Since updating to 6.9.10-arch1-1kernel Firefox has been crashing due to nvidia explicit sync on Wayland.
GraphicsCriticalError: |[0][GFX1-]: Wayland protocol error: wp_linux_drm_syncobj_surface_v1@53: error 4: explicit sync is used, but no acquire point is set
This is part of the error messaging in Firefox when it crashes which is quite long. It was said to change environmental variable from MOZ_ENABLE_WAYLAND=1 to 0 but that doesn’t work and when i boot if i move the cursor on log in it causes EOS to crash.
So i went into Firefox about:config and set nvidia explicit sync to on. Now it seems to be okay but i need to use it more to know for sure. Just wondered if anyone else has run across this yet?
This is the first issue i have had in a while on nvidia with Wayland.
Sorry if this is a dumb question but how or where can i see if firefox is using explicit sync? When checking with nvtop i never see my Nvidia working it’s always my AMD igpu that’s in action. So far the only package i’ve seen using nvidia is mpv player when watching a movie.
Forgot to mention I’m currently on Arch with nvidia-open-dkms drivers and fully updated just now (including firefox 128.0-2)
It’s still crashing frequently here on version 128.0-2 when playing YouTube videos.
I have rebooted the system just in case, but the problem is still here.
[GFX1-]: Detect DeviceReset DeviceResetReason::RESET DeviceResetDetectPlace::WR_POST_UPDATE in Parent process
[GFX1-]: Failed to make render context current during destroying.
[GFX1-]: Wayland protocol error: wp_linux_drm_syncobj_surface_v1#114: error 4: explicit sync is used, but no acquire point is set
ExceptionHandler::GenerateDump attempting to generate:/home/lucas/.mozilla/firefox/tzxbgoxz.default-release/minidumps/2dfd35ca-317a-df2e-d327-9b445144d634.dmp
ExceptionHandler::GenerateDump cloned child 44134
ExceptionHandler::WaitForContinueSignal waiting for continue signal...
ExceptionHandler::SendContinueSignalToChild sent continue signal to child
ExceptionHandler::GenerateDump minidump generation succeeded
pacman -Q | grep -e firefox
firefox 128.0-2
Edit 1: root cause seems to be in Mesa, tracked by this issue:
Edit 2: I’m launching Firefox as an X window until this is resolved:
Of course. I’m not saying it’s totally fixed but since i set nvidia explicit sync true on Firefox it’s not crashing like it was. It never did this before the latest kernel. Firefox was also updated recently.
Edit: So what I’m finding is that it’s not really crashing but it’s when I’m at the login screen for Kde if i move the mouse it jumps to the last entry i think before it loads the login screen which is the tty screen. I have to use REISUB to reboot it. It’s got to be the latest kernel as i didn’t have this issue before. I guess i could try downgrading the kernel.
All seems good in the settings, anyway i was just wondering how it can be checked. As said i never see any action on the nvidia when being in firefox. Next computer will for sure not be nvidia shit anymore even if they decide to fully open source.
Well I’m still having problems with crashing since updating and I know there is a pile of Kde updates in there also so i went back using snapshots and it’s working fine. It could be the kernel or something else? Think I’ll update without the kernel and try.