If a goroutine is started within a user region, internal/trace assigns
the child goroutine a nameless region for its entire lifetime which is
assosciated the same task as the parent's region.
This is not strictly necessary: a child goroutine is not necessarily
related to the task unless it performs some task operation (in which
case it will be associated with the task through the standard means).
However, it can be quite handy to see child goroutines within a region,
which may be child worker goroutines that you simply didn't perform task
operations on.
If the first GC occurs during a region, the GC worker goroutines will
also inherit a child region. We know for sure that these aren't related
to the task, so filter them out from the region list.
Note that we can't exclude system goroutines from setting activeRegions
in EvGoCreate handling, because we don't know the goroutine start
function name until the first EvGoStart.
Fixes#53784.
Change-Id: Ic83d84e23858a8400a76d1ae2f1418ef49951178
Reviewed-on: https://go-review.googlesource.com/c/go/+/416858
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Rebuild selected amd64 syso images with updated LLVM build rules that
avoid the use of SSE4, so as to ensure that the Go race detector
continues to work on older x86 cpus. No changes to the syso files for
openbsd/amd64 (upstream support has been removed in LLVM) or
netbsd/amd64 (work still in progress there).
Fixes#53743.
Change-Id: I738ae4d1e0528c6e06dd4ddb78e7039a30a51779
Reviewed-on: https://go-review.googlesource.com/c/go/+/416857
Reviewed-by: Cherry Mui <cherryyz@google.com>
Run-TryBot: Than McIntosh <thanm@google.com>
CL 377516 made it so that memory metrics are truly monotonic, but also
updated how heapLive tracked allocated memory to also be monotonic.
The result is that cached spans with allocated memory aren't fully
accounted for by the GC, causing it to make a worse assumption (the
exact mechanism is at this time unknown), resulting in a memory
regression, especially for smaller heaps.
This change is a partial revert of CL 377516 that makes heapLive a
non-monotonic overestimate again, which appears to resolve the
regression.
For #53738.
Change-Id: I5c51067abc0b8e0a6b89dd8dbd4a0be2e8c0c1b2
Reviewed-on: https://go-review.googlesource.com/c/go/+/416417
Reviewed-by: Michael Pratt <mpratt@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
In CL 405901 the definition of exit in the Plan 9 go runtime
was changed like so:
- status = append(itoa(tmp[:len(tmp)-1], uint64(e)), 0)
+ sl := itoa(tmp[:len(tmp)-1], uint64(e))
+ // Don't append, rely on the existing data being zero.
+ status = tmp[:len(sl)+1]
However, itoa only puts the converted number "somewhere" in the buffer.
Specifically, it builds it from the end of the buffer towards the start,
meaning the first byte of the buffer is a 0 byte, and the resulting string
that's passed to exits is empty, leading to a falsely successful exit.
This change uses the returned value from itoa, rather than the buffer
that was passed in, so that we start from the correct location in the
string.
Fixes#53669
Change-Id: I63f0c7641fc6f55250857dc17a1eeb12ae0c2e10
Reviewed-on: https://go-review.googlesource.com/c/go/+/415680
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
cbs.lock protects a map. The map implementation is race instrumented
regardless of which package is it called from.
lock/unlock are not automatically race instrumented, so we can trigger
race false positives without manually annotating our lock acquire and
release.
compileCallback is used during initialization before the P is available,
at which point raceacquire will crash during a racecallback to get the
race proc. Thus we skip instrumentation until scheduler initialization
is complete.
Fixes#50249.
Change-Id: Ie49227c9e9210ffbf0aee65f86f2b7b6a2f64638
Reviewed-on: https://go-review.googlesource.com/c/go/+/414518
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Cherry Mui <cherryyz@google.com>
Run-TryBot: Michael Pratt <mpratt@google.com>
metricsSema protects the metrics map. The map implementation is race
instrumented regardless of which package is it called from.
semacquire/semrelease are not automatically race instrumented, so we can
trigger race false positives without manually annotating our lock
acquire and release.
