After this CL, we can reorder or pad internal/abi.ITab fields at will
(keeping Fun last, and updating ITabTypeOff correctly) without
breaking anything.
Change-Id: Ib7bb5828519813e0d1aa36be5092f96fcd62b3be
Reviewed-on: https://go-review.googlesource.com/c/go/+/549516
Reviewed-by: Keith Randall <khr@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: David Chase <drchase@google.com>
CL 525637 added GOARM_x assembly macros based on GOARM value. But
it did not define the macro in cmd/dist, so the macro is not set
during bootstrapping. This CL defines them.
With CL 514907, cfg.GOARM can also take a soft/hardfloat suffix,
like "7,hardfloat". Handle that case.
For #65601.
Change-Id: I60ffe7e8b623ae693d91d6e8595067a6f76565b3
Reviewed-on: https://go-review.googlesource.com/c/go/+/562995
Run-TryBot: Cherry Mui <cherryyz@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
CL 525637 changed to the guard of DMB instruction from the
compiled-in runtime.goarm value to GOARM_7 macro and CPU feature
detection. It missed a place where runtime.goarm is loaded to a
register and reused later. This CL corrects the condition.
Fixes#65601.
Change-Id: I2ddefd03a1eb1048dbec0254c6e234c65b054279
Reviewed-on: https://go-review.googlesource.com/c/go/+/564855
Run-TryBot: Cherry Mui <cherryyz@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
Fixes#52792
Tested: Ran go test cmd/go
Change-Id: Ib7006256f4dca9e9fbfce266c00253c69595d6ab
Cq-Include-Trybots: luci.golang.try:gotip-linux-amd64-longtest,gotip-windows-amd64-longtest
Reviewed-on: https://go-review.googlesource.com/c/go/+/562775
Reviewed-by: Bryan Mills <bcmills@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Change-Id: I38bb68570f7f097a6dfeef1c75896396992d81f7
GitHub-Last-Rev: 10ba48a546
GitHub-Pull-Request: golang/go#53706
Reviewed-on: https://go-review.googlesource.com/c/go/+/416234
Reviewed-by: Than McIntosh <thanm@google.com>
Auto-Submit: Ian Lance Taylor <iant@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Pull in the latest published version of github.com/google/pprof
as part of the continuous process of keeping Go's dependencies
up to date. Done with:
go get github.com/google/pprof
go mod tidy
go mod vendor
For #36905.
Fixes#65741.
Change-Id: Ice7b085c03ff69be97929cbe47bfd91954907529
Cq-Include-Trybots: luci.golang.try:gotip-linux-amd64-longtest
Reviewed-on: https://go-review.googlesource.com/c/go/+/564636
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Cherry Mui <cherryyz@google.com>
When transitioning between the old object loader and the new
object loader, to support both we made loadelf to take symbol
loading functions as function pointers. Now we only have the new
object loader. Change the function pointers back to static calls.
Change-Id: Ia623a6010376a3d7c0be5eacae002144d956f28a
Reviewed-on: https://go-review.googlesource.com/c/go/+/564635
Reviewed-by: Than McIntosh <thanm@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Sometimes we found that benchmark results may strongly depend on
the ordering of functions laid out in the binary. This CL adds a
flag -randlayout=seed, which randomizes the function layout (in a
deterministic way), so can verify the benchmark results against
different function ordering.
Change-Id: I85f33881bbfd4ca6812fbd4bec00bf475755a09e
Reviewed-on: https://go-review.googlesource.com/c/go/+/562157
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Than McIntosh <thanm@google.com>
Enable canRotate for riscv64, enable rotation intrinsics and provide
better rewrite implementations for rotations. By avoiding Lsh*x64
and Rsh*Ux64 we can produce better code, especially for 32 and 64
bit rotations. By enabling canRotate we also benefit from the generic
rotation rewrite rules.
