Closed dmik closed 3 years ago
Fixing https://github.com/bitwiseworks/libc/issues/111 gives us a nice .TRP when this happens:
033BEDB4 1E19F048 LIBCX0 0001:0000F048 interrupt.c#618 _interrupt_request_release + 88 0001:0000EFC0 (D:\Coding\libcx\master\src\libcx\interrupt.c)
Offset Name Type Hex Value
ÄÄÄÄÄÄ ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ ÄÄÄÄÄÄÄÄÄ
8 result pointer to type 0x24F BEFE5540
033BEDE4 1E19E499 LIBCX0 0001:0000E499 handles.c#598 _libcx_take_handles + 1E9 0001:0000E2B0 (D:\Coding\libcx\master\src\libcx\handles.c)
Offset Name Type Hex Value
ÄÄÄÄÄÄ ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ ÄÄÄÄÄÄÄÄÄ
8 handles pointer to type 0x249 204DB488
12 num_handles 32 bit unsigned 1
16 pid 32 bit signed 360
20 flags 32 bit signed 1
-20 result 0x263 0
-16 fd 32 bit signed 0
-12 pFH pointer to type 0x241 20842D18
-16 fd 32 bit signed 0
-12 pFH pointer to type 0x241 20842D18
-16 fd 32 bit signed 0
-12 pFH pointer to type 0x241 20842D18
-16 fd 32 bit signed 0
-12 pFH pointer to type 0x241 20842D18
-20 result 0x263 0
-20 result 0x263 0
-20 result 0x263 0
-20 result 0x263 0
033BEE44 B3B770DC QT5WEBCD 0001:028770DC std::vector<mojo::PlatformHandle, std::allocator<mojo::PlatformHandle> >::_M_default_append(unsigned int) + 42C 0001:02876CB0 (channel_posix.o)
033BEFA4 B3B2064E QT5WEBCD 0001:0282064E mojo::core::Channel::TryDispatchMessage(base::span<char const, 4294967295u>, unsigned int*) + 13E 0001:02820510 (channel.o)
033BF104 B3B213D4 QT5WEBCD 0001:028213D4 mojo::core::Channel::OnReadComplete(unsigned int, unsigned int*) + E4 0001:028212F0 (channel.o)
033BF244 B3B7AFB8 QT5WEBCD 0001:0287AFB8 __ZN7logging17MakeCheckOpStringIyyEEPNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERKT_RKT0_PKc$w$fOFZKZc9jW7bHItF1 + 3E18 0001:028771A0 (channel_posix.o)
033BF3A4 B838B9D5 QT5WEBCD 0001:0708B9D5 base::MessagePumpLibevent::OnLibeventNotification(int, short, void*) + E5 0001:0708B8F0 (base_jumbo_42.o)
033BF524 11B5548B EVNT217 0001:0001548B
033BF594 11B55D5E EVNT217 0001:00015D5E
033BF604 B838DA10 QT5WEBCD 0001:0708DA10 base::MessagePumpLibevent::Run(base::MessagePump::Delegate*) + 90 0001:0708D980 (base_jumbo_42.o)
033BF764 B87EDBD3 QT5WEBCD 0001:074EDBD3 base::sequence_manager::internal::ThreadControllerWithMessagePumpImpl::Run(bool, base::TimeDelta) + 153 0001:074EDA80 (base_jumbo_23.o)
033BF894 B78809E6 QT5WEBCD 0001:065809E6 base::RunLoop::AfterRun() + 1066 0001:0657F980 (base_jumbo_17.o)
033BFA04 B788403A QT5WEBCD 0001:0658403A base::RunLoop::Run() + 3A 0001:06584000 (base_jumbo_17.o)
033BFB14 B7734535 QT5WEBCD 0001:06434535 base::Thread::Run(base::RunLoop*) + A5 0001:06434490 (base_jumbo_28.o)
033BFC44 BE4E078C QT5WEBCD 0001:0D1E078C content::BrowserProcessSubThread::IOThreadRun(base::RunLoop*) + 3C 0001:0D1E0750 (browser_process_sub_thread.o)
033BFC74 BE4E08CC QT5WEBCD 0001:0D1E08CC content::BrowserProcessSubThread::Run(base::RunLoop*) + CC 0001:0D1E0800 (browser_process_sub_thread.o)
033BFD94 B774742E QT5WEBCD 0001:0644742E base::Thread::ThreadMain() + 4EE 0001:06446F40 (base_jumbo_28.o)
033BFF04 B76D797D QT5WEBCD 0001:063D797D base::HighResolutionTimerManager::~HighResolutionTimerManager() + 52D 0001:063D7450 (base_jumbo_39.o)
033BFF34 1DDE11E2 PTHR01 0001:000011E2 my_os2thread.c#111 _pthread_exit + A2 0001:00001140 (D:\Users\dmik\rpmbuild\BUILD\pthread-os2-0.2.4\src\my_os2thread.c)
033BFF64 1E1980D2 LIBCX0 0001:000080D2 beginthread.c#99 __beginthread - 1E 0001:000080F0 (D:\Coding\libcx\master\src\beginthread.c)
033BFFB4 1E07813E LIBCN0 0001:0005813E beginthr.c#38 ___libc_back_threadCreate - 6A 0001:000581A8 (beginthr.obj)
033BFFF4 1FFECE38 DOSCALL1 0004:0000CE38 DOS32R3EXITADDR
Note that I only see it if --enable-features=NetworkServiceInProcess
is in QTWEBENGINE_CHROMIUM_FLAGS so far.
Sometimes when running the 5.15 simplebrowser app after fixing #41 I get an assertion in LIBCx at some arbitrary points, here is the output:
I wonder why I don't also get a .TRP with the stack trace. This must somehow deal with the fact that EXCEPTQ exception handling is now moved to LIBC...