See similar instrumentation on trace.shutdownSema and reflectOffs.lock.
Fixes#53542.
Change-Id: Ia3fd239ac860e037d09c7cb9c4ad267391e70705
Reviewed-on: https://go-review.googlesource.com/c/go/+/414517
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Cherry Mui <cherryyz@google.com>
Requested by the maintainers of the OpenSSL-based fork of Go+BoringCrypto,
to make maintaining that fork easier.
Change-Id: I770e70ecc12b589034da31edecf59c73b2c6e1dd
Reviewed-on: https://go-review.googlesource.com/c/go/+/407135
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Auto-Submit: Russ Cox <rsc@golang.org>
This lets us generate identical copies of fastlog2table.go on all hosts.
Tested by regenerating fastlog2table.go on linux-amd64 and darwin-arm64.
Fixes#49891
Change-Id: I279d6b5abb5a5290c049d9658050fd9c8d0c0190
Reviewed-on: https://go-review.googlesource.com/c/go/+/413976
Run-TryBot: Ian Lance Taylor <iant@google.com>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Ian Lance Taylor <iant@google.com>
These functions can be inserted by the compiler into the code to be
instrumented. This may result in these functions having callers that
are nosplit. That is why they must be nosplit.
This is a followup for CL 410034 in order to fix#53190.
Change-Id: I03746208a2a302a581a1eaad6c9d0672bb1e949a
GitHub-Last-Rev: 6506d86f22
GitHub-Pull-Request: golang/go#53544
Reviewed-on: https://go-review.googlesource.com/c/go/+/413978
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Run-TryBot: Dmitri Shuralyov <dmitshur@golang.org>
When runtime.sigprof encounters a stack that gentraceback is unable to
process, it synthesizes a call stack with a sentinel function (such as
runtime._System) at the leaf.
The test to confirm that runtime/trace and runtime/pprof have similar
views of CPU profile samples has trouble with those stacks. The test
confirms that the samples match by confirming that their symbolized
forms match, and the symbolization procedure is very different for the
two packages.
Skip the samples that the CPU profiler's view symbolizes to include one
of runtime.sigprof's sentinel functions at the leaf. (The test design
expects the CPU profiler to under-report samples relative to the
execution tracer.)
Fixes#53378
Change-Id: I60c27de4c69b454057d28a3b6e12d70369de4c4f
Reviewed-on: https://go-review.googlesource.com/c/go/+/413457
Reviewed-by: Bryan Mills <bcmills@google.com>
Run-TryBot: Rhys Hiltner <rhys@justin.tv>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Also, add a benchmark for another case that was originally tested.
Also also, remove all the dead code this now creates.
Fixes#53428.
Change-Id: Idbba88d3d31d38a8854fd5ed99001e394da27300
Reviewed-on: https://go-review.googlesource.com/c/go/+/412878
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Bryan Mills <bcmills@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
This test originally existed as two tests in test/locklinear.go, but
this checked against actual locks and was flaky. The test was checking
a property of a deep part of the runtime but from a much higher level,
and it's easy for nondeterminism due to scheduling to completely mess
that up, especially on an oversubscribed system.
That test was then moved to the sync package with a more rigorous
testing methodology, but it could still flake pretty easily.
Finally, this CL makes semtable more testable, exports it in
export_test.go, then writes a very direct scalability test for exactly
the situation the original test described. As far as I can tell, this is
much, much more stable, because it's single-threaded and is just
checking exactly the algorithm we need to check.
Don't bother trying to bring in a test that checks for O(log n) behavior
on the other kind of iteration. It'll be perpetually flaky because the
underlying data structure is a treap, so it's only _expected_ to be
O(log n), but it's very easy for it to get unlucky without a large
number of iterations that's too much for a simple test.
Fixes#53381.
