The parallel-rdp-standalone's discuss from themaister

parallel-rdp-standalone's Issues

Failure to shutdown gracefully on devices that don't have required vulkan extensions

There seems to be a potential race condition during shutdown on devices that don't support all the correct vulkan extensions, like some Android devices.

Mainly I'm getting this stack trace sometimes:

Crashed: Thread: SIGABRT  0x0000000000000000
#00 pc 0x2216c libc.so 
#01 pc 0x22140 libc.so 
#02 pc 0x919e8 libc.so 
#03 pc 0x91050 libc.so 
#04 pc 0x90ecc libc.so 
#05 pc 0x4546d4 libmupen64plus-video-parallel.so (<name omitted> [__threading_support:367])
#06 pc 0x4069b4 libmupen64plus-video-parallel.so (RDP::Renderer::maintain_queues_idle() [__mutex_base:91])
#07 pc 0x3fdabc libmupen64plus-video-parallel.so (RDP::CommandProcessor::enqueue_command_direct(unsigned int, unsigned int const*) [rdp_device.cpp:939])
#08 pc 0x3fa5dc libmupen64plus-video-parallel.so (RDP::CommandRing::thread_loop() [command_ring.cpp:126])
#09 pc 0x3fad84 libmupen64plus-video-parallel.so (void* std::__ndk1::__thread_proxy<std::__ndk1::tuple<std::__ndk1::unique_ptr<std::__ndk1::__thread_struct, std::__ndk1::default_delete<std::__ndk1::__thread_struct> >, void (RDP::CommandRing::*)(), RDP::CommandRing*> >(void*) [thread:273])
#10 pc 0x90328 libc.so 
#11 pc 0x23a28 libc.so

I would take the line numbers with a grain of salt since these are stack traces coming out of optimized builds. Some of the stack is also missing due to the build being optimized. Anyways, an invalid mutex seems to be getting locked during shutdown. It almost seems like the thread is still running after the Renderer object is destroyed.

Repo not proper cleaning old files (wsi_timing)?

@Themaister I noticed that vulkan/wsi_timing.cpp is still here, even though it was removed from Granite. I'm not sure if there are other cases of leftover files.

Recommend Projects