Bug 23996 - [r128] DRI causes lock-up
Summary: [r128] DRI causes lock-up
Status: RESOLVED WONTFIX
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/rage128 (show other bugs)
Version: 7.2 (2007.02)
Hardware: x86 (IA32) Linux (All)
: medium major
Assignee: Xorg Project Team
QA Contact: Xorg Project Team
URL:
Whiteboard: 2011BRB_Reviewed
Keywords:
: 5562 (view as bug list)
Depends on:
Blocks:
 
Reported: 2009-09-17 07:36 UTC by Rhys Ickeringill
Modified: 2014-04-09 18:38 UTC (History)
4 users (show)

See Also:
i915 platform:
i915 features:


Attachments
DRM debug output from foobillard crash (47.89 KB, text/plain)
2009-09-17 07:36 UTC, Rhys Ickeringill
no flags Details
Kernel log from about time of foobillard system lockup (1.24 KB, text/plain)
2009-09-17 07:40 UTC, Rhys Ickeringill
no flags Details
Kernel log associated with mplayer+xv+dma X freeze (2.24 KB, text/plain)
2009-09-17 07:44 UTC, Rhys Ickeringill
no flags Details
X config file (3.57 KB, text/plain)
2009-09-17 07:45 UTC, Rhys Ickeringill
no flags Details
Xorg log associated with wine semi-random X freezes (43.47 KB, text/plain)
2009-09-17 08:12 UTC, Rhys Ickeringill
no flags Details

Description Rhys Ickeringill 2009-09-17 07:36:58 UTC
Created attachment 29631 [details]
DRM debug output from foobillard crash

Am seeing X and sometimes entire system freezes when using various DRI  dependent apps including:

1. mplayer 

outputting via Xv [with DMAforXv enabled] results in X freezing for ~30sec, after which display is momentarily updated before feezing again. Keyboard & mouse are locked, but system responds to ACPI events. No feezing occurs if DMAforXv is disabled.


2. wine 

results in semi-random ~30sec X lockups after which X resume normal operations until next semi-random freeze. Strictly may not be a wine issue - something wine does initiates the problem, but it continues after all wine apps have been closed.


3. foobillard 

Usually results in complete system lockup.


With mplayer and wine I see the following repeated in my Xorg.0.log:

(EE) R128(0): R128CCEWaitForIdle: (DEBUG) CCE idle took i = 1025
(EE) R128(0): Idle timed out, resetting engine...

This bug has close cousins documented: <a href='http://bugs.freedesktop.org/show_bug.cgi?id=5562'>5562</a> and <a href='https://bugs.freedesktop.org/show_bug.cgi?id=3493'>3493</a>

After viewing these bugs, I've collected various logs including a DRM debug log, which I have attached.
Comment 1 Rhys Ickeringill 2009-09-17 07:40:18 UTC
Created attachment 29632 [details]
Kernel log from about time of foobillard system lockup

DRM debug and this kernel log apply to the same foobillard system lockup (note the timestamps). Last line of this file is my futile attempt to reboot my system via ACPI...
Comment 2 Rhys Ickeringill 2009-09-17 07:44:09 UTC
Created attachment 29635 [details]
Kernel log associated with mplayer+xv+dma X freeze

Not sure about exact timing of the error logged with respect to the X freeze - it might be before, at or after the time of the freeze. However it *is* associated in some way, as i don't see this unless X freezes due to mplayer usage.
Comment 3 Rhys Ickeringill 2009-09-17 07:45:03 UTC
Created attachment 29636 [details]
X config file
Comment 4 Rhys Ickeringill 2009-09-17 08:12:13 UTC
Created attachment 29637 [details]
Xorg log associated with wine semi-random X freezes

R128CCEWaitForIdle errors are near EOF. Log captured prior to addition of DMAforXv setting.
Comment 5 Rhys Ickeringill 2009-09-17 08:23:16 UTC
Comment on attachment 29631 [details]
DRM debug output from foobillard crash

Am guessing that the repeated logging of "[drm:r128_freelist_get] returning NULL!" towards the DRM debug log EOF is significant. Perhaps the multiple calls to drm_lock near the beginning of this same log are too?
Comment 6 Matt Turner 2010-12-02 12:32:44 UTC
*** Bug 5562 has been marked as a duplicate of this bug. ***
Comment 7 Adam Jackson 2014-04-09 18:38:38 UTC
There's no longer an r128 DRI driver.


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.