Change-Id: Ia1cd2d2b0e36d552d5a8ae137077260a16016602
Reviewed-on: https://go-review.googlesource.com/c/go/+/412875
Reviewed-by: Michael Pratt <mpratt@google.com>
All loong64 syscalls return values only via R4/A0, and R5/A1 may contain unrelated
content. Always zero the second return value.
Change-Id: I62af59369bece5bd8028b937c74f4694150f7a55
Reviewed-on: https://go-review.googlesource.com/c/go/+/411615
Run-TryBot: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Austin Clements <austin@google.com>
Previously we stole a bit from the field offset to encode whether
a struct field was embedded.
Instead, encode that bit in the name field, where we already have
some unused bits to play with. The bit associates naturally with
the name in any case.
This leaves a full uintptr to specify field offsets. This will make
the fix for #52740 cleaner.
Change-Id: I0bfb85564dc26e8c18101bc8b432f332176d7836
Reviewed-on: https://go-review.googlesource.com/c/go/+/412138
Reviewed-by: Cherry Mui <cherryyz@google.com>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
We occassionally see _beginthread failing with EACCES, meaning
"insufficient resources" according to the Microsoft documentation.
Exactly which resources is unclear.
Similar to pthread_create on unix systems, we can wait a bit and retry
to try to get success. The alternative is to abort, so we may as well
give it a try.
Fixes#52572.
Change-Id: I6e05add53b4ae36c61e53b1ee3fed6bc74e17dfa
Reviewed-on: https://go-review.googlesource.com/c/go/+/410355
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
The bodies of cgo_sys_thread_start (and x_cgo_sys_thread_create) are
nearly identical on all of the windows ports.
Create a single _cgo_beginthread implementation that contains the body
and is used on all ports. This will reduce churn in an upcoming CL to
add retry logic.
We could theoretically have a single implementation of
_cgo_sys_thread_start shared by all ports, but I keep them separate for
ease of searching. Right now every single port implements this function
in their gcc_GOOS_GOARCH.c file, so it is nice to keep this symmetry.
_cgo_dummy_export must move out of libcgo_windows.h because it is a
definition and the inclusion of libcgo_windows.h in multiple files
creates duplicate definitions.
For #52572.
Change-Id: I9fa22009389349c754210274c7db2631b061f9c7
Reviewed-on: https://go-review.googlesource.com/c/go/+/410354
Run-TryBot: Michael Pratt <mpratt@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
In faketime mode, checkdead is responsible for jumping time forward to
the next timer expiration, and waking an M to handle the newly ready
timer.
Currently it pulls the exact P that owns the next timer off of the pidle
list. In theory this is efficient because that P is immediately eligible
to run the timer without stealing. Unfortunately it is also fraught with
peril because we are skipping all of the bookkeeping in pidleget:
* Skipped updates to timerpMask mean that our timers may not be eligible
for stealing, as they should be.
* Skipped updates to idlepMask mean that our runq may not be eligible
for stealing, as they should be.
* Skipped updates to sched.npidle may break tracking of spinning Ms,
potentially resulting in lost work.
* Finally, as of CL 410122, skipped updates to p.limiterEvent may affect
the GC limiter, or cause a fatal throw when another event occurs.
The last case has finally undercovered this issue since it quickly
results in a hard crash.
We could add all of these updates into checkdead, but it is much more
maintainable to keep this logic in one place and use pidleget here like
everywhere else in the runtime. This means we probably won't wake the
P owning the timer, meaning that the P will need to steal the timer,
which is less efficient, but faketime is not a performance-sensitive
build mode. Note that the M will automatically make itself a spinning M
to make it eligible to steal since it is the only one running.
Fixes#53294
For #52890
Change-Id: I4acc3d259b9b4d7dc02608581c8b4fd259f272e9
Reviewed-on: https://go-review.googlesource.com/c/go/+/411119
Run-TryBot: Michael Pratt <mpratt@google.com>
Auto-Submit: Michael Pratt <mpratt@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Very rarely, GDB will successfully run the whole test and the inferior
will exit successfully, and then GDB itself hangs and never exits.
