Bug 25078

Summary: "x11perf -wdcircle100 -time 1 -repeat 1 -rop GXxor" crashes X
Product: xorg Reporter: David H <david.hedberg>
Component: Server/Acceleration/EXAAssignee: Xorg Project Team <xorg-team>
Status: RESOLVED FIXED QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: medium CC: analyzer1, jlt_fdo, pva, trueurssian
Version: unspecified   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
Log/backtrace from X
none
Xorg.log including the backtrace none

Description David H 2009-11-13 09:11:59 UTC
Created attachment 31178 [details]
Log/backtrace from X

Running "x11perf -wdcircle100 -time 1 -repeat 1 -rop GXxor" seems to be a sure way to crash X. The crash was originally found triggered by running a certain application in wine, but as far as I could determine back when I investigated it, this is the exact same crash.

I'm using a HD3200 IGP, and while I'm currently running xf86-video-ati from todays git and xorg-server-1.7.1, this same (I think) bug has actually been present for at least a few months. Log from Xorg including backtrace attached.
Comment 1 Michel Dänzer 2009-11-13 09:43:01 UTC
This is an EXA bug, a fix is being discussed on the xorg-devel mailing list.
Comment 2 Maarten Maathuis 2009-11-16 10:58:18 UTC
This appears to be a bug not (obviously) related to exa. At least not when several other things were fixed (so it's possible you're crashing on something else). For me it crashed in mi/miarc.c:1213, upon dereferencing a arcData0. My initial impression is that arcIndex is out of bounds, but why, i don't know.

rop GXCopy works for me now, rop GXxor not, but these patches haven't made it upstream yet.
Comment 3 Michel Dänzer 2009-11-20 01:06:25 UTC
(In reply to comment #2)
I'm a little confused by this comment, but it sounds like your patch will actually fix this particular crash, but there's more down the road? If so you can either reassign this back once the patch lands, or resolve it and track the remaining crashes in separate reports.
Comment 4 Michel Dänzer 2009-11-26 02:04:54 UTC
*** Bug 25266 has been marked as a duplicate of this bug. ***
Comment 5 Jens Taprogge 2010-02-06 02:05:19 UTC
Are there any news on this bug?  Any known workarounds?

I am seeing this bug on an ia32 machine with an x300 radeon card.  It started showing when Debian upgraded its xserver-xorg-core package to version 1.7 (and going back fixes the issue).
Comment 6 Marco Salm 2010-02-18 11:40:21 UTC
Created attachment 33395 [details]
Xorg.log including the backtrace

I also see this bug both on my desktop (amd64, ATI Technologies Inc Radeon HD 3200 Graphics) and on my notebook (x86, Radeon Mobility 9200 (AGP)).
It happens if I use xfig and draw a circle.

I use xorg-server-1.7.5 and xf86-video-ati (git master commit a3b730eceb522c7ac1ef3dd6f6c7d773118d03f7).
Comment 7 Michel Dänzer 2010-02-19 06:01:33 UTC
Maarten, any news on the stuff you mentioned in comment #2?
Comment 8 Krzysztof Sachanowicz 2010-02-28 17:37:29 UTC
I can confirm this bug. An obvious workaround is to add
Option "AccelMethod" "XAA"
to xorg.conf. The bug also happens when drawing a circle in xfig.
Comment 9 Adam Jackson 2018-06-11 20:56:45 UTC
Unable to reproduce this with either Xvfb or Xephyr -fakexa.

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.