a 1% error, which can be enough to throw
off further analysis of the output. This CL increases it to four
significant figures. For time values, at least, anything beyond four
significant figures is almost certainly noise.
Fixes #34626.
Change-Id: I3bcf305427130026276e6a4c78167989319f280c
Reviewed-on: https://go-review.googlesource.com/c/go/+/267102
Trust: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Michael Pratt <mpratt@google.com>
Brad Fitzpatrick [Sat, 7 Nov 2020 16:00:15 +0000 (08:00 -0800)]
crypto/x509: keep smaller root cert representation in memory until needed
Instead of parsing the PEM files and then storing the *Certificate
values forever, still parse them to see if they're valid and pick out
some fields, but then only store the decoded pem.Block.Bytes until
that cert is first needed.
Saves about 500K of memory on my (Debian stable) machine after doing a
tls.Dial or calling x509.SystemCertPool.
A more aggressive version of this is still possible: we can not keep
the pem.Block.Bytes in memory either, and re-read them from disk when
necessary. But dealing with files disappearing and even large
multi-cert PEM files changing (with offsets sliding around) made this
conservative version attractive. It doesn't change the
slurp-roots-on-startup semantics. It just does so with less memory
retained.
Change-Id: I3aea333f4749ae3b0026042ec3ff7ac015c72204
Reviewed-on: https://go-review.googlesource.com/c/go/+/230025
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Roland Shoemaker <roland@golang.org> Reviewed-by: Filippo Valsorda <filippo@golang.org>
crypto/x509: add support for CertPool to load certs lazily
This will allow building CertPools that consume less memory. (Most
certs are never accessed. Different users/programs access different
ones, but not many.)
This CL only adds the new internal mechanism (and uses it for the
old AddCert) but does not modify any existing root pool behavior.
(That is, the default Unix roots are still all slurped into memory as
of this CL)
Change-Id: Ib3a42e4050627b5e34413c595d8ced839c7bfa14
Reviewed-on: https://go-review.googlesource.com/c/go/+/229917
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Brad Fitzpatrick <bradfitz@golang.org>
Trust: Roland Shoemaker <roland@golang.org> Reviewed-by: Filippo Valsorda <filippo@golang.org> Reviewed-by: Roland Shoemaker <roland@golang.org>
Joshua M. Clulow [Fri, 6 Nov 2020 21:11:58 +0000 (13:11 -0800)]
os/exec: use "pfiles" for fd debugging on illumos
On illumos (and Solaris) systems, the native "pfiles" tool provides the
best information about open file descriptors for a process:
https://illumos.org/man/1/pfiles
Use that instead of "lsof" when debugging file descriptor leaks.
Updates #42431.
Change-Id: If1250c4e6c9e8adbd076495a09fb1ce63abcc68b
Reviewed-on: https://go-review.googlesource.com/c/go/+/268019
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org> Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Trust: Dmitri Shuralyov <dmitshur@golang.org>
Cholerae Hu [Fri, 31 Jul 2020 05:57:48 +0000 (13:57 +0800)]
cmd/compile: check indirect connection between if block and phi block in addLocalInductiveFacts
CL 244579 added guard clauses to prevent a faulty state that was
possible under the incorrect logic of the uniquePred loop in
addLocalInductiveFacts. That faulty state was still making the
intended optimization, but not for the correct reason.
Removing the faulty state also removed the overly permissive application
of the optimization, and therefore made these two tests fail.
We disabled the tests of this optimization in CL 244579 to allow us to
quickly apply the fix in the CL. This CL now corrects the logic of the
uniquePred loop in order to apply the optimization correctly.
The comment above the uniquePred loop says that it will follow unique
predecessors until it reaches a join point. Without updating the child
node on each iteration, it cannot follow the chain of unique
predecessors more than one step. Adding the update to the child node
on each iteration of the loop allows the logic to follow the chain of
unique predecessors until reaching a join point (because a non-unique
predecessor will signify a join point).
Updates #40502.
Change-Id: I23d8367046a2ab3ce4be969631f9ba15dc533e6c
Reviewed-on: https://go-review.googlesource.com/c/go/+/246157
Run-TryBot: Dmitri Shuralyov <dmitshur@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: David Chase <drchase@google.com>
Trust: Dmitri Shuralyov <dmitshur@golang.org>
Artyom Pervukhin [Wed, 4 Nov 2020 12:28:25 +0000 (15:28 +0300)]
archive/zip: fix documentation to mention fs.FS interface
Fixes #42374
Change-Id: I0ed1eb052d79bcc65810b74bff48f1e615e1dc1e
Reviewed-on: https://go-review.googlesource.com/c/go/+/267657 Reviewed-by: Ian Lance Taylor <iant@golang.org> Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Emmanuel T Odeke [Sat, 18 Feb 2017 10:03:32 +0000 (03:03 -0700)]
runtime: make stack traces of endless recursion print only top and bottom 50
This CL makes it so that instead of printing massive stack traces during
endless recursion, which spams users and aren't useful, it now prints out
the top and bottom 50 frames. If the number of frames <= 100
(_TracebackMaxFrames), we'll just print all the frames out.