Detect this and skip the test as flaky. We could just continue the
test since all of the output we need is there, but by skipping it
we're less likely to notice serious regressions in this test.
Fixes#37405.
Change-Id: I016cbb06f48673f064733da3e3f1ddcbefd58159
Reviewed-on: https://go-review.googlesource.com/c/go/+/411117
Reviewed-by: Cherry Mui <cherryyz@google.com>
libfuzzerHookStrCmp is manually reformatted into a proper go doc list.
We don't always format testdata, but these test programs are standard Go
programs that can be formatted.
Change-Id: I4dde398bca225ae8c72e787e4d43fd0ccfd0a90b
Reviewed-on: https://go-review.googlesource.com/c/go/+/411114
Auto-Submit: Michael Pratt <mpratt@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Michael Pratt <mpratt@google.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
CL 389354 removed the fallback to pipe on all platforms with pipe2. This
is the case for linux. Thus, pipe and setNonblock are no longer needed
on linux/loong64 too.
Change-Id: I089adf918d0fd8de5d4d61a893707a2660f89183
Reviewed-on: https://go-review.googlesource.com/c/go/+/410736
Reviewed-by: Ian Lance Taylor <iant@google.com>
Auto-Submit: Tobias Klauser <tobias.klauser@gmail.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Run-TryBot: Ian Lance Taylor <iant@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Use the program counter to compute the address of the first instruction
of the ret sled. The ret sled is located after 5 instructions from the
MOVD instruction saving the value of the program counter.
Change-Id: Ie7ae7a0807785d6fea035cf7a770dba7f37de0ec
GitHub-Last-Rev: 2719208c6a
GitHub-Pull-Request: golang/go#53039
Reviewed-on: https://go-review.googlesource.com/c/go/+/407895
Reviewed-by: Cherry Mui <cherryyz@google.com>
Reviewed-by: Alan Donovan <adonovan@google.com>
A few of these are copied from the memory model doc.
Many are entirely new, following discussion on #47141.
See https://research.swtch.com/gomm for background.
The rule we are establishing is that each type that is meant
to help synchronize a Go program should document its
happens-before guarantees.
For #50859.
Change-Id: I947c40639b263abe67499fa74f68711a97873a39
Reviewed-on: https://go-review.googlesource.com/c/go/+/381316
Auto-Submit: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
Currently runtime.Breakpoint generates a SIGSEGV in ppc64.
The solution is an unconditional trap similar to what clang and gcc do. It is documented in the section C.6 of the ABI Book 3.
Fixes#52101
Change-Id: I071d2f2679b695ef268445b04c9222bd74e1f9af
GitHub-Last-Rev: fff4e5e8ff
GitHub-Pull-Request: golang/go#52102
Reviewed-on: https://go-review.googlesource.com/c/go/+/397554
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Paul Murphy <murp@ibm.com>
Run-TryBot: Paul Murphy <murp@ibm.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
statment -> statement
Change-Id: Ia93a466fdc20157a7d6048903e359fe8717ecb8f
GitHub-Last-Rev: 0a9bc5cab0
GitHub-Pull-Request: golang/go#53231
Reviewed-on: https://go-review.googlesource.com/c/go/+/410374
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Run-TryBot: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Currently the GC CPU limiter only tracks idle GC work time. However, in
very undersubscribed situations, it's possible that all this extra idle
time prevents the enabling of the limiter, since it all gets account for
as mutator time. Fix this by tracking all idle time via pidleget and
pidleput. To support this, pidleget and pidleput also accept and return
"now" parameters like the timer code.
While we're here, let's clean up some incorrect assumptions that some of
the scheduling code makes about "now."
Fixes#52890.
Change-Id: I4a97893d2e5ad1e8c821f8773c2a1d449267c951
Reviewed-on: https://go-review.googlesource.com/c/go/+/410122
Reviewed-by: Michael Pratt <mpratt@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Either due to a new nowritebarrierrec annotation or a change in escape
analysis, printDebuglog can't be called from sighandler anymore.
