Bug 13208

Summary:	r300 crash in stepmania 4 snapshot while trying to switch video modes
Product:	Mesa	Reporter:	Simon <Simon80>
Component:	Drivers/DRI/r300	Assignee:	Default DRI bug account <dri-devel>
Status:	RESOLVED FIXED	QA Contact:
Severity:	critical
Priority:	medium
Version:	unspecified
Hardware:	x86 (IA32)
OS:	Linux (All)
Whiteboard:
i915 platform:		i915 features:

Description Simon 2007-11-12 15:18:31 UTC

Stepmania 4 snapshots in the past year have crashed intermittently during attempts to change the video mode it is displaying a window in.  I always run it in windowed mode, and using alt+tab (I'm running GNOME) seems to bring out the problems rather quickly in this and other applications.

Backtrace below:

#0  0xb3638ace in r300DestroyTexObj () from /usr/lib/dri/r300_dri.so
#1  0xb3621709 in driDestroyTextureObject () from /usr/lib/dri/r300_dri.so
#2  0xb363ab2b in ?? () from /usr/lib/dri/r300_dri.so
#3  0x0ad58a68 in ?? ()
#4  0x0ad26890 in ?? ()
#5  0xb76cd140 in ?? () from /lib/tls/i686/cmov/libc.so.6
#6  0x09eb9fd0 in ?? ()
#7  0xb76cd154 in main_arena () from /lib/tls/i686/cmov/libc.so.6
#8  0x0ad71a08 in ?? ()
#9  0x0a7332e8 in ?? ()
#10 0x00000000 in ?? ()

Comment 1 Michel Dänzer 2007-11-21 00:42:03 UTC

A more complete backtrace (bt full and surrounding info such as the signal caught,  r300_dri.so built with debugging symbols) could be useful.

Comment 2 Simon 2007-11-21 05:57:33 UTC

(In reply to comment #1)
> A more complete backtrace (bt full and surrounding info such as the signal
> caught,  r300_dri.so built with debugging symbols) could be useful.
> 

I'm using the r300_dri.so that comes with the libgl1-mesa-dri package in Ubuntu 7.10, and I have debugging symbols for it installed with the libgl1-mesa-dri-dbg package.  I'm not sure what is causing the missing symbols.  Here is another reproduction of the crash, with some additional info:

[New Thread -1387619440 (LWP 20781)]
[Thread -1387619440 (LWP 20781) exited]
/////////////////////////////////////////
WARNING: X11 Protocol error BadValue (integer parameter out of range for operation) (2) has occurred, caused by request 128,9, resource ID 88081058
/////////////////////////////////////////

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1220700464 (LWP 20772)]
0xb373bace in r300DestroyTexObj () from /usr/lib/dri/r300_dri.so
(gdb) bt
#0  0xb373bace in r300DestroyTexObj () from /usr/lib/dri/r300_dri.so
#1  0xb3724709 in driDestroyTextureObject () from /usr/lib/dri/r300_dri.so
#2  0xb373db2b in ?? () from /usr/lib/dri/r300_dri.so
#3  0x0ad59f88 in ?? ()
#4  0x0ad73110 in ?? ()
#5  0xb776e140 in ?? () from /lib/tls/i686/cmov/libc.so.6
#6  0x0a114fd8 in ?? ()
#7  0xb776e154 in main_arena () from /lib/tls/i686/cmov/libc.so.6
#8  0x0ad3c468 in ?? ()
#9  0x0a731828 in ?? ()
#10 0x00000000 in ?? ()

Comment 3 Maciej Cencora 2009-05-21 06:22:05 UTC

Can you check current mesa?
09c04db3c900e4ed833d060853b48c7ca23697e1 or c28707b50701b1cf8727be29d61e2d939c6ee58f could have fixed this issue.

Comment 4 Simon 2009-05-21 21:16:25 UTC

(In reply to comment #3)
> Can you check current mesa?
> 09c04db3c900e4ed833d060853b48c7ca23697e1 or
> c28707b50701b1cf8727be29d61e2d939c6ee58f could have fixed this issue.
> 

Sorry, the laptop I was running r300 on started having weird video memory corruption as of March 2008 (every other tile had vertical black stripes 1px wide), which forced me to replace it. As a result, I can't test r300 anymore.

Comment 5 Jerome Glisse 2009-05-22 00:56:39 UTC

Closing this bugs as it's not reproducable, anyone reopen if you have same problem with recent mesa 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.