Crash in BrowserChildProcessHostImpl::OnChildDisconnected

[smayoral]

Hi,
I am using cef_binary_94.4.10+g38a7995+chromium-94.0.4606.81_windows32 in Windows 10 and I have following sporadic crash which I am not able to reproduce and also no idea how it happens, what the user does.
Someone has an idea?
Thanks a lot,
Sergio

libcef.dll!logging::LogMessage::~LogMessage() Zeile 889 C++
libcef.dll!content::`anonymous namespace'::IntentionallyCrashBrowserForUnusableGpuProcess() Zeile 418 C++
libcef.dll!content::GpuDataManagerImplPrivate::FallBackToNextGpuMode() Zeile 1717 C++
libcef.dll!content::GpuDataManagerImpl::FallBackToNextGpuMode() Zeile 366 C++
libcef.dll!content::GpuProcessHost::RecordProcessCrash() Zeile 1291 C++
libcef.dll!content::GpuProcessHost::OnProcessCrashed(int exit_code) Zeile 959 C++
> libcef.dll!content::BrowserChildProcessHostImpl::OnChildDisconnected() Zeile 516 C++
libcef.dll!IPC::ChannelMojo::OnPipeError() Zeile 244 C++
libcef.dll!IPC::internal::MessagePipeReader::OnPipeError(unsigned int error) Zeile 179 C++
[Inlineframe] libcef.dll!base::internal::FunctorTraits<void (gpu::raster::RasterImplementation::*)(gpu::raster::RasterImplementation::AsyncYUVReadbackRequest *) __attribute__((thiscall)),void>::Invoke(void(gpu::raster::RasterImplementation::*)(gpu::raster::RasterImplementation::AsyncYUVReadbackRequest *) method, gpu::raster::RasterImplementation * && receiver_ptr, gpu::raster::RasterImplementation::AsyncYUVReadbackRequest * && args) Zeile 509 C++
[Inlineframe] libcef.dll!base::internal::InvokeHelper<0,void>::MakeItSo(void(gpu::raster::RasterImplementation::*)(gpu::raster::RasterImplementation::AsyncYUVReadbackRequest *) && functor, gpu::raster::RasterImplementation * && args, gpu::raster::RasterImplementation::AsyncYUVReadbackRequest * && args) Zeile 648 C++
[Inlineframe] libcef.dll!base::internal::Invoker<base::internal::BindState<void (gpu::raster::RasterImplementation::*)(gpu::raster::RasterImplementation::AsyncYUVReadbackRequest *) __attribute__((thiscall)),base::internal::UnretainedWrapper<gpu::raster::RasterImplementation>,gpu::raster::RasterImplementation::AsyncYUVReadbackRequest *>,void ()>::RunImpl(void(gpu::raster::RasterImplementation::*)(gpu::raster::RasterImplementation::AsyncYUVReadbackRequest *) && functor, std::__1::tuple<base::internal::UnretainedWrapper<gpu::raster::RasterImplementation>,gpu::raster::RasterImplementation::AsyncYUVReadbackRequest *> && bound, std::__1::integer_sequence<unsigned int,0,1>) Zeile 721 C++
libcef.dll!base::internal::Invoker<base::internal::BindState<void (gpu::raster::RasterImplementation::*)(gpu::raster::RasterImplementation::AsyncYUVReadbackRequest *) __attribute__((thiscall)),base::internal::UnretainedWrapper<gpu::raster::RasterImplementation>,gpu::raster::RasterImplementation::AsyncYUVReadbackRequest *>,void ()>::RunOnce(base::internal::BindStateBase * base) Zeile 694 C++
[Inlineframe] libcef.dll!base::OnceCallback<void ()>::Run() Zeile 98 C++
libcef.dll!mojo::InterfaceEndpointClient::NotifyError(const absl::optional<mojo::DisconnectReason> & reason) Zeile 685 C++
libcef.dll!IPC::`anonymous namespace'::ChannelAssociatedGroupController::NotifyEndpointOfError(IPC::`anonymous namespace'::ChannelAssociatedGroupController::Endpoint * endpoint, bool force_async) Zeile 804 C++
libcef.dll!IPC::`anonymous namespace'::ChannelAssociatedGroupController::OnPipeError() Zeile 790 C++
[Inlineframe] libcef.dll!base::OnceCallback<void ()>::Run() Zeile 98 C++
libcef.dll!mojo::Connector::HandleError(bool force_pipe_reset, bool force_async_handler) Zeile 674 C++
libcef.dll!mojo::Connector::OnHandleReadyInternal(unsigned int result) Zeile 441 C++