Benchmark on a StarFive VisionFive 2:
│ rotate.1 │ rotate.2 │
│ sec/op │ sec/op vs base │
RotateLeft-4 14.700n ± 0% 8.016n ± 0% -45.47% (p=0.000 n=10)
RotateLeft8-4 14.70n ± 0% 10.69n ± 0% -27.28% (p=0.000 n=10)
RotateLeft16-4 14.70n ± 0% 12.02n ± 0% -18.23% (p=0.000 n=10)
RotateLeft32-4 13.360n ± 0% 8.016n ± 0% -40.00% (p=0.000 n=10)
RotateLeft64-4 13.360n ± 0% 8.016n ± 0% -40.00% (p=0.000 n=10)
geomean 14.15n 9.208n -34.92%
Change-Id: I1a2036fdc57cf88ebb6617eb8d92e1d187e183b2
Reviewed-on: https://go-review.googlesource.com/c/go/+/560315
Reviewed-by: M Zhuo <mengzhuo1203@gmail.com>
Run-TryBot: Joel Sing <joel@sing.id.au>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Mark Ryan <markdryan@rivosinc.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
Reviewed-by: David Chase <drchase@google.com>
The comment being removed is no longer true as of CL 494916.
Change-Id: I4975866c58efb5c9c41db3640564ac7233f5840f
Reviewed-on: https://go-review.googlesource.com/c/go/+/542415
Auto-Submit: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Tobias Klauser <tobias.klauser@gmail.com>
Reviewed-by: Than McIntosh <thanm@google.com>
Run-TryBot: Kirill Kolyshkin <kolyshkin@gmail.com>
Reviewed-by: Bryan Mills <bcmills@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
The call site calculation in the previous version is incorrect. For
the PGO preprocess file, the compiler should directly use the call
site offset value. Additionly, this change refactors the preprocess
tool to clean up unused fields including startline, the flat and the
cum.
Change-Id: I7bffed3215d4c016d9a9e4034bfd373bf50ab43f
Reviewed-on: https://go-review.googlesource.com/c/go/+/562795
Reviewed-by: Cherry Mui <cherryyz@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
adds a test case for the MarshalJSON and MarshalText method of the Level type in the slog package.
Change-Id: I3f79f0b46c41252ad9d743e03e34503e19998f3e
GitHub-Last-Rev: dab00d4c20
GitHub-Pull-Request: golang/go#65525
Reviewed-on: https://go-review.googlesource.com/c/go/+/561315
Reviewed-by: Damien Neil <dneil@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Jonathan Amsterdam <jba@google.com>
The new wording is slightly more strict than before in that
it guarantees each Write only contains exactly one Record,
while the previous wording opened up the possibility for
multiple Records in a Write call.
We also add a comment about the lack of sorting guarantees for
concurrently logged Records. That is, the obtained lock only covers
the Write call, rather than the combination of the call to time.Now,
JSON/text serialization, and also the Write call.
Change-Id: Ia65c50579215a35a1f5b2952c6954ddb60e7fe66
Reviewed-on: https://go-review.googlesource.com/c/go/+/563976
Reviewed-by: Jonathan Amsterdam <jba@google.com>
TryBot-Bypass: Jonathan Amsterdam <jba@google.com>
Reviewed-by: Alan Donovan <adonovan@google.com>
According to RFC 9110 and RFC 9112, invalid "Content-Length" headers
might involve request smuggling or response splitting, which could
also cause security failures. Currently, `net/http` ignores all
"Content-Length" headers when there is a "Transfer-Encoding" header and
forward the message anyway while other mainstream HTTP implementations
such as Apache Tomcat, Nginx, HAProxy, Node.js, Deno, Tornado, etc. reject
invalid Content-Length headers regardless of the presence of a
"Transfer-Encoding" header and only forward chunked-encoding messages
with either valid "Content-Length" headers or no "Content-Length" headers.
Fixes#65505
Change-Id: I73af2ee0785137e56c7546a4cce4a5c5c348dbc5
Reviewed-on: https://go-review.googlesource.com/c/go/+/561075
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Bryan Mills <bcmills@google.com>
Reviewed-by: Damien Neil <dneil@google.com>
Go 1.23 will require macOS 11 Big Sur or later, even on AMD64.
The comment here suggests the main requirement for the OS and
SDK version is to be recent enough not to break Apple signing,
and recent enough not to cause other problems.
For now, this CL simplifies the code by merging the ARM64 and
AMD64 cases into one, given 1.23 will be the first Go release
with a common minimum macOS version for both architectures so
there's no need to treat them separately here.
For #64207.
Change-Id: I821fcb9a2a316de0703833c8a75abcbaa10b17a3
Cq-Include-Trybots: luci.golang.try:gotip-darwin-amd64_11,gotip-darwin-amd64_14,gotip-darwin-arm64_11,gotip-darwin-arm64_13
Reviewed-on: https://go-review.googlesource.com/c/go/+/563857
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Cherry Mui <cherryyz@google.com>
For #64207.
