https://bugs.freedesktop.org/show_bug.cgi?id=111784
Bug ID: 111784 Summary: Hang when using glWaitSync with multithreaded shared GL contexts Product: DRI Version: XOrg git Hardware: x86-64 (AMD64) OS: Linux (All) Status: NEW Severity: normal Priority: not set Component: DRM/AMDgpu Assignee: dri-devel@lists.freedesktop.org Reporter: emmanueldurand@protonmail.com
Created attachment 145472 --> https://bugs.freedesktop.org/attachment.cgi?id=145472&action=edit Output of dmesg
I develop a tool which uses a separate thread for uploading textures to the GPU, in parallel to the rendering thread. These two threads are synchronized using OpenGL fences, which prevents the rendering to happen while a texture is being copied from a PBO.
On recent AMD hardware (tested on a Vega 56 and a Radeon VII) this setup hangs almost instantaneously. From my tests it seems that it waits for a glWaitSync to finish. The exact same code runs flawlessly on Intel (Mesa driver) and Nvidia (proprietary driver).
I managed to somewhat reproduce the issue in a simpler code, which merely creates two shared OpenGL contexts and does nothing except creating fences and waiting for the other thread. This example hangs with AMDGPU driver, but once again runs fine on Intel (Mesa driver) and Nvidia (proprietary driver).
I'll attach the code to this thread, and it can be found here too: https://gitlab.com/sat-metalab/splash/blob/fix/radeon_test/tests/sandbox/rad....
https://bugs.freedesktop.org/show_bug.cgi?id=111784
--- Comment #1 from Emmanuel Durand emmanueldurand@protonmail.com --- Created attachment 145473 --> https://bugs.freedesktop.org/attachment.cgi?id=145473&action=edit Xorg log
https://bugs.freedesktop.org/show_bug.cgi?id=111784
--- Comment #2 from Emmanuel Durand emmanueldurand@protonmail.com --- Created attachment 145474 --> https://bugs.freedesktop.org/attachment.cgi?id=145474&action=edit Source code exhibiting the issue
https://bugs.freedesktop.org/show_bug.cgi?id=111784
Emmanuel Durand emmanueldurand@protonmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Priority|not set |high
https://bugs.freedesktop.org/show_bug.cgi?id=111784
Emmanuel Durand emmanueldurand@protonmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Version|XOrg git |DRI git
https://bugs.freedesktop.org/show_bug.cgi?id=111784
--- Comment #3 from Pierre-Eric Pelloux-Prayer pierre-eric.pelloux-prayer@amd.com --- Created attachment 145487 --> https://bugs.freedesktop.org/attachment.cgi?id=145487&action=edit output from gdb
Using the env var "GALLIUM_THREAD=0" makes the issue worse (the example hangs at the first iteration).
One app thread is stuck at: glWaitSync(_textureUploadFence, 0, GL_TIMEOUT_IGNORED);
The other thread is stuck waiting for the first thread to release the mutex. Before waiting for the mutex it made a call to: "_textureUploadFence = glFenceSync(GL_SYNC_GPU_COMMANDS_COMPLETE, 0);"
All the mesa internal threads are waiting for work to do.
https://bugs.freedesktop.org/show_bug.cgi?id=111784
Alex Deucher alexdeucher@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Version|DRI git |git Component|DRM/AMDgpu |Drivers/Gallium/radeonsi Product|DRI |Mesa QA Contact| |dri-devel@lists.freedesktop | |.org
https://bugs.freedesktop.org/show_bug.cgi?id=111784
GitLab Migration User gitlab-migration@fdo.invalid changed:
What |Removed |Added ---------------------------------------------------------------------------- Resolution|--- |MOVED Status|NEW |RESOLVED
--- Comment #4 from GitLab Migration User gitlab-migration@fdo.invalid --- -- GitLab Migration Automatic Message --
This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.
You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/mesa/mesa/issues/1430.
dri-devel@lists.freedesktop.org