These are slightly frustrating. First, when a struct is packed, some of
its fields may be underaligned. This is fine for direct access
(foo.bar), but if one takes the address if the field, this creates an
unaligned pointer. Dereferencing that pointer is then UB. (I'm not sure
if creating that pointer is UB.)
Crashpad seemingly doesn't do this, but it uses EXPECT_EQ from GTest.
EXPECT_EQ seems to internally take pointers to its arguments. I'm
guessing it binds them by const reference. This then trips UBSan. To
avoid this, we can copy the value into a temporary before passing to
EXPECT_EQ.
Second, the test to divide by 0 to trigger SIGFPE is undefined behavior.
The compiler is not actually obligated to trip SIGFPE. UBSan prints one
of its errors instead. Instead, since this file is only built on POSIX
anyway, use GCC inline assembly to do the division. That one is
well-defined.
Finally, casting a string to uint32_t* is undefined both by alignment
and by strict aliasing (although Chromium doesn't enable the latter).
Instead, type-punning should be done with memcpy.
Bug: chromium:1394755
Change-Id: I79108773a04ac26f5189e7b88a0acbf62eb4401d
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4985905
Reviewed-by: Robert Sesek <rsesek@chromium.org>
Commit-Queue: David Benjamin <davidben@chromium.org>
Fixes a pending issue when we eventually move to C++20.
Original author: Dean Sturtevant
Change-Id: I7bb0648c73df6b6a28a3a4debdb4524d3cd27b38
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4979733
Reviewed-by: Justin Cohen <justincohen@chromium.org>
Commit-Queue: Eric Astor <epastor@google.com>
Include check_op.h directly, instead of relying on the transitive
include from logging.h. This transitive include does not exist in
Chromium's //base.
Change-Id: I15962a9cdc26ac206032157b8d2659cf263ad695
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4950200
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Lei Zhang <thestig@chromium.org>
Some users of crashpad load and unload the dll that hosts
crashpad code. crashpad registers a vectored exception handler
to help collect heap corruption crashes. If the dll is
unloaded this handler might still be called.
This CL adds a scoped handler for such registrations and
uses it on Windows crashpad client. To allow this to
be stored, RegisterHandler() on the client needs to move
onto the client object from being a helper function.
Bug: crashpad:462
Change-Id: I5d77c056e2a9a61ddcfa9d0186ab4bfd85a19bff
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4898263
Reviewed-by: Ben Hamilton <benhamilton@google.com>
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Commit-Queue: Alex Gough <ajgo@chromium.org>
This rolls mini_chromium to the version that has more files in
base/apple, and adjusts the code to match.
Bug: chromium:1444927
Change-Id: I9642698c8c16151bd0aaca7b46745a59d6e5e6d3
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4791121
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Avi Drissman <avi@chromium.org>
This CL rolls mini_chromium to pick up the move of a bunch of files
to base/apple, and makes changes to adjust.
Bug: chromium:1444927
Change-Id: Ib692e2a1628e2c0c8228795eaecdb7f35b1c09fa
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4786387
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Avi Drissman <avi@chromium.org>
This reverts commit ca6d64d0ae4905ad7033adab0a28273a0741ee5c.
Reason for revert: The changes did not actually fix the problem once combined with the latest changes from mini_chromium.
Original change's description:
> [fuchsia][mac] Fix build errors
>
> A recent CL [1] broke Fuchsia's Crashpad roller due to duplicate build
> argument declarations. This CL ensures that sysroot.gni is only imported once.
>
> [1] https://chromium-review.googlesource.com/c/chromium/mini_chromium/+/4651973
>
> Fixed: fuchsia:131454
> Change-Id: Idcf6ac65cdffee2c9a9551559a8aab0063044428
> Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4743381
> Reviewed-by: Joshua Peraza <jperaza@chromium.org>
> Commit-Queue: Thomas Gales <tgales@google.com>
Change-Id: Id3dc42484fbd87e242756c8d2889d2e404370ac7
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4753637
Commit-Queue: Thomas Gales <tgales@google.com>
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
ARC is now enabled by default, so there’s no need to enforce it
against files being put into non-ARC targets.
