GpuProcessHost Fails to Init in --single-process |
|||
Issue description
We currently are having a crash when trying to run mash_browser_tests with --single-process:
hide_tests ./out/${OUT_DIR}/browser_tests --run-in-mash --single-process --gtest_filter=BrowserTest.NoTitle
[22757:22780:0525/164228.915583:ERROR:proxy_service_factory.cc(116)] Cannot use V8 Proxy resolver in single process mode.
Received signal 11 SEGV_MAPERR 000000000000
#0 0x7f00bfb11eeb base::debug::StackTrace::StackTrace()
#1 0x7f00bfb10c2c base::debug::StackTrace::StackTrace()
#2 0x7f00bfb119ff base::debug::(anonymous namespace)::StackDumpSignalHandler()
#3 0x7f00bff73330 <unknown>
#4 0x7f00b75021be content::GpuProcessHost::Init()
#5 0x7f00b7501b28 content::GpuProcessHost::Get()
#6 0x7f00b74ca865 content::BrowserGpuChannelHostFactory::EstablishRequest::EstablishOnIO()
#7 0x7f00b6504a87 _ZN4base8internal13FunctorTraitsIMN7content17WebFileSystemImpl23WaitableCallbackResultsEFvvEvE6InvokeIRK13scoped_refptrIS4_EJEEEvS6_OT_DpOT0_
#8 0x7f00b74cd591 _ZN4base8internal12InvokeHelperILb0EvE8MakeItSoIRKMN7content28BrowserGpuChannelHostFactory16EstablishRequestEFvvEJRK13scoped_refptrIS6_EEEEvOT_DpOT0_
#9 0x7f00b74cd532 _ZN4base8internal7InvokerINS0_9BindStateIMN7content28BrowserGpuChannelHostFactory16EstablishRequestEFvvEJ13scoped_refptrIS5_EEEEFvvEE7RunImplIRKS7_RKSt5tupleIJS9_EEJLm0EEEEvOT_OT0_NS_13IndexSequenceIJXspT1_EEEE
#10 0x7f00b74cd47c _ZN4base8internal7InvokerINS0_9BindStateIMN7content28BrowserGpuChannelHostFactory16EstablishRequestEFvvEJ13scoped_refptrIS5_EEEEFvvEE3RunEPNS0_13BindStateBaseE
#11 0x7f00bfad132e _ZNO4base8CallbackIFvvELNS_8internal8CopyModeE0ELNS2_10RepeatModeE0EE3RunEv
#12 0x7f00bfb173e1 base::debug::TaskAnnotator::RunTask()
#13 0x7f00bfba1a1e base::MessageLoop::RunTask()
#14 0x7f00bfba1c87 base::MessageLoop::DeferOrRunPendingTask()
#15 0x7f00bfba1edd base::MessageLoop::DoWork()
#16 0x7f00bfbb3e0c base::MessagePumpLibevent::Run()
#17 0x7f00bfba14f3 base::MessageLoop::Run()
#18 0x7f00bfc3f827 base::RunLoop::Run()
#19 0x7f00bfce4692 base::Thread::Run()
#20 0x7f00b70f9158 content::BrowserThreadImpl::IOThreadRun()
#21 0x7f00b70f93f1 content::BrowserThreadImpl::Run()
#22 0x7f00bfce4e83 base::Thread::ThreadMain()
#23 0x7f00bfccd00a base::(anonymous namespace)::ThreadFunc()
#24 0x7f00bff6b184 start_thread
#25 0x7f00a5035bed clone
r8: 0000000000000000 r9: 00007f00954fd1c0 r10: fffffffffffffeb8 r11: 00007f00a50c1cf0
r12: 0000000000000000 r13: 0000000000000000 r14: 00007f00955019c0 r15: 00007f0095501700
di: 00007f00954fddb8 si: 0000000000000000 bp: 00007f00954fe1e0 bx: 0000000000000000
dx: 00007f00954fdc48 ax: 0000000000000000 cx: 00007f00afd63158 sp: 00007f00954fdce0
ip: 00007f00b75021be efl: 0000000000010246 cgf: 0000000000000033 erf: 0000000000000004
trp: 000000000000000e msk: 0000000000000000 cr2: 0000000000000000
[end of stack trace]
,
May 25 2017
I would not expect this configuration to work?
,
May 29 2017
That's what I wasn't too sure about. The test framework is setup in a way to try to create an in-process service manager, and services. I wasn't sure if the various services were ready yet. Do we plan for them to be? Or should we remove the test support for single process?
,
May 29 2017
per conversation in hall: given that we want browser tests running on ozone/drm and support for in-process gpu in ozone/drm is unlikely, I think it's reasonable to remove single process test support.
,
May 29 2017
I will follow up to remove the single process test setup. This issue is therefore not needed
,
Feb 26 2018
|
|||
►
Sign in to add a comment |
|||
Comment 1 by sadrul@chromium.org
, May 25 2017