Bug 826000

Summary: The X server crashes when using MGA driver and EXA acceleration
Product: [Fedora] Fedora Reporter: Göran Uddeborg <goeran>
Component: xorg-x11-serverAssignee: X/OpenGL Maintenance List <xgl-maint>
Status: CLOSED WONTFIX QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 17CC: xgl-maint
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2013-08-01 13:32:02 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
Xorg log from a crashed session none

Description Göran Uddeborg 2012-05-29 10:10:32 UTC
Created attachment 587377 [details]
Xorg log from a crashed session

Description of problem:
After the XAA acceleration was removed, the MGA driver doesn't behave well.  (Also noted in bug 824569.)  It is significantly slower, and there are several small bugs in the rendering.

I tried to use the EXA acceleration instead, by setting

	Option "AccelMethod" "EXA"

in my xorg.conf.  After that, the server crashes when I try to start a session.  Just starting an empty server doesn't trigger the problem, and some clients can connect to it without errors.  But many other clients makes it crash.  A simple test case I've been using has been to start xclock.



Version-Release number of selected component (if applicable):
xorg-x11-server-Xorg-1.12.0-5.fc17.x86_64
xorg-x11-drv-mga-1.4.13-19.20120104git9223c44a7.fc17.x86_64
xorg-x11-apps-7.6-5.fc17.x86_64

How reproducible:
Every time

Steps to Reproduce:
1. Xorg :42
2. xclock  -display :42

Actual results:
The server crashes

Expected results:
The server should run, showing an xclock

Additional info:
I've tried to look at this in a debugger, and I get the backtrace shown below.  My limited understanding is that the call from exaCopyDirty() in exa_migration_classic.c to mgaDownloadFromScreen() in mga_exa.c breaks, most likely because some inconsistency in the argument.  If the bug is exaCopyDirty sending wrong arguments, or mgaDownloadFromScreen treating them incorrectly, I don't know, so assigning this to the server package is just a guess from my side.  AFAIK, it could just as well be a bug in the MGA driver.

#0  __memcpy_sse2 () at ../sysdeps/x86_64/memcpy.S:197
#1  0x00007fa3f0b1d2fe in memcpy (__len=32, __src=0x14650, __dest=0x238afd8) at /usr/include/bits/string3.h:52
#2  mgaDownloadFromScreen (pSrc=pSrc@entry=0x237dd40, x=<optimized out>, y=<optimized out>, w=<optimized out>, h=<optimized out>, dst=0x238afd8 "", dst_pitch=656) at mga_exa.c:745
#3  0x00007fa3f06eba7b in exaCopyDirty (migrate=migrate@entry=0x7fff9ebc7550, pValidDst=0x237ddc8, pValidSrc=0x237ddd8, transfer=0x7fa3f0b1d290 <mgaDownloadFromScreen>, fallback_index=fallback_index@entry=1, sync=0x7fa3f06ea6e0 <exaWaitSync>) at exa_migration_classic.c:220
#4  0x00007fa3f06ebdc4 in exaCopyDirtyToSys (migrate=migrate@entry=0x7fff9ebc7550) at exa_migration_classic.c:281
#5  0x00007fa3f06ebe38 in exaDoMoveOutPixmap (migrate=migrate@entry=0x7fff9ebc7550) at exa_migration_classic.c:396
#6  0x00007fa3f06ec914 in exaDoMigration_classic (pixmaps=0x7fff9ebc7550, npixmaps=<optimized out>, can_accel=0) at exa_migration_classic.c:704
#7  0x00007fa3f06eca14 in exaPrepareAccessReg_classic (pPixmap=0x237dd40, index=0, pReg=<optimized out>) at exa_migration_classic.c:742
#8  0x00007fa3f06f7f68 in ExaPrepareCompositeReg (height=2, width=8, yDst=81, xDst=148, yMask=0, xMask=0, ySrc=0, xSrc=0, pDst=0x2350210, pMask=0x21df840, pSrc=0x23504d0, op=3 '\003', pScreen=<optimized out>) at exa_unaccel.c:599
#9  ExaCheckComposite (op=op@entry=3 '\003', pSrc=pSrc@entry=0x23504d0, pMask=pMask@entry=0x21df840, pDst=pDst@entry=0x2350210, xSrc=xSrc@entry=0, ySrc=ySrc@entry=0, xMask=xMask@entry=0, yMask=yMask@entry=0, xDst=148, yDst=81, width=width@entry=8, height=height@entry=2) at exa_unaccel.c:624
#10 0x00007fa3f06f4898 in exaComposite (op=3 '\003', pSrc=0x23504d0, pMask=0x21df840, pDst=0x2350210, xSrc=<optimized out>, ySrc=<optimized out>, xMask=0, yMask=0, xDst=<optimized out>, yDst=<optimized out>, width=8, height=2) at exa_render.c:1066
#11 0x0000000000501989 in damageComposite (op=3 '\003', pSrc=0x23504d0, pMask=0x21df840, pDst=0x2350210, xSrc=0, ySrc=0, xMask=0, yMask=0, xDst=148, yDst=81, width=8, height=2) at damage.c:569
#12 0x00007fa3f06f595f in exaTrapezoids (op=3 '\003', pSrc=0x23504d0, pDst=0x2350210, maskFormat=<optimized out>, xSrc=0, ySrc=<optimized out>, ntrap=<optimized out>, traps=0x22c58f8) at exa_render.c:1181
#13 0x00000000004fb4cb in ProcRenderTrapezoids (client=0x22c9220) at render.c:777
#14 0x000000000043439a in Dispatch () at dispatch.c:439
#15 0x00000000004233e5 in main (argc=2, argv=0x7fff9ebc7ad8, envp=<optimized out>) at main.c:287

Comment 1 Fedora End Of Life 2013-07-04 04:25:52 UTC
This message is a reminder that Fedora 17 is nearing its end of life.
Approximately 4 (four) weeks from now Fedora will stop maintaining
and issuing updates for Fedora 17. It is Fedora's policy to close all
bug reports from releases that are no longer maintained. At that time
this bug will be closed as WONTFIX if it remains open with a Fedora 
'version' of '17'.

Package Maintainer: If you wish for this bug to remain open because you
plan to fix it in a currently maintained version, simply change the 'version' 
to a later Fedora version prior to Fedora 17's end of life.

Bug Reporter:  Thank you for reporting this issue and we are sorry that 
we may not be able to fix it before Fedora 17 is end of life. If you 
would still like  to see this bug fixed and are able to reproduce it 
against a later version  of Fedora, you are encouraged  change the 
'version' to a later Fedora version prior to Fedora 17's end of life.

Although we aim to fix as many bugs as possible during every release's 
lifetime, sometimes those efforts are overtaken by events. Often a 
more recent Fedora release includes newer upstream software that fixes 
bugs or makes them obsolete.

Comment 2 Fedora End Of Life 2013-08-01 13:32:11 UTC
Fedora 17 changed to end-of-life (EOL) status on 2013-07-30. Fedora 17 is 
no longer maintained, which means that it will not receive any further 
security or bug fix updates. As a result we are closing this bug.

If you can reproduce this bug against a currently maintained version of 
Fedora please feel free to reopen this bug against that version.

Thank you for reporting this bug and we are sorry it could not be fixed.