The PGO-based devirtualization helper pgoir.addIndirectEdges makes a
series of calls into the unified IR reader to import functions that
would not normally be imported but may be the target of a hot indirect
call from the current package. This importing primarily targets at
non-generic functions and methods, but as part of the process we can
encounter types that have methods (including generic methods) whose
bodies need to be read in. When the reader encounters an inlinable
func of this sort, it may (depending on the context) decide not to
read the body right away, but instead adds the func to a list
("todoBodies") to be read in later on in a more convenient context.
In the bug in question, a hot method lookup takes place in
pgoir.addIndirectEdges, and as part of the import process we wind up
with a type T with method M that is in this partially created state,
and in addition T gets added to the unified IR's list of types that
may need method wrappers. During wrapper generation we create a new
wrapper "(*T).M" whose body has a call to "T.M", then farther on down
the pike during escape analysis we try to analyze the two functions;
this causes a crash due to "T.M" being in partially constructed state.
As a fix, add a new "PostLookupCleanup" hook (in the unified IR
reader) that pgoir.addIndirectEdges can invoke that takes care of
reading in the bodies of any functions that have been added to the
"todoBodies" list.
[Note: creating a test case for this problem is proving to be very
tricky; a new test will be added in a subsequent patch].
Fixes#67746.
Change-Id: Ibc47ee79e08a55421728d35341df80a865231cff
Reviewed-on: https://go-review.googlesource.com/c/go/+/591075
Reviewed-by: Cherry Mui <cherryyz@google.com>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
For #61308.
For #65614.
Change-Id: I36b4f2392075d5a3fb9f53a28bd19b997e7be363
Reviewed-on: https://go-review.googlesource.com/c/go/+/592197
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
TryBot-Bypass: Dmitri Shuralyov <dmitshur@golang.org>
The notes don't define identifiers m and seq. Describe in words
instead.
While here, rename 3-structs.md to 4-structs.md.
For #65614.
Change-Id: I6690f552903d1870c70e28673c87ac84b7c5c041
Reviewed-on: https://go-review.googlesource.com/c/go/+/592195
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
It reportedly breaks the DNS server on some modems.
For #6464
For #21160
For #44135
For #51127
For #51153Fixes#67925
Change-Id: I54a11906159f00246d08a54cc8be7327e9ebfd2c
Reviewed-on: https://go-review.googlesource.com/c/go/+/591995
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Damien Neil <dneil@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Currently the runtime.end symbol is put into the noptrbss section,
which is usually the last section, except that when fuzzing is
enabled, the last section is actually .go.fuzzcntrs. The
runtime.end symbol has the value pointing to the end of the data
segment, so if it is not in the last section, the value will not
actually be in the range of the section. This causes an assertion
failure in the new Apple linker. This CL fixes this by putting it
in the last section.
Fixes#65169.
Change-Id: I5c991c46a0483a96e5f6e0255a3b444953676026
Reviewed-on: https://go-review.googlesource.com/c/go/+/592095
Reviewed-by: Than McIntosh <thanm@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
For processes that don't exist at lookup time, CL 570036 and CL 588675
make Wait return unconditionally return ErrProcessDone when using pidfd,
rather than attempting to make a wait system call.
This is consistent with Signal/Kill, but inconsistent with the previous
behavior of Wait, which would pass through the kernel error,
syscall.ECHILD.
Switch the ErrProcessDone case to return syscall.ECHILD instead for
consistency with previous behavior.
That said, I do think a future release should consider changing ECHILD
to ErrProcessDone in all cases (including when making an actual wait
system call) for better consistency with Signal/Kill/FindProcess.
Fixes#67926.
Cq-Include-Trybots: luci.golang.try:gotip-darwin-amd64_14,gotip-solaris-amd64,gotip-openbsd-amd64
Change-Id: I1f688a5751d0f3aecea99c3a5b35c7894cfc2beb
Reviewed-on: https://go-review.googlesource.com/c/go/+/591816
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Kirill Kolyshkin <kolyshkin@gmail.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
The existing implementation logs some errors to the results file created in TestBogoSuite.
This change would additionally log json errors to the results file.
Change-Id: Ib1a6f612ed83f6c5046531ee259c4e85dd71402a
Reviewed-on: https://go-review.googlesource.com/c/go/+/591379
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Auto-Submit: Filippo Valsorda <filippo@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
For #65614.