[Inlineframe] libcef.dll!base::internal::FunctorTraits<void (network::mojom::CookieManager_GetAllCookies_ProxyToResponder::*)(const std::__1::vector<net::CanonicalCookie,std::__1::allocator<net::CanonicalCookie>> &) __attribute__((thiscall)),void>::Invoke(void(network::mojom::CookieManager_GetAllCookies_ProxyToResponder::*)(const std::__1::vector<net::CanonicalCookie,std::__1::allocator<net::CanonicalCookie>> &) method, std::__1::unique_ptr<network::mojom::CookieManager_GetAllCookies_ProxyToResponder,std::__1::default_delete<network::mojom::CookieManager_GetAllCookies_ProxyToResponder>> && receiver_ptr, const std::__1::vector<net::CanonicalCookie,std::__1::allocator<net::CanonicalCookie>> & args) Zeile 509 C++
[Inlineframe] libcef.dll!base::internal::InvokeHelper<0,void>::MakeItSo(void(network::mojom::CookieManager_GetAllCookies_ProxyToResponder::*)(const std::__1::vector<net::CanonicalCookie,std::__1::allocator<net::CanonicalCookie>> &) && functor, std::__1::unique_ptr<network::mojom::CookieManager_GetAllCookies_ProxyToResponder,std::__1::default_delete<network::mojom::CookieManager_GetAllCookies_ProxyToResponder>> && args, const std::__1::vector<net::CanonicalCookie,std::__1::allocator<net::CanonicalCookie>> & args) Zeile 648 C++
[Inlineframe] libcef.dll!base::internal::Invoker<base::internal::BindState<void (network::mojom::CookieManager_GetAllCookies_ProxyToResponder::*)(const std::__1::vector<net::CanonicalCookie,std::__1::allocator<net::CanonicalCookie>> &) __attribute__((thiscall)),std::__1::unique_ptr<network::mojom::CookieManager_GetAllCookies_ProxyToResponder,std::__1::default_delete<network::mojom::CookieManager_GetAllCookies_ProxyToResponder>>>,void (const std::__1::vector<net::CanonicalCookie,std::__1::allocator<net::CanonicalCookie>> &)>::RunImpl(void(network::mojom::CookieManager_GetAllCookies_ProxyToResponder::*)(const std::__1::vector<net::CanonicalCookie,std::__1::allocator<net::CanonicalCookie>> &) && functor, std::__1::tuple<std::__1::unique_ptr<network::mojom::CookieManager_GetAllCookies_ProxyToResponder,std::__1::default_delete<network::mojom::CookieManager_GetAllCookies_ProxyToResponder>>> && bound, std::__1::integer_sequence<unsigned int,0>, const std::__1::vector<net::CanonicalCookie,std::__1::allocator<net::CanonicalCookie>> & unbound_args) Zeile 721 C++
libcef.dll!base::internal::Invoker<base::internal::BindState<void (network::mojom::CookieManager_GetAllCookies_ProxyToResponder::*)(const std::__1::vector<net::CanonicalCookie,std::__1::allocator<net::CanonicalCookie>> &) __attribute__((thiscall)),std::__1::unique_ptr<network::mojom::CookieManager_GetAllCookies_ProxyToResponder,std::__1::default_delete<network::mojom::CookieManager_GetAllCookies_ProxyToResponder>>>,void (const std::__1::vector<net::CanonicalCookie,std::__1::allocator<net::CanonicalCookie>> &)>::RunOnce(base::internal::BindStateBase * base, const std::__1::vector<net::CanonicalCookie,std::__1::allocator<net::CanonicalCookie>> & unbound_args) Zeile 694 C++
[Inlineframe] libcef.dll!base::RepeatingCallback<void (unsigned int)>::Run(unsigned int args) Zeile 166 C++
libcef.dll!mojo::SimpleWatcher::DiscardReadyState(const base::RepeatingCallback<void (unsigned int)> & callback, unsigned int result, const mojo::HandleSignalsState & state) Zeile 189 C++
[Inlineframe] libcef.dll!base::internal::FunctorTraits<void (*)(const base::RepeatingCallback<void (unsigned int)> &, unsigned int, const mojo::HandleSignalsState &),void>::Invoke(void(*)(const base::RepeatingCallback<void (unsigned int)> &, unsigned int, const mojo::HandleSignalsState &) & function, const base::RepeatingCallback<void (unsigned int)> & args, unsigned int && args, const mojo::HandleSignalsState & args) Zeile 404 C++