Bug: chromium:1468376
Change-Id: I58bbb4d1736293a6e9977954ce932dcfe2bafa54
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4750419
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Mark Mentovai <mark@chromium.org>
CaptureContext isn't actually used on Fuchsia and there is a desire to
remove `ucontext_t` from Fuchsia as it isn't a real concept on Fuchsia
and was only added as a placeholder. Moreover, `ucontext_t` won't ever
be added to Fuchsia for RISC-V.
Bug: fuchsia:123052
Fixed: fuchsia:131112
Fixed: fuchsia:127655
Tested: `fx test crashpad` on core.x64 emulator
Tested: `fx test crashpad` on ARM64 device
Tested: `fx shell crasher` @ 16b19a9891978487 on ARM64 device, ran
through Breakpad stackwalker locally as well
Tested: `fx build crashpad_tests` for minimal.riscv64
Change-Id: I4695054426df78a9deff8c9ea9c478b5bf9701b1
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4717085
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Thomas Gales <tgales@google.com>
Only RV64GC is supported.
RISC-V Fuchsia is not able to serve packages yet so unit testing is not
possible.
Bug: fuchsia:127655
Tested: `crasher` with crashpad added to crashsvc, ran minidump through
Breakpad stackwalker
Change-Id: I1b6d79128759281aee348e333ea15434ab397001
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4602412
Reviewed-by: Mark Mentovai <mark@chromium.org>
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Only RV64GC is supported.
Bug: fuchsia:127655
Tested: `python build/run_tests.py` on RISC-V emulator
Tested: Created minidump via self-induced crash on RISC-V emulator,
ran through Breakpad stackwalker
Change-Id: I713797cd623b0a758269048e01696cbce502ca6c
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4581050
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Mac OS X Server has been discontinued as a separate operating system
flavor since 10.6. Current minimal requirements for both Crashpad and
Chromium are above that.
Change-Id: Ia9063be2e55a48e45d9f9974ac2e51bac004f37d
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4584570
Reviewed-by: Mark Mentovai <mark@chromium.org>
__has_feature is a clang extension. GCC errors out on the test.
Define a helper macro to make the code working with other compilers.
Bug: chromium:819294
Change-Id: I359150acd4700e65b4faf5f297b29664c18000d3
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4418706
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Commit-Queue: Joshua Peraza <jperaza@chromium.org>
Reviewed-by: Ben Hamilton <benhamilton@google.com>
Always reset the file descriptor to -1, even if FlushWriteBuffer or
RawLoggingCloseFile fails.
Bug: 1431760
Change-Id: I193f526d65f477bba002dd9faf68996020e48a3b
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4406657
Reviewed-by: Ben Hamilton <benhamilton@google.com>
Commit-Queue: Justin Cohen <justincohen@chromium.org>
Reviewed-by: Robert Sesek <rsesek@chromium.org>
Pointer Authentication works by adding a signature to the top bits of
an instruction or data pointer (only instruction pointers on the stack
are currently signed in Chromium). This can confuse range checks,
because they need to strip the top bits. Masking these bits during sanitization range checks prevents confusion.
Test: Testing was done manually on a device with pointer authentication enabled.
Bug: crashpad:364
Bug: 919548
Change-Id: I2e739cadb2844cfaf73a75596d664135aeb5faac
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4387271
Commit-Queue: Adam Walls <avvall@google.com>
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Reviewed-by: Ben Hamilton <benhamilton@google.com>
Several tests in filesystem_test.cc create symbol links. The privilege
needed to do this is not enabled on all Windows systems so several of
the tests check for the privilege and are skipped if it is not
available.
However, two tests that created symbol links were not doing this check
and therefore failed on some Windows machines. This corrects those
failures by adding the checks.