Updates #61476.
Change-Id: Id677aa6d2a59366ab75a26f08a383d2d253f270e
Reviewed-on: https://go-review.googlesource.com/c/go/+/591895
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Joel Sing <joel@sing.id.au>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
For #65614.
Updates #60905.
Change-Id: I2dd9df3c7066357cf06268d918bad3c255b38aed
Reviewed-on: https://go-review.googlesource.com/c/go/+/591875
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Joel Sing <joel@sing.id.au>
Change-Id: I418953a766db22b134a9569161f06cf8682c1eef
Reviewed-on: https://go-review.googlesource.com/c/go/+/591336
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
When using time.RFC1123Z to parse the date header value out of an email,
an error is returned for dates that occur in the first 9 days of a
month. This is because the format strings for RFC 1123 defined in the
time package indicate that the day should be prefixed with a leading 0.
Reading the spec, the line that talks about it seems to indicate that
days can be either 1 or 2 digits:
`date = 1*2DIGIT month 2*4DIGIT`
So a date header with a day like `7` with no leading zero should be
accepted.
Fixes#67887
Change-Id: Ie7ee40d94da2c8c0417957e8b89f9987314949c8
GitHub-Last-Rev: 22a5a52fcb
GitHub-Pull-Request: golang/go#67888
Reviewed-on: https://go-review.googlesource.com/c/go/+/591335
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
When unlink fails, it is not yet known if the argument is a directory or not.
Since CL 588495, we figure out if it's a directory when trying to open
it (and, for a directory, return the original unlink error).
The (very minor) issue is, in case of a symlink, a different error is
returned -- usually it's ELOOP, but some systems use other values. Let's
account for that error code, too.
This is a followup to CL 588495.
Change-Id: I4ee10fe9b57f045fbca02f13e5c9ea16972803bc
Reviewed-on: https://go-review.googlesource.com/c/go/+/589376
Reviewed-by: Michael Pratt <mpratt@google.com>
Run-TryBot: Kirill Kolyshkin <kolyshkin@gmail.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
The libc docs say that these functions can return ENOENT, and
per issue #67912, this does happen in practice sometimes. Handle
both the ENOENT and !found cases the same way, for Lookup,
LookupId, LookupGroup, LookupGroupId.
Fixes#67912
Change-Id: I993935af44c83ad785b6cd735fc313a3647daa19
Reviewed-on: https://go-review.googlesource.com/c/go/+/591555
Reviewed-by: Ian Lance Taylor <iant@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
The URL has changed.
Change-Id: I295020c59aaf64a89c4307fca527163095148a67
Reviewed-on: https://go-review.googlesource.com/c/go/+/591897
Reviewed-by: Joel Sing <joel@sing.id.au>
Reviewed-by: M Zhuo <mengzhuo1203@gmail.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Mark Ryan <markdryan@rivosinc.com>
I think we want to keep the "experimental" as we did before. Also
added "64-bit" to be more specific.
For #65614.
Change-Id: I928d0ba6f1b53efa1f882f81f68bfb2446c7ab42
Reviewed-on: https://go-review.googlesource.com/c/go/+/591896
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Joel Sing <joel@sing.id.au>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
For #65614.
Change-Id: I1a2a4bc18601526053840a280e0604a8e1028ce6
Reviewed-on: https://go-review.googlesource.com/c/go/+/591899
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Than McIntosh <thanm@google.com>
The unification code has "early exits" when the compared
types are pointer-identical.
Because of Alias nodes, we cannot simply compare x == y but we
must compare Unalias(x) == Unalias(y). Still, in the common case
there are no aliases, so as a minor optimization we write:
x == y || Unalias(x) == Unalias(y)
to test whether x and y are (pointer-) identical.
Add the missing Unalias calls in the place where we forgot them.
Fixes#67872.
Change-Id: Ia26ffe7205b0417fc698287a4aeb1c900d30cc0d
Reviewed-on: https://go-review.googlesource.com/c/go/+/591975
Reviewed-by: Robert Findley <rfindley@google.com>
Auto-Submit: Robert Griesemer <gri@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Robert Griesemer <gri@google.com>
There are several issues with pidfd handling today:
* The zero value of a Process makes the handle field appear valid, so
methods attempt to use it as a pidfd rather than falling back to the
PID as they should (#67634).
