Bug 102414 - Hawaii gpu (R9 290) and the troubles with the amd-staging-drm-next branch
Summary: Hawaii gpu (R9 290) and the troubles with the amd-staging-drm-next branch
Status: RESOLVED MOVED
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/Gallium/radeonsi (show other bugs)
Version: unspecified
Hardware: Other All
: medium normal
Assignee: Default DRI bug account
QA Contact: Default DRI bug account
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2017-08-25 18:16 UTC by Vadim Girlin
Modified: 2019-09-25 17:59 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
dmesg (250.29 KB, text/plain)
2017-08-25 18:16 UTC, Vadim Girlin
Details
lspci (18.00 KB, text/plain)
2017-08-25 18:18 UTC, Vadim Girlin
Details
Xorg.0.log (52.04 KB, text/x-log)
2017-08-25 18:34 UTC, Vadim Girlin
Details
dmesg right after boot (250.84 KB, text/plain)
2017-08-25 18:41 UTC, Vadim Girlin
Details
Patch for suspend/resume issues (1.99 KB, patch)
2017-08-25 18:49 UTC, Harry Wentland
Details | Splinter Review
dmesg after failed resume (106.20 KB, text/plain)
2017-09-11 06:39 UTC, Vadim Girlin
Details

Description Vadim Girlin 2017-08-25 18:16:35 UTC
Created attachment 133782 [details]
dmesg

Current amd-staging-drm-next branch has some issues with hawaii/CI R9 290 gpu, there is a lot of backtrace noise in the dmesg (see attachment) and some related messages during the boot. I'll add a dmesg right after the boot as well.

There are also some issues with suspend/resume which are possibly related, though not sure.

The problem appeared since the beginning of the August, IIRC. Everything was more stable previously, but it's hard to double-check older versions as long as the branch is usually force-pushed.
Comment 1 Vadim Girlin 2017-08-25 18:18:41 UTC
Created attachment 133783 [details]
lspci
Comment 2 Vadim Girlin 2017-08-25 18:34:21 UTC
Created attachment 133784 [details]
Xorg.0.log
Comment 3 Vadim Girlin 2017-08-25 18:41:20 UTC
Created attachment 133785 [details]
dmesg right after boot

There is a lot of noise, so initial stuff is missing anyway.
Please let me know if you need it, I can probably adjust kernel options to save the entire log.
Comment 4 Harry Wentland 2017-08-25 18:49:20 UTC
Created attachment 133786 [details] [review]
Patch for suspend/resume issues

Hopefully this helps with the suspend/resume issues.
Comment 5 Vadim Girlin 2017-08-26 14:52:12 UTC
(In reply to Harry Wentland from comment #4)
> Created attachment 133786 [details] [review] [review]
> Patch for suspend/resume issues
> 
> Hopefully this helps with the suspend/resume issues.

Unfortunately the problem with suspend is still there, basically the system fails to suspend and then any attempt to suspend/reboot/shutdown results in "There's already a shutdown or sleep operation in progress" message, so I have to use magic sysrq keys to reboot.

I'm still not sure if it's directly related to the amdgpu driver issues, but IIRC the problem has appeared at the same time as the backtraces noise in the logs, and I didn't hit it with other kernels, only with the amd-staging-drm-next branch.

It was reproducible for me on Fedora 26 and currently it's the same with the clean installation of Fedora 27 (pre-release).
Comment 6 Vadim Girlin 2017-09-11 06:39:50 UTC
Created attachment 134152 [details]
dmesg after failed resume

The latest amd-staging-drm-next branch fails to resume from suspend properly for me. It worked previously (when it was possible to perform a suspend).

After resume the screen is mostly blue now with some dynamic random noise - horizontal black lines etc. Though I hear the sound via the display port audio from the video that was playing at the moment of suspend, so at least that part seems to work.

There are some amdgpu backtraces in the attached dmesg before and after the suspend/resume, but I didn't notice any real issues before suspend so far, the only problem is that the display is unusable after resume.
Comment 7 Timothy Arceri 2018-04-10 12:25:47 UTC
Is this still a problem?
Comment 8 GitLab Migration User 2019-09-25 17:59:57 UTC
-- 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/1277.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.