[SOLVED] mesa 23.2.1-1: google chrome can't initialise GL

martinc2 · 2023-10-04 10:31:38

(resolved with mesa 23.2.1-2)

Hi, google-chrome 117.0.5938.132-1 barfs when trying to initialize GL, resulting in a session with no hw acceleration and no WebGL. Downgrading mesa to 23.1.8-1 resolves the issue.

pastebin of STDERR log is here.

I'm on intel irix, no particular settings that I can think of that would interfere. I'm not sure if this is a chrome or a mesa issue; admins please feel free to move this post to the AUR subforum if more appropriate.

Last edited by martinc2 (2023-10-07 17:27:08)

Lone_Wolf · 2023-10-05 07:48:47

https://aur.archlinux.org/packages/google-chrome was updated to 117.0.5938.149-1 yesterday evening, does that work witrh latest mesa ?

WorMzy · 2023-10-05 08:45:40

Mod note: moving to AUR Issues

martinc2 · 2023-10-05 10:21:28

Lone_Wolf wrote:

https://aur.archlinux.org/packages/google-chrome was updated to 117.0.5938.149-1 yesterday evening, does that work witrh latest mesa ?

Nope, unfortunately the issue persists.

No issues on a Desktop i7-6700 (skylake, HD Graphics 530).

V1del · 2023-10-05 10:43:40

sometimes chrome doesn't properly invalidate it's GL cache on mesa updates and keels over in various ways and forms, try removing ~/.config/google-chrome/ShaderCache (and/or ~/.config/google-chrome/Default/GPUCache ) (I don't know the actual chrome path/this was subsituted from chromium)

FWIW can you reproduce your issues on chromium? If you can, we likely have a mesa issue.

martinc2 · 2023-10-05 12:42:26

For chrome, removing the Cache files makes no difference.

The same issues persist with chromium 117.0.5938.149 (though STDERR is less verbose)

[4569:4569:1005/133723.482086:ERROR:policy_logger.cc(154)] :components/enterprise/browser/controller/chrome_browser_cloud_management_controller.cc(163) Cloud management controller initialization aborted as CBCM is not enabled.
[4569:4569:1005/133725.059720:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=139
[4669:1:1005/133726.046209:ERROR:command_buffer_proxy_impl.cc(129)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
[4569:4569:1005/133726.139413:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=139
[4569:4569:1005/133727.204222:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=139
[4670:7:1005/133727.330560:ERROR:command_buffer_proxy_impl.cc(129)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
[4669:7:1005/133727.331251:ERROR:command_buffer_proxy_impl.cc(129)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.

Lone_Wolf · 2023-10-06 08:25:44

You already mentioned downgrading mesa helped, but this could be something that is exposed by mesa, not caused.

STDERR reports when things have gone wrong, but doesn't help to determine how that happened.
Please post full lspci -k and journal (as root) from the most recent fail.

martinc2 · 2023-10-06 08:53:42

Thanks @Lone_Wolf for your help.

(running chromium):

seth · 2023-10-06 14:20:12

Oct 06 09:41:50 jeff systemd[1]: Created slice Slice /system/systemd-coredump.
Oct 06 09:41:50 jeff systemd[1]: Started Process Core Dump (PID 2492/UID 0).
Oct 06 09:41:51 jeff systemd-coredump[2493]: Process 2378 (chromium) of user 1000 dumped core.
                                             
                                             Stack trace of thread 2378:
                                             #0  0x0000000000000000 n/a (n/a + 0x0)
                                             ELF object binary architecture: AMD x86-64
Oct 06 09:41:51 jeff systemd[1]: systemd-coredump@0-2492-0.service: Deactivated successfully.
Oct 06 09:41:51 jeff krunner[2330]: [2330:2330:1006/094151.267733:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=139
Oct 06 09:41:51 jeff systemd[1]: Started Process Core Dump (PID 2540/UID 0).
Oct 06 09:41:51 jeff systemd[1]: systemd-hostnamed.service: Deactivated successfully.
Oct 06 09:41:52 jeff systemd-coredump[2541]: Process 2510 (chromium) of user 1000 dumped core.
                                             
