Bug 80066 - With kernel 3.14, terrible radeonsi performance with a RadeonHD 7770 resuming from suspend/sleep - "*ERROR* UVD not responding, trying to reset the VCPU!!!"
Summary: With kernel 3.14, terrible radeonsi performance with a RadeonHD 7770 resuming...
Status: NEW
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/Radeon (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: medium major
Assignee: xf86-video-ati maintainers
QA Contact: Xorg Project Team
URL: https://bugzilla.redhat.com/show_bug....
Depends on:
Reported: 2014-06-15 20:41 UTC by Jean-François Fortin Tam
Modified: 2014-06-16 14:12 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Note You need to log in before you can comment on or make changes to this bug.
Description Jean-François Fortin Tam 2014-06-15 20:41:13 UTC
Initially filed downstream on Fedora 20: https://bugzilla.redhat.com/show_bug.cgi?id=1097045

See that bug report for details, systemd journal logs, etc.
journalctl output: https://bugzilla.redhat.com/attachment.cgi?id=894972

Since kernel 3.14 on Fedora 20, I can't suspend my desktop computer anymore. The suspend/resume operation "seems to work" generally speaking, but it borks the state of the graphics driver. This used to work fine in kernel 3.12 (and I think 3.13 too).

Symptoms: extremely sluggish operations for anything that involves 3D (ex: gnome shell animations). This persists until a complete reboot (reloading gnome shell, or using "systemctl isolate multi-user.target && systemctl isolate graphical.target" doesn't solve it).

The journalctl output seems to indicate that the radeon driver is freaking out about UVD, among other things:

> [drm:uvd_v1_0_start] *ERROR* UVD not responding, trying to reset the VCPU!!!
> [drm:uvd_v1_0_start] *ERROR* UVD not responding, giving up!!!
> [drm:si_startup] *ERROR* radeon: failed initializing UVD (-1).
Comment 1 Alex Deucher 2014-06-16 14:12:56 UTC
Does it work properly with 3.15?  There were some suspend and resume fixes in 3.15 that should make their way back to stable kernels.  If not, can you bisect?

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.