For #65614.
Change-Id: Ia5365b4090060e0657c255341751d3e67691f836
Reviewed-on: https://go-review.googlesource.com/c/go/+/563856
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
This is another attempt at CL 558895, but without adding stale pollDescs
protection, which deviates from the original goal of the CL and adds
complexity without proper testing.
It is currently not possible to distinguish between a netpollBreak,
an internal/poll WSA operation, and an external WSA operation (as
in #58870). This can cause spurious wakeups when external WSA operations
are retrieved from the queue, as they are treated as netpollBreak
events.
This CL makes use of completion keys to identify the source of the
event.
While here, fix TestWSASocketConflict, which was not properly
exercising the "external WSA operation" case.
Change-Id: I91f746d300d32eb7fed3c8f27266fef379360d98
Reviewed-on: https://go-review.googlesource.com/c/go/+/561895
Reviewed-by: Than McIntosh <thanm@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
CL 561937 taught os.Stat about IO_REPARSE_TAG_AF_UNIX and added a
test for it, TestModeSocket. This test fails on Windows older than
10.0.17063, in which AF_UNIX support was added. Skip the test on those
versions.
Some CI builders use Windows 10.0.14393, so CL 561937 broke them,
e.g. https://build.golang.org/log/5ea4f6422779f32eccfef3a25df54283ddd4e65e.
Change-Id: I6c21a78a1454d2d88321478288c0da1b8a93e590
Reviewed-on: https://go-review.googlesource.com/c/go/+/563256
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Bryan Mills <bcmills@google.com>
The internal/asan package contains helper functions for manually
instrumenting code for the address sanitizer. It reexports the asan
routines in runtime unconditionally, making the functions a no-op if the
build flag "asan" is not present.
For #64611
Change-Id: Ie79e698aea7a6d969afd2a5f008c084c9545b1a5
GitHub-Last-Rev: e658670c14
GitHub-Pull-Request: golang/go#64635
Reviewed-on: https://go-review.googlesource.com/c/go/+/548695
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Do not add a trailing slash and redirect if the path already
ends in a slash.
Also, and unrelatedly, add a test for cleanPath.
Fixes#65624.
Change-Id: Ifcf9edc929d2eb6db88132c09d2bade85c5dda3d
Reviewed-on: https://go-review.googlesource.com/c/go/+/562557
Reviewed-by: Damien Neil <dneil@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Change-Id: I20e72d421d89095c460495001969291b99cdf59e
Reviewed-on: https://go-review.googlesource.com/c/go/+/563139
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Damien Neil <dneil@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
A typo without consequences, but confusing nonetheless. The last
line prints micro2 and then micro, instead of micro2 twice.
One-character fix.
Fixes#65666
Change-Id: I61d636382a2223d53eac58d6ddbcc7c15b4efc85
Reviewed-on: https://go-review.googlesource.com/c/go/+/563275
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
Run-TryBot: Rob Pike <r@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Probably a day 1 oversight, and almost always inconsequential, but
there is evidence of occasional trouble. There is no reason not to
clear them.
I tried and failed to write a test to catch this, but the change should
be harmless and is all but certain to fix the problem.
Fixes#61913
Change-Id: I0f7bbb4ab2780d8999d3ff7a35255dc07fb5c7e1
Reviewed-on: https://go-review.googlesource.com/c/go/+/556215
Run-TryBot: Rob Pike <r@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
The %#g format prints a "Go-syntax representation", but there is
no such thing for IEEE754 infinities and NaNs, so just document
what happens, which is that it prints +Inf, -Inf, or NaN. We could
show something like math.Inf(1) and math.Nan(), but that doesn't
sit right, and anyway for NaNs you can't even recover the original
value. Simpler and more honest to give up.
Fixes#51486
Change-Id: I8d4e8186f5d7acc3e0e7b51d0b322142908ea0a2
Reviewed-on: https://go-review.googlesource.com/c/go/+/557235
Run-TryBot: Rob Pike <r@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
The default case in x_cgo_getstackbound does not actually get the stack bound of
the current thread, but estimates the bound based on the default stack size. Add
a comment noting this.