[Inlineframe] libcef.dll!base::internal::InvokeHelper<0,void>::MakeItSo(void(*)(const base::RepeatingCallback<void (unsigned int)> &, unsigned int, const mojo::HandleSignalsState &) & functor, const base::RepeatingCallback<void (unsigned int)> & args, unsigned int && args, const mojo::HandleSignalsState & args) Zeile 648 C++
[Inlineframe] libcef.dll!base::internal::Invoker<base::internal::BindState<void (*)(const base::RepeatingCallback<void (unsigned int)> &, unsigned int, const mojo::HandleSignalsState &),base::RepeatingCallback<void (unsigned int)>>,void (unsigned int, const mojo::HandleSignalsState &)>::RunImpl(void(*)(const base::RepeatingCallback<void (unsigned int)> &, unsigned int, const mojo::HandleSignalsState &) & functor, const std::__1::tuple<base::RepeatingCallback<void (unsigned int)>> & bound, std::__1::integer_sequence<unsigned int,0>, unsigned int && unbound_args, const mojo::HandleSignalsState & unbound_args) Zeile 721 C++
libcef.dll!base::internal::Invoker<base::internal::BindState<void (*)(const base::RepeatingCallback<void (unsigned int)> &, unsigned int, const mojo::HandleSignalsState &),base::RepeatingCallback<void (unsigned int)>>,void (unsigned int, const mojo::HandleSignalsState &)>::Run(base::internal::BindStateBase * base, unsigned int unbound_args, const mojo::HandleSignalsState & unbound_args) Zeile 703 C++
[Inlineframe] libcef.dll!base::RepeatingCallback<void (unsigned int, const mojo::HandleSignalsState &)>::Run(unsigned int args, const mojo::HandleSignalsState & args) Zeile 166 C++
libcef.dll!mojo::SimpleWatcher::OnHandleReady(int watch_id, unsigned int result, const mojo::HandleSignalsState & state) Zeile 278 C++
libcef.dll!mojo::SimpleWatcher::Context::Notify(unsigned int result, MojoHandleSignalsState signals_state, unsigned int flags) Zeile 94 C++
libcef.dll!mojo::SimpleWatcher::Context::CallNotify(const MojoTrapEvent * event) Zeile 63 C++
libcef.dll!mojo::core::WatcherDispatcher::InvokeWatchCallback(unsigned int context, unsigned int result, const mojo::core::HandleSignalsState & state, unsigned int flags) Zeile 94 C++
libcef.dll!mojo::core::Watch::InvokeCallback(unsigned int result, const mojo::core::HandleSignalsState & state, unsigned int flags) Zeile 79 C++
libcef.dll!mojo::core::RequestContext::~RequestContext() Zeile 73 C++
libcef.dll!mojo::core::NodeChannel::OnChannelError(mojo::core::Channel::Error error) Zeile 853 C++
libcef.dll!mojo::core::`anonymous namespace'::ChannelWin::OnIOCompleted(base::MessagePumpForIO::IOContext * context, unsigned long bytes_transfered, unsigned long error) Zeile 262 C++
libcef.dll!base::MessagePumpForIO::WaitForIOCompletion(unsigned long timeout) Zeile 797 C++
libcef.dll!base::MessagePumpForIO::WaitForWork(base::MessagePump::Delegate::NextWorkInfo next_work_info) Zeile 770 C++
libcef.dll!base::MessagePumpForIO::DoRunLoop() Zeile 752 C++
libcef.dll!base::MessagePumpWin::Run(base::MessagePump::Delegate * delegate) Zeile 80 C++
libcef.dll!base::sequence_manager::internal::ThreadControllerWithMessagePumpImpl::Run(bool application_tasks_allowed, base::TimeDelta timeout) Zeile 470 C++
libcef.dll!base::RunLoop::Run(const base::Location & location) Zeile 136 C++
libcef.dll!base::Thread::Run(base::RunLoop * run_loop) Zeile 341 C++
libcef.dll!content::BrowserProcessIOThread::IOThreadRun(base::RunLoop * run_loop) Zeile 128 C++
libcef.dll!base::Thread::ThreadMain() Zeile 415 C++
libcef.dll!base::`anonymous namespace'::ThreadFunc(void * params) Zeile 113 C++