Bug: chromium:1418165
Change-Id: I6621796b462b8db02271ad5a05e0c29ee047f648
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4348801
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Bruce Dawson <brucedawson@chromium.org>
Missed this the first time around because it was Windows-only.
Bug: chromium:691162
Change-Id: Ic98a5943957f77fbf17d92a93409eaa35910ae0e
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4297482
Commit-Queue: David Benjamin <davidben@chromium.org>
Reviewed-by: Robert Sesek <rsesek@chromium.org>
This also significantly simplifies the implementation, since we don't
really need the ThreadLogMessagesMaster class at all.
Bug: chromium:1416710
Change-Id: I85849230015f901dfbf084d140e639f14cb872a7
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4313281
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Peter Kasting <pkasting@chromium.org>
This is the only change needed to build crashpad against musl, yay! The
reason this change is needed is that user_vfp is bionic-specific, and
does not exist in glibc, dietlibc, uclibc, or musl.
I have not (yet) tried running the tests against another libc.
Bug: chromium:1380656
Change-Id: I2247352e1611a300dff995156d393508c8257039
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4255370
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Elly Fong-Jones <ellyjones@chromium.org>
This CL introduces a new class ScopedVMMap, a fork of ScopedVMRead
which maps the memory using vm_remap() instead of reading it.
This is useful for Annotations which use ScopedSpinGuard to
protect reads from simultaneous writes; the in-process intermediate
dump handler can try to take the spin guard when reading such
an Annotation and skip reading it if it the spin guard could not
be obtained.
Change-Id: I60d7a48d1ba4e5d2dfdb44307b78b4d9ffb73560
Bug: crashpad:437
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4114550
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Ben Hamilton <benhamilton@google.com>
ARM64 supports storing pointer authentication codes in the upper bits of
a pointer. This mask can be used by LLDB to mimic ptrauth_strip and
strip the pointer authentication codes. To recover an address from
pointer with an authentication code, `AND` this mask with the pointer.
If the platform does not support pointer authentication, or the range of
valid addressees for a pointer was unaccessible, this field will be 0
and should be ignored.
Change-Id: Ie5cef90802dd1e892d456195ab8874223eac6a1b
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/2773358
Commit-Queue: Justin Cohen <justincohen@chromium.org>
Reviewed-by: Mark Mentovai <mark@chromium.org>
The try and catch macros were conditionally defined by libstdc++ pre-gcc
4.4 (2009-04-21), fixed in
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=25191. Surely none of this
code would build with such an old libstdc++ any more, since Crashpas has
adopted modern C++ (C++11 and later). Remove this obsolete nod to
history.
Change-Id: Ie3cea1ecc1cfd358f27ea48f8111791e7f08bfa5
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4136890
Reviewed-by: Robert Sesek <rsesek@chromium.org>
Commit-Queue: Mark Mentovai <mark@chromium.org>
... as we move crashpad from //third_party/crashpad to
//third_party/crashpad/src
Change-Id: I081520ad44334cc83397234e5d16535d0db4806d
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4132465
Reviewed-by: Francois Rousseau <frousseau@google.com>
Commit-Queue: Jay Zhuang <jayzhuang@google.com>
To support a new crashpad::RingBufferAnnotation type which can be safely
written to and read from simultaneously by different threads/processes,
this CL introduces a new class ScopedSpinGuard, which is a simple RAII
wrapper around an atomic boolean.
Change-Id: I5bafe6927a8dc2a3e25734cb941fd9fce9a8d139
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4031729
Commit-Queue: Ben Hamilton <benhamilton@google.com>
Reviewed-by: Robert Sesek <rsesek@chromium.org>
This bug was found when trying to upgrading the MSAN bots from Ubuntu
18.04 (where this codepath was not hit) to 20.04. The following MSAN
error is produced when running HTTPTransport/HTTPTransport.*
==3496553==WARNING: MemorySanitizer: use-of-uninitialized-value
#0 0x5616c540ad7d in __is_long buildtools/third_party/libc++/trunk/include/string:1674:33
#1 0x5616c540ad7d in size buildtools/third_party/libc++/trunk/include/string:1069:17
#2 0x5616c540ad7d in crashpad::(anonymous namespace)::HTTPTransportLibcurl::WriteResponseBody(char*, unsigned long, unsigned long, void*) third_party/crashpad/crashpad/util/net/http_transport_libcurl.cc:528:50
...