Fix this by avoiding a string allocation that's the primary culprit.
Change-Id: Ic84873a453f45852b0443a46597ed3ab8c9443fd
Reviewed-on: https://go-review.googlesource.com/c/go/+/410121
Run-TryBot: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Currently the GC CPU limiter consumes CPU time from a few pools, but
because the events that flush to those pools may overlap, rather than be
strictly contained within, the update window for the GC CPU limiter, the
limiter's accounting is ultimately sloppy.
This sloppiness complicates accounting for idle time more completely,
and makes reasoning about the transient behavior of the GC CPU limiter
much more difficult.
To remedy this, this CL adds a field to the P struct that tracks the
start time of any in-flight event the limiter might care about, along
with information about the nature of that event. This timestamp is
managed atomically so that the GC CPU limiter can come in and perform a
read of the partial CPU time consumed by a given event. The limiter also
updates the timestamp so that only what's left over is flushed by the
event itself when it completes.
The end result of this change is that, since the GC CPU limiter is aware
of all past completed events, and all in-flight events, it can much more
accurately collect the CPU time of events since the last update. There's
still the possibility for skew, but any leftover time will be captured
in the following update, and the magnitude of this leftover time is
effectively bounded by the update period of the GC CPU limiter, which is
much easier to consider.
One caveat of managing this timestamp-type combo atomically is that they
need to be packed in 64 bits. So, this CL gives up the top 3 bits of the
timestamp and places the type information there. What this means is we
effectively have only a 61-bit resolution timestamp. This is fine when
the top 3 bits are the same between calls to nanotime, but becomes a
problem on boundaries when those 3 bits change. These cases may cause
hiccups in the GC CPU limiter by not accounting for some source of CPU
time correctly, but with 61 bits of resolution this should be extremely
rare. The rate of update is on the order of milliseconds, so at worst
the runtime will be off of any given measurement by only a few
CPU-milliseconds (and this is directly bounded by the rate of update).
We're probably more inaccurate from the fact that we don't measure real
CPU time but only approximate it.
For #52890.
Change-Id: I347f30ac9e2ba6061806c21dfe0193ef2ab3bbe9
Reviewed-on: https://go-review.googlesource.com/c/go/+/410120
Reviewed-by: Michael Pratt <mpratt@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
This change forces mark and scavenge assists to be cancelled early if
the limiter is enabled. This avoids goroutines getting stuck in really
long assists if the limiter happens to be disabled when they first come
into the assist. This can get especially bad for mark assists, which, in
dire situations, can end up "owing" the GC a really significant debt.
For #52890.
Change-Id: I4bfaa76b8de3e167d49d2ffd8bc2127b87ea566a
Reviewed-on: https://go-review.googlesource.com/c/go/+/408816
Run-TryBot: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Cgo TSAN (not the Go race detector) intercepts signals and calls
the signal handler at a later time. When the signal handler is
called, the memory may have changed, but the signal context
remains old. As the signal context and the memory don't match, it
is unsafe to unwind the stack from the signal PC and SP. We have
to ignore the signal.
It is probably also not safe to do async preemption, which relies
on the signal PC, and inspects and even writes to the stack (for
call injection).
We also inspect the stack for fatal signals (e.g. SIGSEGV), but I
think they are not delayed. For other signals we don't inspect
the stack, so they are probably fine.
Fixes#27540.
