Bug 87728

Summary: Starting X produces a screen of static and GPU lockups (Radeon HD 7870)
Product: xorg Reporter: John Steele Scott <toojays>
Component: Driver/RadeonAssignee: xf86-video-ati maintainers <xorg-driver-ati>
Status: RESOLVED DUPLICATE QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: medium    
Version: unspecified   
Hardware: Other   
OS: All   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
lspci -v
none
syslog
none
xorg.log
none
syslog using radeonsi_dri.so from 10.3 branch at 5878e18, "atombios stuck" none

Description John Steele Scott 2014-12-26 02:34:54 UTC
I am still unable to get the radeonsi driver working with my Radeon HD 7970. Bug 84186 was my previous attempt.

Right now I have installed:

kernel from Dave Airlie's drm-next, at commit 4e0cd68 (December 13 2014)
libgl1-mesa-glx 10.3.2-1
xserver-xorg-core 2:1.16.2.901-1
xserver-xorg-video-radeon 1:7.5.0-1

Apart from the kernel, the rest are Debian (jessie) packages.

With this configuration, X starts, but the display is a mess of grey static. Through the static I can discern a large square which is apparently a mouse cursor (moves when I move my mouse).

My syslog contains many complaints from the radeon driver which look like:

Dec 26 12:41:57 kano kernel: radeon 0000:01:00.0: ring 0 stalled for more than 10004msec
Dec 26 12:41:57 kano kernel: radeon 0000:01:00.0: GPU lockup (current fence id 0x0000000000000001 last fence id 0x0000000000000051 on ring 0)
Dec 26 12:41:58 kano kernel: [drm:si_gpu_check_soft_reset] MC busy: 0x00000548, clearing.
Comment 1 John Steele Scott 2014-12-26 02:37:03 UTC
Created attachment 111345 [details]
lspci -v
Comment 2 John Steele Scott 2014-12-26 02:38:26 UTC
Oops, it's actually a 7870 (according to the box it came in). Not sure why the audio part in lspci reports as 7970.
Comment 3 John Steele Scott 2014-12-26 02:38:49 UTC
Created attachment 111346 [details]
syslog
Comment 4 John Steele Scott 2014-12-26 02:39:10 UTC
Created attachment 111347 [details]
xorg.log
Comment 5 Michel Dänzer 2014-12-26 03:12:49 UTC
The fix from bug 60879 might help.
Comment 6 John Steele Scott 2014-12-28 00:35:25 UTC
(In reply to Michel Dänzer from comment #5)
> The fix from bug 60879 might help.

Thanks Michel. I tried replacing my radeonsi_dri.so (only) with one built from the 10.3 branch at 5878e18, which includes the fix for bug 60879. The display is still grey static, but now there is no "mouse cursor" kind of thing. Periodically it cycles between static and blank screen. After a couple of minutes of this, the box locks up hard.

The syslog output is a little different as well. Now I see some lines like: 

Dec 28 10:57:00 kano kernel: [drm:atom_op_jump [radeon]] *ERROR* atombios stuck in loop for more than 5secs aborting
Dec 28 10:57:00 kano kernel: [drm:atom_execute_table_locked [radeon]] *ERROR* atombios stuck executing C238 (len 254, WS 0, PS 4) @ 0xC262
Dec 28 10:57:00 kano kernel: [drm:atom_execute_table_locked [radeon]] *ERROR* atombios stuck executing B798 (len 106, WS 0, PS 8) @ 0xB7ED

which were not there before. I'll attach the full syslog shortly.
Comment 7 John Steele Scott 2014-12-28 00:37:22 UTC
Created attachment 111413 [details]
syslog using radeonsi_dri.so from 10.3 branch at 5878e18, "atombios stuck"
Comment 8 Michel Dänzer 2015-11-20 00:51:17 UTC

*** This bug has been marked as a duplicate of bug 60879 ***

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.