                                             Stack trace of thread 2510:
                                             #0  0x0000000000000000 n/a (n/a + 0x0)
                                             ELF object binary architecture: AMD x86-64
Oct 06 09:41:52 jeff NetworkManager[989]: <warn>  [1696581712.2260] platform-linux: do-add-ip6-address[3: fe80::c1a7:e56b:209d:2910]: failure 95 (Operation not supported)
Oct 06 09:41:52 jeff systemd[1]: systemd-coredump@1-2540-0.service: Deactivated successfully.
Oct 06 09:41:52 jeff krunner[2330]: [2330:2330:1006/094152.255329:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=139
Oct 06 09:41:52 jeff krunner[2434]: [2434:8:1006/094152.441983:ERROR:command_buffer_proxy_impl.cc(129)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
Oct 06 09:41:52 jeff systemd[1]: Started Process Core Dump (PID 2597/UID 0).
Oct 06 09:41:53 jeff systemd-coredump[2598]: Process 2567 (chromium) of user 1000 dumped core.
                                             
                                             Stack trace of thread 2567:
                                             #0  0x0000000000000000 n/a (n/a + 0x0)
                                             ELF object binary architecture: AMD x86-64
Oct 06 09:41:53 jeff systemd[1]: systemd-coredump@2-2597-0.service: Deactivated successfully.
Oct 06 09:41:53 jeff krunner[2435]: [2435:8:1006/094153.220270:ERROR:command_buffer_proxy_impl.cc(321)] GPU state invalid after WaitForGetOffsetInRange.
Oct 06 09:41:53 jeff krunner[2435]: [2435:8:1006/094153.220978:ERROR:command_buffer_proxy_impl.cc(129)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
Oct 06 09:41:53 jeff krunner[2330]: [2330:2330:1006/094153.221109:ERROR:gpu_process_host.cc(989)] GPU process exited unexpectedly: exit_code=139
Oct 06 09:41:53 jeff krunner[2434]: [2434:8:1006/094153.369319:ERROR:command_buffer_proxy_impl.cc(129)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
Oct 06 09:41:53 jeff krunner[2435]: [2435:8:1006/094153.373584:ERROR:command_buffer_proxy_impl.cc(129)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.

Looks like https://bbs.archlinux.org/viewtopic.php?id=289333 and then there's also https://bbs.archlinux.org/viewtopic.php … 5#p2124845

martinc2 · 2023-10-07 17:26:23

mesa 23.2.1-2 resolves the issue

Arch Linux

#1 2023-10-04 10:31:38

[SOLVED] mesa 23.2.1-1: google chrome can't initialise GL

#2 2023-10-05 07:48:47

Re: [SOLVED] mesa 23.2.1-1: google chrome can't initialise GL

#3 2023-10-05 08:45:40

Re: [SOLVED] mesa 23.2.1-1: google chrome can't initialise GL

#4 2023-10-05 10:21:28

Re: [SOLVED] mesa 23.2.1-1: google chrome can't initialise GL

#5 2023-10-05 10:43:40

Re: [SOLVED] mesa 23.2.1-1: google chrome can't initialise GL

#6 2023-10-05 12:42:26

Re: [SOLVED] mesa 23.2.1-1: google chrome can't initialise GL

#7 2023-10-06 08:25:44

Re: [SOLVED] mesa 23.2.1-1: google chrome can't initialise GL

#8 2023-10-06 08:53:42

Re: [SOLVED] mesa 23.2.1-1: google chrome can't initialise GL

#9 2023-10-06 14:20:12

Re: [SOLVED] mesa 23.2.1-1: google chrome can't initialise GL

#10 2023-10-07 17:26:23

Re: [SOLVED] mesa 23.2.1-1: google chrome can't initialise GL

Board footer