Closed Bug 1275199 Opened 9 years ago Closed 9 years ago

Crashes when running mochitest with GCC 6.1 build

Categories

(Core :: JavaScript Engine: JIT, defect)

defect
Not set
normal

Tracking

()

RESOLVED DUPLICATE of bug 1245783

People

(Reporter: xidorn, Unassigned)

Details

After applying patches from bug 1269171, I successfully build firefox with GCC 6.1.1 on Arch Linux. But I cannot run mochitest normally. The test usually stops very quick with > IOError: Connection to Marionette server is lost. Check gecko.log (desktop firefox) or logcat (b2g) for errors. > 0 ERROR Automation Error: Received unexpected exception while running application And if e10s is not disabled, also: > #01: mozilla::ipc::ProcessLink::OnChannelError() (/data/mozilla/central/ipc/glue/MessageLink.cpp:404) > #02: event_process_active_single_queue (/data/mozilla/central/ipc/chromium/src/third_party/libevent/event.c:1355) > #03: base::MessagePumpLibevent::Run(base::MessagePump::Delegate*) (/data/mozilla/central/ipc/chromium/src/base/message_pump_libevent.cc:372) > #04: MessageLoop::RunInternal() (/data/mozilla/central/ipc/chromium/src/base/message_loop.cc:236) > #05: MessageLoop::AutoRunState::~AutoRunState() (/data/mozilla/central/ipc/chromium/src/base/message_loop.cc:493) > #06: base::Thread::ThreadMain() (/data/mozilla/central/ipc/chromium/src/base/thread.cc:183) > #07: ThreadFunc(void*) (/data/mozilla/central/ipc/chromium/src/base/platform_thread_posix.cc:40) > #08: start_thread (pthread_create.c:?) > #09: __clone (/usr/lib/libc.so.6) > #10: ??? (???:???) > [Child 5454] ###!!! ABORT: Aborting on channel error.: file /data/mozilla/central/ipc/glue/MessageChannel.cpp, line 2050 > Hit MOZ_CRASH() at /data/mozilla/central/memory/mozalloc/mozalloc_abort.cpp:33 And when I specify --debugger=gdb, I see the following backtrace when it crashes: > Thread 1 "firefox" received signal SIGSEGV, Segmentation fault. > 0x00007fffc493d9ff in ?? () > (gdb) bt > #0 0x00007fffc493d9ff in ?? () > #1 0x00007fffe94628cf in js::jit::SnapshotIterator::numAllocations (this=0x7fffffffa9f0) at /data/mozilla/central/js/src/jit/JitFrames.cpp:2022 > #2 js::jit::IonFrameStackDepthOp::IonFrameStackDepthOp (frame=..., this=<optimized out>) at /data/mozilla/central/js/src/jit/JitFrames.cpp:358 > #3 js::jit::TryNoteIterIon::TryNoteIterIon (frame=..., cx=0x7fffe3587400, this=0x7fffffffa970) at /data/mozilla/central/js/src/jit/JitFrames.cpp:369 > #4 js::jit::HandleExceptionIon (overrecursed=0x7fffffffa8df, rfe=0x7fffffffae78, frame=..., cx=0x7fffe3587400) at /data/mozilla/central/js/src/jit/JitFrames.cpp:416 > #5 js::jit::HandleException (rfe=<optimized out>) at /data/mozilla/central/js/src/jit/JitFrames.cpp:791 > #6 0x00007ffff7e9363d in ?? () > #7 0x0000000000000008 in ?? () > #8 0x00007fffffffae78 in ?? () > #9 0x0000000000000000 in ?? () And this is why I set the component to JavaScriptEngine: JIT here. I can stably reproduce this with |./mach mochitest dom/tests/mochitest/pointerlock/|, either with or without e10s. |./mach mochitest| sometimes runs fine for a long time, sometimes crashes immediately. I'm using an install of Arch Linux inside VMware Workstation 11.1.3 build-3206955. My compiler is GCC 6.1.1 20160501.
It should work with these 2 patches (one bug is still on inbound after 2-3 days): https://bugzilla.mozilla.org/show_bug.cgi?id=1245783#c36 Can you confirm that fixes it?
Flags: needinfo?(bugzilla)
Yes, that fixes it. Thanks!
Status: NEW → RESOLVED
Closed: 9 years ago
Flags: needinfo?(bugzilla)
Resolution: --- → DUPLICATE
You need to log in before you can comment on or make changes to this bug.