Modified gentraceback to return counts of:
* ntotalframes
* nregularframes
which allows us to get accurate counts of the various kinds of frames.
While here, also fixed a bug that resulted from CL 37222, in which we
no longer accounted for decrementing requested frame skips, and assumed
that when printing, that skip would always be 0. The fix is instead to add
precondition that we'll only print if skip <= 0, but also decrement skip
as we iterate.
Fixes #7181.
Fixes #24628.
Change-Id: Ie31ec6413fdfbe43827b254fef7d99ea26a5277f
Reviewed-on: https://go-review.googlesource.com/c/go/+/37222
Run-TryBot: Emmanuel Odeke <emmanuel@orijtech.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
Trust: Emmanuel Odeke <emmanuel@orijtech.com>
Matthew Dempsky [Thu, 29 Oct 2020 08:13:16 +0000 (01:13 -0700)]
cmd/compile: support inlining of type switches
This CL adds support for inlining type switches, including exporting
and importing them.
Type switches are represented mostly the same as expression switches.
However, if the type switch guard includes a short variable
declaration, then there are two differences: (1) there's an ONONAME
(in the OTYPESW's Left) to represent the overall pseudo declaration;
and (2) there's an ONAME (in each OCASE's Rlist) to represent the
per-case variables.
For simplicity, this CL simply writes out each variable separately
using iimport/iiexport's normal Vargen mechanism for disambiguating
identically named variables within a function. This could be improved
somewhat, but inlinable type switches are probably too uncommon to
merit the complexity.
While here, remove "case OCASE" from typecheck1. We only type check
"case" clauses as part of a "select" or "switch" statement, never as
standalone statements.
Fixes #37837
Change-Id: I8f42f6c9afdd821d6202af4a6bf1dbcbba0ef424
Reviewed-on: https://go-review.googlesource.com/c/go/+/266203
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
Trust: Matthew Dempsky <mdempsky@google.com>
This commit is a copy of filepath.WalkDir adapted to use fs.FS
instead of the native OS file system. It is the last implementation
piece of the io/fs proposal.
The original io/fs proposal was to adopt filepath.Walk, but we
have since introduced the more efficient filepath.WalkDir (#42027),
so this CL adopts that more efficient option instead.
(The changes in path/filepath bring the two copies more in line
with each other. The main change is unembedding the field
in statDirEntry, so that the fs.DirEntry passed to the WalkDirFunc
for the root of the tree does not have any extra methods.)
For #41190.
Change-Id: I9359dfcc110338c0ec64535f22cafb38d0b613a6
Reviewed-on: https://go-review.googlesource.com/c/go/+/243916
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Rob Pike <r@golang.org>
Russ Cox [Thu, 29 Oct 2020 16:20:53 +0000 (12:20 -0400)]
path/filepath: add WalkDir
WalkDir is like Walk but can use ReadDir to read directories,
instead of Readdirnames + Lstat on every entry,
which is usually a significant performance improvement.
(The Lstat can still happen if the walk function calls d.Info.)
Fixes #42027.
[Replay of CL 266240 after it was reverted due to accidentally
enabling on Windows a test that does not work on Windows.
The original code only ran the test on os.Getuid() > 0.
The rolled-back CL skipped the test on os.Getuid() == 0.
But on Windows, os.Getuid(), it turns out, always returns -1.
So what looked like a test for root was also excluding Windows.
This CL revises the test to skip Windows explicitly.]
Change-Id: I9b3661013d6449b7486532445d934ae91e5393ef
Reviewed-on: https://go-review.googlesource.com/c/go/+/267887
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Dmitri Shuralyov <dmitshur@golang.org> Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Michael Munday [Mon, 11 May 2020 16:44:48 +0000 (09:44 -0700)]
cmd/compile: optimize shift pairs and masks on s390x
Optimize combinations of left and right shifts by a constant value
into a 'rotate then insert selected bits [into zero]' instruction.
Use the same instruction for contiguous masks since it has some
benefits over 'and immediate' (not restricted to 32-bits, does not
overwrite source register).
To keep the complexity of this change under control I've only
implemented 64 bit operations for now.
There are a lot more optimizations that can be done with this
instruction family. However, since their function overlaps with other
instructions we need to be somewhat careful not to break existing
optimization rules by creating optimization dead ends. This is
particularly true of the load/store merging rules which contain lots
of zero extensions and shifts.
