Bug 21400 - GPU locked up on M26 (RV410)
Summary: GPU locked up on M26 (RV410)
Status: RESOLVED FIXED
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/r300 (show other bugs)
Version: unspecified
Hardware: Other All
: medium normal
Assignee: Default DRI bug account
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2009-04-25 03:53 UTC by Tormod Volden
Modified: 2009-11-18 07:30 UTC (History)
2 users (show)

See Also:
i915 platform:
i915 features:


Attachments
gdb full backtrace of Xorg (1.21 KB, text/plain)
2009-04-25 03:53 UTC, Tormod Volden
Details
Xorg.0.log (44.88 KB, text/plain)
2009-04-25 03:57 UTC, Tormod Volden
Details
use 1/12 subpixel mode in the ddx (5.07 KB, patch)
2009-05-03 09:43 UTC, Alex Deucher
Details | Splinter Review

Description Tormod Volden 2009-04-25 03:53:08 UTC
Created attachment 25130 [details]
gdb full backtrace of Xorg

When using googleearth it does not take long to lock up this laptop. Switching between fullscreen and window made triggered this hang. I can ssh in and Xorg is stuck. There is nothing in dmesg. See attached gdb full backtrace. (Some symbols are missing, I will fix this for next time).

Is there anything else I can do at this point? Is there a tool to see what has been sent to the cp?

libdrm2 2.4.5-0ubuntu4
libgl1-mesa-dri 7.4-0ubuntu3
linux-image-2.6.30-020630rc3-generic 2.6.30-020630rc3
xserver-xorg-core 2:1.6.0-0ubuntu14
xserver-xorg-video-ati 1:6.12.2.99+git20090415.5abcea88-0ubuntu0tormod
Comment 1 Tormod Volden 2009-04-25 03:57:22 UTC
Created attachment 25131 [details]
Xorg.0.log
Comment 2 Tormod Volden 2009-05-03 07:39:43 UTC
After the "R3xx-R5xx: don't set TX_OFFSET_RS in RS_INST_COUNT" I have the impression I can keep on longer before it crashes, but I can not be sure.
Comment 3 Alex Deucher 2009-05-03 09:43:21 UTC
Created attachment 25391 [details] [review]
use 1/12 subpixel mode in the ddx

Does this patch help any more?
Comment 4 Tormod Volden 2009-05-03 14:20:41 UTC
Thanks for the patch! Applied on master head. Again, I felt I could keep on longer (placebo?), but eventually it died. This time I saw an omen in form of a cursor corruption (a fuzzy square) a second before it got slow and froze. Then I got blooming and sysrq did not help.
Comment 5 Maciej Cencora 2009-05-19 06:39:21 UTC
I think the problem is radeon ddx and r300 dri driver interaction. I can easily lock up the GPU by opening some 3d app in windowed mode and then doing some stuff that heavily utilizes EXA acceleration while 3d app window is still visible.
Comment 6 Xavier Bestel 2009-05-19 07:37:35 UTC
Could you try the "carousel" screensaver ? It locks the GPU without interaction in a few minutes here.

Thanks,
        Xav
Comment 7 Fabio Pedretti 2009-06-26 04:33:27 UTC
This could be the same issue of bug 21849. Upgrading to 2.6.30 should fix the problem.
Comment 8 Tormod Volden 2009-06-29 01:45:27 UTC
No, these issues are older than 2.6.30 and the alignment patch in that bug report. OTOH it seems like the radeon-rewrite and ttm/dri2 make it much less prone to lock up. I will test some more and close the bug if I can not reproduce any longer.
Comment 9 Fabio Pedretti 2009-11-14 05:13:07 UTC
Is this still an issue?
Comment 10 Tormod Volden 2009-11-18 07:30:49 UTC
No, this problem has gone away, thanks!


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.