* If a process doesn't exist, FindProcess returns a Process with Pid ==
-2, which is not a compatible change (#67640).
* pidfd close is racy as-is. A Release call or successful Wait will
clear the handle field and close the pidfd. However, a concurrent call
may have already loaded the handle field and could then proceed to use
the closed FD (which could have been reopened as a different pidfd,
targeting a different process) (#67641).
This CL performs multiple structural changes to the internals of
Process.
First and foremost, each method is refactored to clearly select either
pidfd or raw pid mode. Previously, raw pid mode was structured as a
fallback when pidfd mode is unavailable. This works fine, but it does
not make it clear that a given Process object either always uses pidfd
or always uses raw pid. Since each mode needs to handle different race
conditions, it helps to make it clear that we can't switch between modes
within a single Process object.
Second, pidfd close safety is handled by reference counting uses of the
FD. The last user of the FD will close the FD. For example, this means
that with concurrent Release and Signal, the Signal call may be the one
to close the FD. This is the bulk of this CL, though I find the end
result makes the overall implementation easier to reason about.
Third, the PID path handles a similar race condtion between Wait and
Kill: Wait frees the PID value in the kernel, which could be reallocated
causing Kill to target the wrong process. This is handled with a done
flag and a mutex. The done flag now shares the same state field used for
the handle.
Similarly, the Windows implementation reuses all of the handle reference
counting that Linux uses. This means the implementations more
consistent, and make Windows safe against the same handle reuse
problems. (Though I am unsure if Windows ever reuses handles).
Wait has a slight behavior change on Windows: previously Wait after
Release or an earlier Wait would hang indefinitely (WaitForSingleObject
on syscall.InvalidHandle waits indefinitely). Now it returns the same
errors as Linux (EINVAL and ErrProcessDone, respectively).
Similarly, Release on Windows no longer returns close errors, as it may
not actually be the place where the close occurs.
Fixes#67634.
Fixes#67640.
Fixes#67641.
Updates #67642.
Cq-Include-Trybots: luci.golang.try:gotip-linux-amd64-longtest,gotip-windows-amd64-longtest
Change-Id: I2ad998f7b67d32031e6f870e8533dbd55d3c3d10
Reviewed-on: https://go-review.googlesource.com/c/go/+/588675
Reviewed-by: Austin Clements <austin@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
CL 586975 added support to the compiler back end to emit a synthetic
".closureptr" variable in range func bodies, plus code to spill the
incoming context pointer to that variable's location on the stack.
This patch fixes up the code in the back end that generates DWARF
location lists for incoming parameters (which sometimes arrive in
registers) in the "-l -N" no-optimization case to also create a
correct DWARF location list for ".closureptr", a two-piece list
reflecting the fact that its value arrives in a register and then is
spilled to the stack in the prolog.
Fixes#67918.
Change-Id: I029305b5248b8140253fdeb6821b877916fbb87a
Reviewed-on: https://go-review.googlesource.com/c/go/+/591595
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Alessandro Arzilli <alessandro.arzilli@gmail.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
The creation of a bytes.Buffer in one code path is missing causing a nil
pointer dereference.
Changed (as rec. by Bryan Mills) to use fmt.Appendf() on []byte instead of
fmt.Fprintf on *bytes.Buffer - simpler and avoids duplicated code (but
requires Go 1.19 or later).
Added test to verify the change (as rec. by Michael Matloob) at
src\cmd\go\testdata\script\build_repeated_godebug_issue62346.txt
Fixes#62346
Change-Id: Ic3267d878a6f7ebedb1cde64e6206de404176b10
Reviewed-on: https://go-review.googlesource.com/c/go/+/523836
Reviewed-by: Michael Matloob <matloob@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Auto-Submit: Michael Matloob <matloob@golang.org>
Change-Id: Ia73b15b6df5e6f88741f2b54258d7cbad0239ddd
Reviewed-on: https://go-review.googlesource.com/c/go/+/591396
Reviewed-by: Robert Findley <rfindley@google.com>
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>
The lack of links to https://go.dev/ref/mem in the sync package
documentation makes it difficult to read for people who have no previous
knowledge of that page. This PR includes the links where needed.