This CL does interfere with the store merging rules when an operand
is shifted left before it is stored:
binary.BigEndian.PutUint64(b, x << 1)
This is unfortunate but it's not critical and somewhat complex so
I plan to fix that in a follow up CL.
Change-Id: I7591cce41e06867ba10a745daae9333513062746
Reviewed-on: https://go-review.googlesource.com/c/go/+/233317
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
Trust: Michael Munday <mike.munday@ibm.com>
Joel Sing [Thu, 5 Nov 2020 05:41:32 +0000 (16:41 +1100)]
cmd/dist: remove openbsd/mips64 from incomplete ports
Remove openbsd/mips64 from incomplete ports lists - all of the necessary code
has landed and we want to run tests so we can see/deal with remaining failures.
Update #40995
Change-Id: I5d4f89af82ff3abe57570a9a8abf889498093d32
Reviewed-on: https://go-review.googlesource.com/c/go/+/267606
Trust: Joel Sing <joel@sing.id.au> Reviewed-by: Cherry Zhang <cherryyz@google.com>
Jonathan Swinney [Wed, 4 Nov 2020 16:18:23 +0000 (16:18 +0000)]
cmd/compile: improve atomic swap intrinsics on arm64
ARMv8.1 has added new instructions for atomic memory operations. This
change builds on the previous change which added support for atomic add, 0a7ac93c27c9ade79fe0f66ae0bb81484c241ae5, to include similar support for
atomic-compare-and-swap, atomic-swap, atomic-or, and atomic-and
intrinsics. Since the new instructions are not guaranteed to be present,
we guard their usages with a branch on a CPU feature.
Peformance on an ARMv8.1 machine:
name old time/op new time/op delta
CompareAndSwap-16 37.9ns ±16% 24.1ns ± 4% -36.44% (p=0.000 n=10+9)
CompareAndSwap64-16 38.6ns ±15% 24.1ns ± 3% -37.47% (p=0.000 n=10+10)
name old time/op new time/op delta
Swap-16 46.9ns ±32% 12.5ns ± 6% -73.40% (p=0.000 n=10+10)
Swap64-16 53.4ns ± 1% 12.5ns ± 6% -76.56% (p=0.000 n=10+10)
Than McIntosh [Thu, 5 Nov 2020 19:19:47 +0000 (14:19 -0500)]
cmd/link: report error if builtin referenced but not defined
When the compiler refers to a runtime builtin, it emits an indexed
symbol reference in the object file via predetermined/preassigned ID
within the PkgIdxBuiltin pseudo-package. At link time when the loader
encounters these references, it redirects them to the corresponding
defined symbol in the runtime package. This redirection process
currently assumes that if a runtime builtin is referenced, we'll
always have a definition for it. This assumption holds in most cases,
however for the builtins "runtime.racefuncenter" and
"runtime.racefuncexit", we'll only see definitions if the runtime
package we're linking against was built with "-race".
In the bug in question, build passes "-gcflags=-race" during
compilation of the main package, but doesn't pass "-race" directly to
'go build', and as a result the final link combines a
race-instrumented main with a non-race runtime; this results in R_CALL
relocations with zero-valued target symbols, resulting in a panic
during stack checking.
This patch changes the loader's resolve method to detect situations
where we're asking for builtin "runtime.X", but the runtime package
read in doesn't contain a definition for X.
Fixes #42396.
Change-Id: Iafd38bd3b0f7f462868d120ccd4d7d1b88b27436
Reviewed-on: https://go-review.googlesource.com/c/go/+/267881
Trust: Than McIntosh <thanm@google.com>
Run-TryBot: Than McIntosh <thanm@google.com> Reviewed-by: Jeremy Faller <jeremy@golang.org> Reviewed-by: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Rob Findley [Tue, 27 Oct 2020 02:44:54 +0000 (22:44 -0400)]
go/types: add unexported start and end positions to type checker errors
Tools often need to associate errors not with a single position, but
with a span of source code. For example, gopls currently estimates
diagnostic spans using heuristics to expand the positions reported by
the type checker to surrounding source code. Unfortunately this is often
inaccurate.
This CL lays the groundwork to solve this within go/types by adding a
start and end position to type checker errors. This is an experimental
API, both because we are uncertain of the ideal representation for these
spans and because their initial positioning is naive. In most cases this
CL simply expands errors to the surrounding ast.Node being typechecked,
if available. This might not be the best error span to present to the
user. For these reasons the API is unexported -- gopls can read these
positions using reflection, allowing us to gain experience and improve
them during the next development cycle.