Change-Id: I7d886461f0bbc795834bed37b554417cf3837a2d
Reviewed-on: https://go-review.googlesource.com/c/go/+/563376
Auto-Submit: Michael Pratt <mpratt@google.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Fixes#63337
Change-Id: I239315047e6e4325e2f471108fd764f8dbb7d5b2
GitHub-Last-Rev: cacdf0a279
GitHub-Pull-Request: golang/go#64952
Reviewed-on: https://go-review.googlesource.com/c/go/+/553616
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
The reflect.Type.Elem method is somewhat slow,
which is unfortunate since the reflect.TypeOf((*T)(nil)).Elem()
trick is only needed if T is an interface.
Optimize for concrete types by doing the faster reflect.TypeOf(v)
call first and only falling back on the Elem method if needed.
Performance:
name old time/op new time/op delta
TypeForString-24 9.10ns ± 1% 1.78ns ± 2% -80.49% (p=0.000 n=10+10)
TypeForError-24 9.55ns ± 1% 9.78ns ± 1% +2.39% (p=0.000 n=10+9)
Updates #60088
Change-Id: I2ae76988c9a3dbcbae10d2c19b55db3c8d4559bf
Reviewed-on: https://go-review.googlesource.com/c/go/+/555597
Auto-Submit: Joseph Tsai <joetsai@digital-static.net>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
Run-TryBot: Joseph Tsai <joetsai@digital-static.net>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Mauri de Souza Meneguzzo <mauri870@gmail.com>
Unix sockets are identified by the IO_REPARSE_TAG_AF_UNIX reparse tag.
Teach fileStat.Mode() to recognize this tag and set the os.ModeSocket
bit in such case.
Note that there is a bug starting in Windows 19H1 until 20H1 that
makes the IO_REPARSE_TAG_AF_UNIX tag not being set for unix sockets.
This CL doesn't provide a workaround for this bug.
Fixes#33357.
Change-Id: Iea8f24b20672c8d4b03f55ef298d128431dc3fac
Reviewed-on: https://go-review.googlesource.com/c/go/+/561937
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Bryan Mills <bcmills@google.com>
When cross-compiling a cgo program with CC=clang for Linux/ARMv5,
atomic warnings cause build errors, as cgo uses -Werror.
These warnings seem to be harmless and come from the usage of
__atomic_load_n, which is emulated due to the lack of atomic
instructions in armv5.
Fixes#65290
Change-Id: Ie72efb77468f06888f81f15850401dc8ce2c78f9
GitHub-Last-Rev: fbad847b96
GitHub-Pull-Request: golang/go#65588
Reviewed-on: https://go-review.googlesource.com/c/go/+/562348
Auto-Submit: Ian Lance Taylor <iant@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Fixes#65419
Change-Id: I8f9f82ab0387d8bb39aaca4f9e60e36ee15c587d
Reviewed-on: https://go-review.googlesource.com/c/go/+/560137
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
This ensures the trace buffers are as up-to-date as possible right
before crashing. It increases the chance of finding the culprit for the
crash when looking at core dumps, e.g. if slowness is the cause for the
crash (monitor kills process).
Fixes#65319.
Change-Id: Iaf5551911b3b3b01ba65cb8749cf62a411e02d9c
Reviewed-on: https://go-review.googlesource.com/c/go/+/562616
Auto-Submit: Michael Knyszek <mknyszek@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
UDP messages may be truncated:
https://www.rfc-editor.org/rfc/rfc1035#section-4.2.1
> Messages carried by UDP are restricted to 512 bytes (not counting
> the IP or UDP headers). Longer messages are truncated and the TC
> bit is set in the header.
However, TCP also have a size limitation of 65535 bytes
https://www.rfc-editor.org/rfc/rfc1035#section-4.2.2
> The message is prefixed with a two byte length field which gives
the message length, excluding the two byte length field.
These limitations makes that the maximum possible number of A records
per RRSet is ~ 4090.
There are environments like Kubernetes that may have larger number of
records (5000+) that does not fit in a single message. In this cases,
the DNS server sets the Truncated bit on the message to indicate that
it could not send the full answer despite is using TCP.
We should only retry when the TC bit is set and the connection is UDP,
otherwise, we'll never being able to get an answer and the client will
receive an errNoAnswerFromDNSServer, that is a different behavior than
the existing in the glibc resolver, that returns all the existing
addresses in the TCP truncated response.