Fixes#67891
Change-Id: I0e1344cc6d7b702f4cb2e55fe0fcee3eb089391a
GitHub-Last-Rev: 427cf58aae
GitHub-Pull-Request: golang/go#67892
Reviewed-on: https://go-review.googlesource.com/c/go/+/591395
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Since CL 588076 runc can do fine without the kludge. The code accessing the symbol is now guarded with `go:build !go1.23` in all supported runc branches (main: [1], release-1.1: [2]).
This reverts part of CL 587219.
Updates #67401.
For #66797.
[1]: https://github.com/opencontainers/runc/pull/4290
[2]: https://github.com/opencontainers/runc/pull/4299
Change-Id: I204843a93c36857e21ab9b43bd7aaf046e8b9787
Reviewed-on: https://go-review.googlesource.com/c/go/+/587918
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
CL 512155 fixed#36768, but introduced #62596.
CL 527820 fixed#62596, but meant that the code failed to look up
file extensions on Windows for a relative path.
This CL fixes that problem by recording whether it has already
looked up file extensions.
This does mean that if Path is set manually then we do not update
it with file extensions, as doing that would be racy.
Fixes#66586
Change-Id: I9a0305d1e466c5e07bfbe442566ea12f5255a96e
GitHub-Last-Rev: dc3169f235
GitHub-Pull-Request: golang/go#67035
Reviewed-on: https://go-review.googlesource.com/c/go/+/581695
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Consider the following code snippet:
next, stop := iter.Pull(seq)
stop()
Today, seq will iterate exactly once before it notices that its
iteration is invalid to begin with. This effect is observable in a
variety of ways. For example, if the iterator panics, since that panic
must propagate to the caller of stop. But if the iterator is stateful in
anyway, then it may update some state.
This is somewhat unexpected and because it's observable, can be depended
upon. This behavior does not align well with other possible
implementations of Pull, like CPS performed by the compiler. It's also
just odd to let even one iteration happen, precisely because of
unexpected state modification.
Fix this by not iterating at all of the done flag is set before entering
the iterator.
For #67712.
Change-Id: I18162e29df45a2e8968f68379450d92e1de47c4d
Reviewed-on: https://go-review.googlesource.com/c/go/+/590075
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Austin Clements <austin@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
This change propagates a runtime.Goexit initiated by the iterator into
the caller of next and/or stop.
Fixes#67712.
Change-Id: I5bb8d22f749fce39ce4f587148c5fc71aee2af65
Reviewed-on: https://go-review.googlesource.com/c/go/+/589137
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: David Chase <drchase@google.com>
This change propagates panics from the iterator passed to Pull through
next and stop. Once the panic occurs, next and stop become no-ops (the
iterator is invalidated).
For #67712.
Change-Id: I05e45601d4d10acdf51b53e3164bd891c1b324ac
Reviewed-on: https://go-review.googlesource.com/c/go/+/589136
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Austin Clements <austin@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
RLIWNM does not clear the upper 32 bits of the target register if
the mask wraps around (e.g 0xF000000F). Don't elide MOVWZreg for
such masks. All other usage clears the upper 32 bits.
Fixes#67844.
Change-Id: I11b89f1da9ae077624369bfe2bf25e9b7c9b79bc
Reviewed-on: https://go-review.googlesource.com/c/go/+/590896
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Cherry Mui <cherryyz@google.com>
Reviewed-by: Keith Randall <khr@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
The test fail when $GOROOT/go.env contain GOTOOLCHAIN=local
because GOTOOLCHAIN=local is assumed to be a non-default value.
This CL fixed the test failure
by using go.env from the test as $GOROOT/go.env throughout the test.
Test have also been added to ensure that
when $GOROOT/go.env contain GOTOOLCHAIN=local,
GOTOOLCHAIN=local is not taken as a non-default value.
Fixes#67793
Change-Id: Ibc5057d38d36c6c55726a039de1e7c37d6935b52
GitHub-Last-Rev: 12b62464e6
GitHub-Pull-Request: golang/go#67807
Reviewed-on: https://go-review.googlesource.com/c/go/+/590196
Reviewed-by: Sam Thanawalla <samthanawalla@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Fixes#67728
Change-Id: Ifbdee77dcebd865eac9501ec4615128eb3a9c9c0
Reviewed-on: https://go-review.googlesource.com/c/go/+/590797
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: Michael Matloob <matloob@golang.org>
Commit-Queue: Ian Lance Taylor <iant@google.com>
It actually causes the program to throw.