[amaitland]

Are you distributing SwANGLE and swiftshader dlls?

What GPU does the machine have?

IntentionallyCrashBrowserForUnusableGpuProcess is crashing intentionally.

One possibility is https://www.magpcss.org/ceforum/viewtopic.php?t=18524

[smayoral]

Yes, I am distributing all the libraries.
The application does not crash on startup, just sporadically and quite often during sleep mode or after coming back from this state.
Sadly unable to reproduce it...

[magreenblatt]

There are a number of known/similar crashes in Chromium currently. Sadly I can't share links (they're view restricted), but the Chromium team is aware of them and may fix them in an unspecified upcoming version. If the problem is significant for you I suggest sticking with M93 for the time being (they may also occur with M93 and older, but the problem seems to have become more common starting with M94).

[amaitland]

Does the problem occur with a specific GPU? You can load chrome://gpu in a CefBrowser instance to get a report of the GPU.

If it's limited to a specific GPU then perhaps there might be some options worth testing --disable-gpu etc

[smayoral]

A user reported today again the same crash. Log showed the following:

[1214/175350.787:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175350.953:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.031:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.081:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.120:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.185:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.228:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.283:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.327:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.373:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=1073807364
[1214/175351.373:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 1 time(s)
[1214/175351.379:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.443:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.451:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.451:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 2 time(s)
[1214/175351.495:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.506:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.506:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 3 time(s)
[1214/175351.547:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.562:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.562:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 4 time(s)
[1214/175351.608:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.625:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.625:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 5 time(s)
[1214/175351.662:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.693:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.693:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 6 time(s)
[1214/175351.728:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.755:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.755:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 7 time(s)
[1214/175351.777:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.802:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.802:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 8 time(s)
[1214/175351.864:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.875:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.875:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 9 time(s)
[1214/175351.875:FATAL:gpu_data_manager_impl_private.cc(417)] GPU process isn't usable. Goodbye.

Looking at the stack trace I found following code:

bool disable_crash_limit = base::CommandLine::ForCurrentProcess()->HasSwitch(
switches::kDisableGpuProcessCrashLimit);

This switch sounds promising and hopefully my application does not crash any more.

[ndesktop]

Try to check if --disable-gpu (mostly --disable-gpu --disable-gpu-compositing and --disable-gpu-rasterization) solves or at least improves.
If so, maybe you can expose in your configuration something like "Disable acceleration" and avoid the crashes.

[lwttai]

A user reported today again the same crash. Log showed the following:

[1214/175350.787:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175350.953:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.031:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.081:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.120:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.185:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.228:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.283:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.327:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.373:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=1073807364
[1214/175351.373:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 1 time(s)
[1214/175351.379:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.443:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.451:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.451:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 2 time(s)
[1214/175351.495:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.506:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.506:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 3 time(s)
[1214/175351.547:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.562:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.562:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 4 time(s)
[1214/175351.608:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.625:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.625:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 5 time(s)
[1214/175351.662:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.693:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.693:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 6 time(s)
[1214/175351.728:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.755:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.755:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 7 time(s)
[1214/175351.777:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.802:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.802:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 8 time(s)
[1214/175351.864:ERROR:network_service_instance_impl.cc(333)] Network service crashed, restarting service.
[1214/175351.875:ERROR:gpu_process_host.cc(956)] GPU process exited unexpectedly: exit_code=-1073741205
[1214/175351.875:WARNING:gpu_process_host.cc(1269)] The GPU process has crashed 9 time(s)
[1214/175351.875:FATAL:gpu_data_manager_impl_private.cc(417)] GPU process isn't usable. Goodbye.

Looking at the stack trace I found following code:

bool disable_crash_limit = base::CommandLine::ForCurrentProcess()->HasSwitch(
switches::kDisableGpuProcessCrashLimit);

This switch sounds promising and hopefully my application does not crash any more.

I have been found same situation with same gpu-process exit codes: First 1073807364(STATUS_CONTROL_C_EXIT) then several -1073741205(STATUS_DLL_INIT_FAILED_LOGOFF)
It's easy to reproduce on some machine by shutdown or restart the Windows while the browser app still running. Not every machine, still not figure out what's the different.
In my opinion, maybe chromium should make an exception for this situation for GpuProcessCrashLimit as default enable?

[tapineb371]

try the in-process-gpu switch