Change-Id: I5c80a7512265b8ea4a91422954dbff32c6c3a0d0
Reviewed-on: https://go-review.googlesource.com/c/go/+/408218
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Cherry Mui <cherryyz@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Fixes "GOARCH=loong64 go vet runtime/internal/syscall"
Change-Id: I5879eec3ff07b0c69a5a8ac8e854733261e98fbf
Reviewed-on: https://go-review.googlesource.com/c/go/+/408695
Reviewed-by: Ian Lance Taylor <iant@google.com>
Run-TryBot: Ian Lance Taylor <iant@google.com>
Reviewed-by: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: xiaodong liu <teaofmoli@gmail.com>
Auto-Submit: Ian Lance Taylor <iant@golang.org>
Co-authored-by: limeidan <limeidan@loongson.cn>
Change-Id: I0011a10f831e6c2b0da96265682212b0747f0e2a
Reviewed-on: https://go-review.googlesource.com/c/go/+/407774
Reviewed-by: Cherry Mui <cherryyz@google.com>
Reviewed-by: David Chase <drchase@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: WANG Xuerui <git@xen0n.name>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Run-TryBot: Ian Lance Taylor <iant@google.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
The current value is appropriate for an early in-house version of
Linux/LoongArch, but for the upstream version it is very likely
"LINUX_5.10" instead, per the latest upstream submission [1].
[1]: https://lore.kernel.org/all/20220518095709.1313120-3-chenhuacai@loongson.cn/
Change-Id: Ia97e5cae82a5b306bd3eea86b9e442441da07973
Reviewed-on: https://go-review.googlesource.com/c/go/+/407755
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: David Chase <drchase@google.com>
Run-TryBot: Ian Lance Taylor <iant@google.com>
Reviewed-by: abner chenc <chenguoqi@loongson.cn>
libFuzzer provides a special mode known as “value profiling” in which it
tracks the bit-wise progress made by the fuzzer in satisfying tracked
comparisons. Furthermore, libFuzzer uses the value of the return address
in its hooks to distinguish the progress for different comparisons.
The original implementation of the interception for integer comparisons
in Go simply called the libFuzzer hooks from a function written in Go
assembly. The libFuzzer hooks thus always see the same return address
(i.e., the address of the call instruction in the assembly snippet) and
thus can’t distinguish individual comparisons anymore. This drastically
reduces the usefulness of value profiling.
This is fixed by using an assembly trampoline that injects synthetic but
valid return addresses on the stack before calling the libFuzzer hook,
otherwise preserving the calling convention of the respective platform
(for starters, x86_64 Windows or Unix). These fake PCs are generated
deterministically based on the location of the compare instruction in
the IR representation.
Change-Id: Iea68057c83aea7f9dc226fba7128708e8637d07a
GitHub-Last-Rev: f9184baafd
GitHub-Pull-Request: golang/go#51321
Reviewed-on: https://go-review.googlesource.com/c/go/+/387336
Reviewed-by: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
IR string compares as well as calls to string comparison functions such
as `strings.EqualFold` are intercepted and the corresponding libFuzzer
callbacks are invoked with the corresponding arguments. As a result, the
compared strings will be added to libFuzzer’s table of recent compares,
which feeds future mutations performed by the fuzzer and thus allow it
to reach into branches guarded by string comparisons.
The list of methods to intercept is maintained in
`cmd/compile/internal/walk/expr.go` and can easily be extended to cover
more standard library functions in the future.
Change-Id: I5c8b89499c4e19459406795dea923bf777779c51
GitHub-Last-Rev: 6b8529b555
GitHub-Pull-Request: golang/go#51319
Reviewed-on: https://go-review.googlesource.com/c/go/+/387335
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
Run-TryBot: Keith Randall <khr@golang.org>
Currently, physical-page-aligned allocations for stacks (where the
physical page size is greater than the runtime page size) first
overallocates some memory, then frees the unaligned portions back to the
heap.
However, because allocating via h.pages.alloc causes scavenged bits to
get cleared, we need to account for that memory correctly in heapFree
and heapReleased. Currently that is not the case, leading to throws at
runtime.
Trying to get that accounting right is complicated, because information
about exactly which pages were scavenged needs to get plumbed up.
Instead, find the oversized region first, and then only allocate the
aligned part. This avoids any accounting issues.