SUMMARY: MemorySanitizer: use-of-uninitialized-value buildtools/third_party/libc++/trunk/include/string:1674:33 in __is_long
ORIGIN: invalid (0). Might be a bug in MemorySanitizer origin tracking.
The memory is initialized in http_transport_test.cc:293, but MSAN gets
confused. Given the message output by MSAN (ORIGIN: invalid (0).
Might be a bug in MemorySanitizer origin tracking), this appears
to be a bug in MSAN, not crashpad, so this CL suppresses the error.
Bug: chromium: 1260217
Change-Id: I2d6a46e3489816270cc1fee776793ffafe0147e4
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/4015160
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Commit-Queue: Thomas Anderson <thomasanderson@chromium.org>
Copy of crrev.com/c/3952963.
Fixes locking not working on some Android filesystems due to flock not
being available. Instead, we now use the same approach as Fuchsia with
a dedicated lock file. This is an issue when running tests on
non-rooted Android devices, as we need files to be written to a
location accessible without root, but the chosen location might not
have flock support.
Bug: chromium:1358240
Change-Id: Ie910481be472403a8b0e9e36100594b0618f85e6
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/3999273
Commit-Queue: Brian Sheedy <bsheedy@chromium.org>
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
sed -i '' -E -e 's/Copyright (.+) The Crashpad Authors\. All rights reserved\.$/Copyright \1 The Crashpad Authors/' $(git grep -El 'Copyright (.+) The Crashpad Authors\. All rights reserved\.$')
Bug: chromium:1098010
Change-Id: I8d6138469ddbe3d281a5d83f64cf918ec2491611
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/3878262
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Commit-Queue: Mark Mentovai <mark@chromium.org>
registration_protocol_win.h includes <string>, which adds an
unacceptable dependency on libc++ in //components/crash/win:chrome_wer
in Chrome as that file is included in crashpad_wer.cc. Rather than
remove <string>, which would require doing a lot of transitive
refactoring work in Crashpad, we just extract the data structures into
another file, as crashpad_wer.cc only includes
registration_protocol_win.h for its struct definitions.
Bug: chromium:1357827
Change-Id: Ic20c2952be07ea75d063702cd346cdca0ab65038
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/3864251
Commit-Queue: Alan Zhao <ayzhao@google.com>
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
* crashpad_http_transport_impl is "socket" when targeting Fuchsia
so the dependency on //third_party/curl:libcurl isn't actually
ever added - we might as well remove it to prevent confusion
Bug: fuchsia:107235
TESTED=`fx build` in Fuchsia checkout
Change-Id: I75da6e7505f8ab09f9978472e93c48600f4c35cb
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/3840964
Commit-Queue: Francois Rousseau <frousseau@google.com>
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Passing -1 (or size_t max) to ScopedVMRead would succeed, because the
amount of memory to be read would overflow vm_address_t/vm_size_t and
turn into something reasonable. ScopedVMRead would return true having
only read a miniscule subset of the requested data length.
Bug: 1348341
Change-Id: I061a1d86928f211c541a6378a78ee045d489a838
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/3791710
Commit-Queue: Justin Cohen <justincohen@chromium.org>
Reviewed-by: Mark Mentovai <mark@chromium.org>
Adds a 4K buffer to the intermediate dump writer. Aside from the final
flush, only write in multiples of 4K. This saves between 30ms and 50ms
on an iPhone 12 Pro.