For golang/go#42290
Change-Id: I39a04d70ea2bb2134b4d4c937f32b2ddb4456430
Reviewed-on: https://go-review.googlesource.com/c/go/+/265250
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Robert Griesemer <gri@golang.org>
Trust: Robert Findley <rfindley@google.com>
Trust: Robert Griesemer <gri@golang.org>
Michael Matloob [Tue, 3 Nov 2020 18:12:19 +0000 (13:12 -0500)]
cmd/go: account for flags when parsing regexps in TestScript
Test script expects the regexp argument for stdout, stderr, and cmp
to be the first argument after the command, but that might not be the
case if the -q or -count flags are provided. Treat the first argument
after a flag as a regexp instead.
For #39958
Change-Id: I369926109ec10cca8b2c3baca27e7a3f7baf364b
Reviewed-on: https://go-review.googlesource.com/c/go/+/267877
Trust: Michael Matloob <matloob@golang.org>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Bryan C. Mills <bcmills@google.com>
Michael Pratt [Mon, 2 Nov 2020 21:34:51 +0000 (16:34 -0500)]
runtime: disable preemption in startm
startm contains a critical section from when it takes ownership of a P
(either on function entry or call to pidleput) until it wakes the M
receiving the P. If preempted in this critical section, the owned P is
left in limbo. If preempted for a GC stop, there will be nothing to stop
the owned P and STW will wait forever.
golang.org/cl/232298 introduced the first call to startm that is not on
the system stack (via a wakep call), introducing the possibility of
preemption. Disable preemption in startm to ensure this remains
non-preemptible.
Since we're not always on the system stack anymore, we also need to be
careful in allocm.
Updates #42237
Change-Id: Icb95eef9eb262121856485316098331beea045da
Reviewed-on: https://go-review.googlesource.com/c/go/+/267257 Reviewed-by: Austin Clements <austin@google.com> Reviewed-by: Michael Knyszek <mknyszek@google.com>
Trust: Michael Pratt <mpratt@google.com>
cmd/go/internal/modget: resolve paths at the requested versions
Previously, we resolved each argument to 'go get' to a package path or
module path based on what was in the build list at existing versions,
even if the argument specified a different version explicitly. That
resulted in bugs like #37438, in which we variously resolved the wrong
version or guessed the wrong argument type for what is unambiguously a
package argument at the requested version.
We were also using a two-step upgrade/downgrade algorithm, which could
not only upgrade more that is strictly necessary, but could also
unintentionally upgrade *above* the requested versions during the
downgrade step.
This change instead uses an iterative approach, with an explicit
disambiguation step for the (rare) cases where an argument could match
the same package path in multiple modules. We use a hook in the
package loader to halt package loading as soon as an incorrect version
is found — preventing over-resolving — and verify that the result
after applying downgrades successfully obtained the requested versions
of all modules.
Making 'go get' be correct and usable is especially important now that
we are defaulting to read-only mode (#40728), for which we are
recommending 'go get' more heavily.
While I'm in here refactoring, I'm also reworking the API boundary
between the modget and modload packages. Previously, the modget
package edited the build list directly, and the modload package
accepted the edited build list without validation. For lazy loading
(#36460), the modload package will need to maintain additional
metadata about the requirement graph, so it needs tighter control over
the changes to the build list.
As of this change, modget no longer invokes MVS directly, but instead
goes through the modload package. The resulting API gives clearer
reasons in case of updates, which we can use to emit more useful
errors.
Fixes #37438
Updates #36460
Updates #40728
Change-Id: I596f0020f3795870dec258147e6fc26a3292c93a
Reviewed-on: https://go-review.googlesource.com/c/go/+/263267
Trust: Bryan C. Mills <bcmills@google.com>
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Russ Cox <rsc@golang.org>
Bryan C. Mills [Sat, 31 Oct 2020 01:00:41 +0000 (21:00 -0400)]
cmd/go/internal/modload: fix (*mvsReqs).Max when the second argument is the empty string
As far as I can tell, this bug had gone unnoticed because everything
that uses Max so far happened to only ever present the empty string as
the first argument.
For #37438
Change-Id: Ie8c42313157d2c2c17e4058c53b5bb026b95a1c1
Reviewed-on: https://go-review.googlesource.com/c/go/+/266860
Trust: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Michael Matloob <matloob@golang.org>
Bryan C. Mills [Fri, 30 Oct 2020 16:24:51 +0000 (12:24 -0400)]
cmd/go/internal/modload: return a module-only result from QueryPattern
This allows a single QueryPattern call to resolve a path that could be
either a package or a module. It is important to be able to make a
single QueryPattern call — rather than a QueryPattern followed by a
Query for the specific module path — to provide appropriate fallback
behavior: if the proxy returns package results but does not contain a
module result, we don't want to fall back to the next proxy to look
for the (probably-nonexistent) module.
For #37438
Change-Id: I419b8bb3ab4565f443bb5cee9a8b206f453b9801
Reviewed-on: https://go-review.googlesource.com/c/go/+/266657
Trust: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org> Reviewed-by: Jay Conrod <jayconrod@google.com>