Fixes#67729
Change-Id: Id970baff631616a4dc4e434827e622e1b16f2724
Reviewed-on: https://go-review.googlesource.com/c/go/+/590915
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Commit-Queue: Ian Lance Taylor <iant@google.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
Auto-Submit: Ian Lance Taylor <iant@golang.org>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Just like for tls.Config.GetCertificate the http.Server.ServeTLS method
should be checking tls.Config.GetConfigForClient before trying top open
the specified certFile/keyFile.
This was previously fixed for crypto/tls when using tls.Listen in
CL205059, but the same change for net/http was missed. I've added a
comment src/crypto/tls/tls.go in the relevant section in the hope that
any future changes of a similar nature consider will consider updating
net/http as needed as well.
Change-Id: I312303bc497d92aa2f4627fe2620c70779cbcc99
GitHub-Last-Rev: 6ed29a9008
GitHub-Pull-Request: golang/go#66795
Reviewed-on: https://go-review.googlesource.com/c/go/+/578396
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Reviewed-by: Damien Neil <dneil@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
Currently, the compiler generates the argument stack map based on
the function signature for bodyless function declarations, if it
is not linknamed. The assumption is that linknamed function is
provided by (Go code in) another package, so its args stack map
will be generated when compiling that package.
Now we have linknames added to declarations of assembly functions,
to signal that this function is accessed externally. Examples
include runtime.morestack_noctxt, math/big.addVV. In the current
implementation the compiler does not generate its args stack map.
That causes the assembly function's args stack map missing.
Instead, change it to generate the stack map if it is a
declaration of an ABI0 function, which can only be defined in
assembly and passed to the compiler through the -symabis flag. The
stack map generation currently only works with ABI0 layout anyway,
so we don't need to handle ABIInternal assembly functions.
Change-Id: Ic9da3b4854c604e64ed01584da3865994f5b95b8
Reviewed-on: https://go-review.googlesource.com/c/go/+/587928
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Than McIntosh <thanm@google.com>
In shared build mode, we include all symbols. This includes
function symbols that are deadcoded by the compiler. They don't
really get compiled, and their metadata may be missing, causing
linker failures. Skip them.
Fixes#67635.
Change-Id: Ic0e64bd032be499cca26da5e9e3ffbe9998bac05
Reviewed-on: https://go-review.googlesource.com/c/go/+/588316
Reviewed-by: Than McIntosh <thanm@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
A method receiver can be collected during the execution of that method.
This does make sense when thinking about how the GC would work, but
also seems not very obvious, so a point in the docs can increase the
chance of avoiding issues caused by missing KeepAlive of method
receivers.
Change-Id: I6817237dd022b5729dbdcda1b9f70c7059605575
GitHub-Last-Rev: 878bf3fde7
GitHub-Pull-Request: golang/go#67777
Reviewed-on: https://go-review.googlesource.com/c/go/+/589735
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Keith Randall <khr@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
When sending a request with an "Expect: 100-continue" header,
we must send the request body before sending any further requests
on the connection.
When receiving a non-1xx response to an "Expect: 100-continue" request,
send the request body if the connection isn't being closed after
processing the response. In other words, if either the request
or response contains a "Connection: close" header, then skip sending
the request body (because the connection will not be used for
further requests), but otherwise send it.
Correct a comment on the server-side Expect: 100-continue handling
that implied sending the request body is optional. It isn't.
For #67555
Change-Id: Ia2f12091bee697771087f32ac347509ec5922d54
Reviewed-on: https://go-review.googlesource.com/c/go/+/591255
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Jonathan Amsterdam <jba@google.com>
Between Go 1.18 and Go 1.22 go get printed a fatal error if -d was
explicitly set to false. That behavior was reverted in CL 572176, when
we made the -d flag a no-op, but it would make it easier to remove the
-d flag in the future if we continue to print a fatal error if -d is
explicitly set to false.
This change brings back the fatal error for -d=false while keeping the
warning printed for -d=true.
For #43684
Change-Id: I38ae3a3619d408c0237ff485ddee4403b8188abd
Reviewed-on: https://go-review.googlesource.com/c/go/+/591135
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Sam Thanawalla <samthanawalla@google.com>
In addition, add a line number to the go.work error for multiple use statements
for the same directory. Also clean up the error prefix for go.work
errors now containing line numbers.