Change-Id: Icc4b222477bd91fd6952c7cf43b105e1f7a50adb
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/3764243
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Justin Cohen <justincohen@chromium.org>
On iOS, holding a lock during a slow upload can lead to watchdog kills
if the app is suspended mid-upload. Instead, if the client can obtain
the lock, the database sets a lock-time file attribute and releases the
flock. The file attribute is cleared when the upload is completed. The
lock-time attribute can be used to prevent file access from other
processes, or to discard reports that likely were terminated mid-upload.
Bug:chromium:1342051
Change-Id: Ib878f6ade8eae467ee39acb52288296759c84582
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/3739019
Reviewed-by: Robert Sesek <rsesek@chromium.org>
Commit-Queue: Justin Cohen <justincohen@chromium.org>
Reviewed-by: Mark Mentovai <mark@chromium.org>
Adds a new IOSIntermediateDumpWriter::AddPropertyCString method which
takes an address to a cstring of unknown length and page-by-page
searches for a NUL-byte terminator.
This is necessary because currently WriteModuleInfo calls strlen
directly on the dyld and module filePath without first using vm_read.
On iOS14 this occasionally crashes, and is generally unwise. Instead,
use AddPropertyCString.
This patch also removes WriteDyldErrorStringAnnotation, as it's no
longer used going forward with iOS 15.
Bug: 1332862
Change-Id: I3801693bc39259a0127e5175dccf286a1cd97ba7
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/3689516
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Justin Cohen <justincohen@chromium.org>
This adds a runtime exception helper (& test module) for Windows and
plumbing to allow the module to be registered by the crashpad client,
and to trigger the crashpad handler. Embedders can build their own
module to control which exceptions are passed to the handler.
See: go/chrome-windows-runtime-exception-helper for motivation.
When registered (which is the responsibility of the embedding
application), the helper is loaded by WerFault.exe when Windows
Error Reporting receives crashes that are not caught by crashpad's
normal handlers - for instance a control-flow violation when a
module is compiled with /guard:cf.
Registration:
The embedder must arrange for the full path to the helper to
be added in the appropriate Windows Error Reporting\
RuntimeExceptionHelperModules registry key.
Once an embedder's crashpad client is connected to a crashpad
handler (e.g. through SetIpcPipeName()) the embedder calls
RegisterWerModule. Internally, this registration includes handles
used to trigger the crashpad handler, an area reserved to hold an
exception and context, and structures needed by the crashpad handler.
Following a crash:
WerFault.exe handles the crash then validates and loads the helper
module. WER hands the helper module a handle to the crashing target
process and copies of the exception and context for the faulting thread.
The helper then copies out the client's registration data and
duplicates handles to the crashpad handler, then fills back the various structures in the paused client that the crashpad handler will need.
The helper then signals the crashpad handler, which collects a dump then
notifies the helper that it is done.
Support:
WerRegisterExceptionHelperModule has been availble since at least
Windows 7 but WerFault would not pass on the exceptions that crashpad
could not already handle. This changed in Windows 10 20H1 (19041),
which supports HKCU and HKLM registrations, and passes in more types of
crashes. It is harmless to register the module for earlier versions
of Windows as it simply won't be loaded by WerFault.exe.
Tests:
snapshot/win/end_to_end_test.py has been refactored slightly to
group crash generation and output validation in main() by breaking
up RunTests into smaller functions.
As the module works by being loaded in WerFault.exe it is tested
in end_to_end_test.py.
Bug: crashpad:133, 866033, 865632
Change-Id: Id668bd15a510a24c79753e1bb03e9456f41a9780
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/3677284
Reviewed-by: Joshua Peraza <jperaza@chromium.org>
Commit-Queue: Alex Gough <ajgo@chromium.org>
This is a reland of 460943dd9a71dc76f68182a8ede766d5543e5341
Original change's description:
> The DoubleForkAndExec() function was taking over 622 milliseconds to run
> on macOS 11 (BigSur) on Intel i5-1038NG7. I did some debugging by adding
> some custom traces and found that the fork() syscall is the bottleneck
> here, i.e., the first fork() takes around 359 milliseconds and the
> nested fork() takes around 263 milliseconds. Replacing the nested fork()
> and exec() with posix_spawn() reduces the time consumption to 257
> milliseconds!