However, this does come with some performance cost, because we don't
update searchAddr (which is safe, it just means the next allocation
potentially must look harder) and we skip the fast path that
h.pages.alloc has for simplicity.
Fixes#52682.
Change-Id: Iefa68317584d73b187634979d730eb30db770bb6
Reviewed-on: https://go-review.googlesource.com/c/go/+/407502
Run-TryBot: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
By using libFuzzer’s 8-bit counters instead of extra counters, the
coverage instrumentation in libFuzzer mode is improved in three ways:
1- 8-bit counters are supported on all platforms, including macOS and
Windows, with all relevant versions of libFuzzer, whereas extra
counters are a Linux-only feature that only recently received
support on Windows.
2- Newly covered blocks are now properly reported as new coverage by
libFuzzer, not only as new features.
3- The NeverZero strategy is used to ensure that coverage counters
never become 0 again after having been positive once. This resolves
issues encountered when fuzzing loops with iteration counts that
are multiples of 256 (e.g., larger powers of two).
Change-Id: I9021210d7fbffd07c891ad08750402ee91cb3df5
GitHub-Last-Rev: 9057e4b21d
GitHub-Pull-Request: golang/go#51318
Reviewed-on: https://go-review.googlesource.com/c/go/+/387334
Reviewed-by: Keith Randall <khr@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
For #48409.
Change-Id: I056afcdbc417ce633e48184e69336213750aae28
Reviewed-on: https://go-review.googlesource.com/c/go/+/406575
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Contributors to the loong64 port are:
Weining Lu <luweining@loongson.cn>
Lei Wang <wanglei@loongson.cn>
Lingqin Gong <gonglingqin@loongson.cn>
Xiaolin Zhao <zhaoxiaolin@loongson.cn>
Meidan Li <limeidan@loongson.cn>
Xiaojuan Zhai <zhaixiaojuan@loongson.cn>
Qiyuan Pu <puqiyuan@loongson.cn>
Guoqi Chen <chenguoqi@loongson.cn>
This port has been updated to Go 1.15.6:
https://github.com/loongson/go
Updates #46229
Change-Id: I61bef32b38ab07543a147cf172b169eae21b26cf
Reviewed-on: https://go-review.googlesource.com/c/go/+/368082
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Ian Lance Taylor <iant@google.com>
Contributors to the loong64 port are:
Weining Lu <luweining@loongson.cn>
Lei Wang <wanglei@loongson.cn>
Lingqin Gong <gonglingqin@loongson.cn>
Xiaolin Zhao <zhaoxiaolin@loongson.cn>
Meidan Li <limeidan@loongson.cn>
Xiaojuan Zhai <zhaixiaojuan@loongson.cn>
Qiyuan Pu <puqiyuan@loongson.cn>
Guoqi Chen <chenguoqi@loongson.cn>
This port has been updated to Go 1.15.6:
https://github.com/loongson/go
Updates #46229
Change-Id: Ide01fb8a39fe3e890f6cbc5d28f4a1d47eb5d79b
Reviewed-on: https://go-review.googlesource.com/c/go/+/368081
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Run-TryBot: Ian Lance Taylor <iant@google.com>
Contributors to the loong64 port are:
Weining Lu <luweining@loongson.cn>
Lei Wang <wanglei@loongson.cn>
Lingqin Gong <gonglingqin@loongson.cn>
Xiaolin Zhao <zhaoxiaolin@loongson.cn>
Meidan Li <limeidan@loongson.cn>
Xiaojuan Zhai <zhaixiaojuan@loongson.cn>
Qiyuan Pu <puqiyuan@loongson.cn>
Guoqi Chen <chenguoqi@loongson.cn>
This port has been updated to Go 1.15.6:
https://github.com/loongson/go
Updates #46229
Change-Id: I252ecd0b13580c5e71723715023b1951985045f3
Reviewed-on: https://go-review.googlesource.com/c/go/+/342322
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Run-TryBot: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
This benchmark is added to test improvements in memclr_amd64.