Fixes#67623
Change-Id: Ia7edcc50f7d7ec907b4a9eb4fe270c75d04c1fa5
Reviewed-on: https://go-review.googlesource.com/c/go/+/590135
Reviewed-by: Sam Thanawalla <samthanawalla@google.com>
Auto-Submit: Michael Matloob <matloob@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Before this change, we didn't initialize the overlays in the fsys
package or use the fsys logic to read the files, so overlays were not
respected for go.work files. Initialize fsys before loading the go.work
file (initialization is idempotent) and use the new fsys.ReadFile
function to read the file instead of os.ReadFile.
fsys.ReadFile just opens the file with fsys.Open and then calls
io.ReadAll on it. (This is less efficient than what os.ReadFile does:
os.ReadFile reads into a buffer it allocated that's the file's size
while io.ReadAll doesn't know how big the file is so it just reads in
512 byte chunks.)
Change-Id: Ic40bcbb483a16c5d4dd1d896306ea99a16f370f3
Reviewed-on: https://go-review.googlesource.com/c/go/+/590755
Reviewed-by: Sam Thanawalla <samthanawalla@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Michael Matloob <matloob@golang.org>
Fixes#66358.
Change-Id: Ic9bde88eabfb2a446d32e1dc5ac404a51ef49f11
Reviewed-on: https://go-review.googlesource.com/c/go/+/590635
Auto-Submit: Robert Griesemer <gri@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: Robert Griesemer <gri@google.com>
This test can fail due to objects being incorrectly retained due
to conservative scanning. Allow a bit of slop (1 accidentally
retained object) to prevent flaky failures.
Fixes#67204
"fixes" is a bit too strong a word. More like, hopefully reduces
the false positive rate to something approaching 0.
Change-Id: I09984f0cce50d8209aef19f3d89b0e295c86f8d1
Reviewed-on: https://go-review.googlesource.com/c/go/+/590615
Reviewed-by: Keith Randall <khr@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
For #65614.
Change-Id: I9487521817348053170da293c01ac55f6e3dcd20
Reviewed-on: https://go-review.googlesource.com/c/go/+/590895
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Robert Griesemer <gri@google.com>
Auto-Submit: Robert Griesemer <gri@google.com>
TryBot-Bypass: Robert Griesemer <gri@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
The language change for the accepted range-over-func proposal #61405
was documented in CL 590616. Remove the corresponding 'TODO' entry.
Also improve formatting slightly, and switch to preferred relative
links. They'll work better in the long term and in more contexts.
While here, also simplify the suggested line to preview release notes
locally: setting the -content='' flag explicitly is no longer required
as of CL 589936.
For #65614.
Change-Id: I6cee951b9ede33900bca48c9f709e3b2c5e87337
Reviewed-on: https://go-review.googlesource.com/c/go/+/590756
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: Robert Griesemer <gri@google.com>
Use the default frame scheduler (currently round-robin)
rather than overriding the default with the priority scheduler.
The priority scheduler is slow, known buggy, and implements
a deprecated stream prioritization mechanism. The default
changed in x/net about a year ago, but we missed that net/http
is overriding that default.
Fixes#67706
Change-Id: I6d76dd0cc8c55eb5dec5cd7d25a5084877e8e8d9
Reviewed-on: https://go-review.googlesource.com/c/go/+/590796
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Jonathan Amsterdam <jba@google.com>
Auto-Submit: Damien Neil <dneil@google.com>
For #65614
For #67470
Change-Id: Iba2f263f8ca1fb10c383e12ff3455aa86b26421d
Reviewed-on: https://go-review.googlesource.com/c/go/+/590795
Commit-Queue: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
For #65614.
Change-Id: I05b20a80b1163e1d3927c2f763ef6b7d20e6a937
Reviewed-on: https://go-review.googlesource.com/c/go/+/590617
Reviewed-by: Robert Griesemer <gri@google.com>
Auto-Submit: Robert Griesemer <gri@google.com>
TryBot-Bypass: Robert Griesemer <gri@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
The arch name of loong64 should be loongarch64
Fixes#67832
Change-Id: Ic5cf0f0274262a164bba78c426813a8b8851033b
Reviewed-on: https://go-review.googlesource.com/c/go/+/590695
Reviewed-by: Than McIntosh <thanm@google.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: abner chenc <chenguoqi@loongson.cn>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Commit-Queue: Ian Lance Taylor <iant@google.com>