Fixes#64896
Signed-off-by: Antonio Ojea <aojea@google.com>
Change-Id: I1bc2c85f67668765fa60b5c0378c9e1e1756dff2
Reviewed-on: https://go-review.googlesource.com/c/go/+/552418
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Gudger <ian@iangudger.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Mateusz Poliwczak <mpoliwczak34@gmail.com>
Reviewed-by: Mauri de Souza Meneguzzo <mauri870@gmail.com>
Reviewed-by: Damien Neil <dneil@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Make C_S32CON, C_U32CON, and C_32CON distinct classifiers to allow
more specific matching of 32 bit constants. C_U31CON is added to
support C_S32CON.
Likewise, add C_16CON which is the union of C_S16CON and C_U16CON
classification. This wil allow simplifying MOVD/MOVW optab entries
in a future patch.
Change-Id: I193acc0ded8f3edd91d306e39c3e7e55a9811e04
Reviewed-on: https://go-review.googlesource.com/c/go/+/562346
Reviewed-by: Lynn Boger <laboger@linux.vnet.ibm.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
Run-TryBot: Paul Murphy <murp@ibm.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
This would have caught #65607.
On my machine:
internal/trace/v2: ~13s
internal/trace/v2: -short: ~4s
internal/trace/v2: -race -short: ~20s
internal/trace/v2: -race: ~46s
For #65607.
Change-Id: I03a555ebaf8c761c98a82b6d35fa16d7a99aa9e3
Reviewed-on: https://go-review.googlesource.com/c/go/+/562998
Auto-Submit: Michael Pratt <mpratt@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Multiple goroutines all writing to the same sink triggers the race detector,
rightfully so.
Change-Id: Ia64836d0d88c0f587a6cb96ed747f656a3c1804a
Reviewed-on: https://go-review.googlesource.com/c/go/+/562997
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Michael Pratt <mpratt@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
traceReadCPU calls profBuf.read, which does a raceacquire. g0 does not
have a race context, so this crashes when running on the system stack.
We could borrow a race context, but it is simpler to just move
traceReadCPU off of the system stack.
Fixes#65607.
Change-Id: I335155b96d683aebb92b2f4e1eea063dd139f2d5
Reviewed-on: https://go-review.googlesource.com/c/go/+/562996
Auto-Submit: Michael Pratt <mpratt@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
PGO uses noder.LookupFunc to look for devirtualization targets in
export data. LookupFunc does not support type-parameterized
functions, and will currently fail the build when attempting to lookup
a type-parameterized function because objIdx is passed the wrong
number of type arguments.
This doesn't usually come up, as a PGO profile will report a generic
function with a symbol name like Func[.go.shape.foo]. In export data,
this is just Func, so when we do LookupFunc("Func[.go.shape.foo]")
lookup simply fails because the name doesn't exist.
However, if Func is not generic when the profile is collected, but the
source has since changed to make Func generic, then LookupFunc("Func")
will find the object successfully, only to fail the build because we
failed to provide type arguments.
Handle this with a objIdxMayFail, which allows graceful failure if the
object requires type arguments.
Bumping the language version to 1.21 in pgo_devirtualize_test.go is
required for type inference of the uses of mult.MultFn in
cmd/compile/internal/test/testdata/pgo/devirtualize/devirt_test.go.
Fixes#65615.
Change-Id: I84d9344840b851182f5321b8f7a29a591221b29f
Reviewed-on: https://go-review.googlesource.com/c/go/+/562737
Reviewed-by: Cherry Mui <cherryyz@google.com>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Pending a resolution to #65606, this CL marks clang's ASAN runtime as
unable to symbolize stack traces to unblock the LUCI clang builder.
For #65606.
Fixes#65469.
Change-Id: I649773085aff30e5703e7f7ac2c72a0430a015c2
Cq-Include-Trybots: luci.golang.try:gotip-linux-amd64-clang15
Reviewed-on: https://go-review.googlesource.com/c/go/+/562675
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
This CL separates the pass that computes inlinability from the pass
that performs inlinability. In particular, the latter can now happen
in any flat order, rather than bottom-up order. This also allows
inlining of calls exposed by devirtualization.
Change-Id: I389c0665fdc8288a6e25129a6744bfb1ace1eff7
Reviewed-on: https://go-review.googlesource.com/c/go/+/562319
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