As it is stated in Intel Optimization Manual 15.16.3.3, AVX2-implemented
memclr can produce a skewed result with the branch predictor being
trained by the large loop iteration count.
This benchmark generates sizes between some specified range. This should
help to measure how memclr works when branch predictors may be incorrectly
trained.
Change-Id: I14d173cafe43ca47198ed920e655547a66b3909f
Reviewed-on: https://go-review.googlesource.com/c/go/+/373362
Reviewed-by: Keith Randall <khr@golang.org>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Keith Randall <khr@golang.org>
Auto-Submit: Michael Pratt <mpratt@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Reviewed-by: Keith Randall <khr@google.com>
On 32-bit systems the result of hogCount*factor can overflow.
Use division instead to do comparison.
Update #52207
Change-Id: I429fb9dc009af645acb535cee5c70887527ba207
Reviewed-on: https://go-review.googlesource.com/c/go/+/407415
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Keith Randall <khr@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
Contributors to the loong64 port are:
Weining Lu <luweining@loongson.cn>
Lei Wang <wanglei@loongson.cn>
Lingqin Gong <gonglingqin@loongson.cn>
Xiaolin Zhao <zhaoxiaolin@loongson.cn>
Meidan Li <limeidan@loongson.cn>
Xiaojuan Zhai <zhaixiaojuan@loongson.cn>
Qiyuan Pu <puqiyuan@loongson.cn>
Guoqi Chen <chenguoqi@loongson.cn>
This port has been updated to Go 1.15.6:
https://github.com/loongson/go
Updates #46229
Change-Id: I8ef0e7f17d6ada3d2f07c81524136b78457e7795
Reviewed-on: https://go-review.googlesource.com/c/go/+/342319
Run-TryBot: Ian Lance Taylor <iant@google.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Run-TryBot: David Chase <drchase@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
Contributors to the loong64 port are:
Weining Lu <luweining@loongson.cn>
Lei Wang <wanglei@loongson.cn>
Lingqin Gong <gonglingqin@loongson.cn>
Xiaolin Zhao <zhaoxiaolin@loongson.cn>
Meidan Li <limeidan@loongson.cn>
Xiaojuan Zhai <zhaixiaojuan@loongson.cn>
Qiyuan Pu <puqiyuan@loongson.cn>
Guoqi Chen <chenguoqi@loongson.cn>
This port has been updated to Go 1.15.6:
https://github.com/loongson/go
Updates #46229
Change-Id: I0333503db044c6f39df2d7f8d9dff213b1361d6c
Reviewed-on: https://go-review.googlesource.com/c/go/+/342320
Reviewed-by: David Chase <drchase@google.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Make sure that all the targets of 64-bit atomic operations
are actually aligned to 8 bytes. This has been a source of
bugs on 32-bit systems. (e.g. CL 399754)
The strategy is to have a simple test that just checks the
alignment of some explicitly listed fields and global variables.
Then there's a more complicated test that makes sure the list
used in the simple test is exhaustive. That test has some
limitations, but it should catch most cases, particularly new
uses of atomic operations on new or existing fields.
Unlike a runtime assert, this check is free and will catch
accesses that occur even in very unlikely code paths.
Change-Id: I25ac78df471ac33b57cb91375bd8453d6ce2814f
Reviewed-on: https://go-review.googlesource.com/c/go/+/407034
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Reviewed-by: David Chase <drchase@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Currently gctrace and gcpacertrace recompute the heap goal for
end-of-cycle information but this is incorrect.
Because both of these traces are printing stats from the previous cycle
in this case, they should print the heap goal at the end of the previous
cycle.
Change-Id: I967621cbaff9f331cd3e361de8850ddfe0cfc099
Reviewed-on: https://go-review.googlesource.com/c/go/+/407138
Reviewed-by: Michael Pratt <mpratt@google.com>
Reviewed-by: David Chase <drchase@google.com>