>
> See https://github.com/libuv/libuv/pull/3064 to know why fork() is so
> slow on macOS and why posix_spawn() is a better replacement.
>
> Another point to note is that even base::LaunchProcess() from Chromium
> calls posix_spawnp() on macOS -
> 8f8d82dea0:base/process/launch_mac.cc;l=295-296
The reland isolates the change to non-Android POSIX systems because
posix_spawn and posix_spawnp are available in Android NDK 28, but
Chromium is building with version 23.
Change-Id: If44629f5445bb0e3d0a1d3698b85f047d1cbf04f
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/3721655
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Mark Mentovai <mark@chromium.org>
This reverts commit 460943dd9a71dc76f68182a8ede766d5543e5341.
Reason for revert: This fails to compile in Chromium Android.
posix_spawn and posix_spawnp are available in Android NDK 28, but
Chromium is building with version 23.
https://ci.chromium.org/ui/p/chromium/builders/try/android_compile_dbg/1179765/overview
Original change's description:
> posix: Replace DoubleForkAndExec() with ForkAndSpawn()
>
> The DoubleForkAndExec() function was taking over 622 milliseconds to run
> on macOS 11 (BigSur) on Intel i5-1038NG7. I did some debugging by adding
> some custom traces and found that the fork() syscall is the bottleneck
> here, i.e., the first fork() takes around 359 milliseconds and the
> nested fork() takes around 263 milliseconds. Replacing the nested fork()
> and exec() with posix_spawn() reduces the time consumption to 257
> milliseconds!
>
> See https://github.com/libuv/libuv/pull/3064 to know why fork() is so
> slow on macOS and why posix_spawn() is a better replacement.
>
> Another point to note is that even base::LaunchProcess() from Chromium
> calls posix_spawnp() on macOS -
> 8f8d82dea0:base/process/launch_mac.cc;l=295-296
>
> Change-Id: I25c6ee9629a1ae5d0c32b361b56a1ce0b4b0fd26
> Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/3641386
> Reviewed-by: Mark Mentovai <mark@chromium.org>
> Commit-Queue: Mark Mentovai <mark@chromium.org>
Change-Id: I7f6161bc4734c50308438cdde1e193023ee9bfb8
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/3719439
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Justin Cohen <justincohen@chromium.org>
The DoubleForkAndExec() function was taking over 622 milliseconds to run
on macOS 11 (BigSur) on Intel i5-1038NG7. I did some debugging by adding
some custom traces and found that the fork() syscall is the bottleneck
here, i.e., the first fork() takes around 359 milliseconds and the
nested fork() takes around 263 milliseconds. Replacing the nested fork()
and exec() with posix_spawn() reduces the time consumption to 257
milliseconds!
See https://github.com/libuv/libuv/pull/3064 to know why fork() is so
slow on macOS and why posix_spawn() is a better replacement.
Another point to note is that even base::LaunchProcess() from Chromium
calls posix_spawnp() on macOS -
8f8d82dea0:base/process/launch_mac.cc;l=295-296
Change-Id: I25c6ee9629a1ae5d0c32b361b56a1ce0b4b0fd26
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/3641386
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Mark Mentovai <mark@chromium.org>
Sanitizers can prevent the installation of signal handlers, but
sigaction would still return 0 (for success). Detect this by checking
the installed signal handler via a second call to sigaction.
R=mark@chromium.org
Bug: chromium:1328749
Change-Id: I62a5777379ec5c6b1ca2d5a62e7cd3fb8ed1437b
Reviewed-on: https://chromium-review.googlesource.com/c/crashpad/crashpad/+/3702302
Reviewed-by: Mark Mentovai <mark@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
