[g33] (Needs 2.7.99.901+git20090619) X freeze while using firefox (UXA bug)

[Steve Beattie]

Binary package hint: xserver-xorg-video-intel

I was just opening up a firefox tab to fill out an apport crash report when X hung. The result of ubuntu-bug is attached; it starts up xkbcomp which never finishes here, so I had to kill that. Also attached is the gdb backtrace of Xorg process while it was hung, along with an inspection of some of the pointers at the deepest level that I could get debugging information for.

ProblemType: Bug
Architecture: amd64
DistroRelease: Ubuntu 9.04
Package: xserver-xorg-video-intel 2:2.6.3-0ubuntu5
ProcEnviron:
 PATH=(custom, user)
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcVersion: Linux version 2.6.28-11-server (buildd@crested) (gcc version 4.3.3 (Ubuntu 4.3.3-5ubuntu4) ) #40-Ubuntu SMP Fri Apr 3 18:30:08 UTC 2009
SourcePackage: xserver-xorg-video-intel
Uname: Linux 2.6.28-11-server x86_64
Xrandr:

glxinfo:

setxkbmap:

xdpyinfo:

xkbcomp:

#0 0x00007fa9ad24ecd7 in ioctl () from /lib/libc.so.6
#1 0x00007fa9ab1bc3bd in drm_intel_gem_bo_start_gtt_access () from /usr/lib/libdrm_intel.so.1
#2 0x00007fa9ab40cd72 in i830_uxa_prepare_access (pixmap=0x27ab2d0, access=UXA_ACCESS_RW)
    at ../../src/i830_exa.c:817
 scrn = (ScrnInfoPtr) 0x1f5c950
 i830 = (I830Ptr) 0x1f5eaf0
 pScrn = (ScrnInfoPtr) 0x1f5c950
 bo = (drm_intel_bo *) 0x27ab050
#3 0x00007fa9ab4294fe in uxa_check_poly_fill_rect (pDrawable=0xb, pGC=0x1fb75a0, nrect=-1217971184,
    prect=0xffffffffffffffff) at ../../uxa/uxa-unaccel.c:239
#4 0x00007fa9ab426e4e in uxa_create_alpha_picture (pScreen=0x1f8d920, pDst=<value optimized out>,
    pPictFormat=0x1f8ea98, width=200, height=16) at ../../uxa/uxa-render.c:840
 error = 0
 rect = {x = 0, y = 0, width = 200, height = 16}
#5 0x00007fa9ab42744c in uxa_trapezoids (op=3 '\003', pSrc=0x280be20, pDst=0x24efb10, maskFormat=0x1f8ea98,
    xSrc=690, ySrc=-149, ntrap=27, traps=0x2813f44) at ../../uxa/uxa-render.c:908
 pPicture = <value optimized out>
 pScreen = (ScreenPtr) 0x1f8d920
 bounds = {x1 = 690, y1 = 0, x2 = 890, y2 = 16}
#6 0x00000000005349cd in ProcRenderTrapezoids (client=0x23e5000) at ../../render/render.c:782
#7 0x000000000044e304 in Dispatch () at ../../dix/dispatch.c:437
 client = (ClientPtr) 0x23e5000
 nready = 0
 start_tick = 265440
#8 0x0000000000433d8d in main (argc=10, argv=0x7fffb7673f98, envp=<value optimized out>)
    at ../../dix/main.c:397
 i = 1
 alwaysCheckForInput = {0, 1}

[lspci]
00:00.0 Host bridge [0600]: Intel Corporation 82G33/G31/P35/P31 Express DRAM Controller [8086:29c0] (rev 02)
     Subsystem: Holco Enterprise Co, Ltd/Shuttle Computer Device [1297:3106]
00:02.0 VGA compatible controller [0300]: Intel Corporation 82G33/G31 Express Integrated Graphics Controller [8086:29c2] (rev 02)
     Subsystem: Holco Enterprise Co, Ltd/Shuttle Computer Device [1297:3106]

[Steve Beattie (sbeattie)]

Attached is the relavent portion of /var/log/kern.log

[Bryce Harrington (bryce)]

We have a PPA with some new tools for debugging X freezes:

  https://launchpad.net/~ubuntu-x-swat/+archive/x-freeze-test

I know how irritating X freezes can be. They're also typically
quite hard to debug, but the information provided by these new
tools should help upstream figure them out.

You can help by doing the following:

 A. Install the PPA packages on Jaunty and boot kernel 2.6.30-rc2
 B. Reproduce your freeze
 C. ssh into the machine and run the steps to collect the info
 D. Attach the tarball of the results to this bug report

With this information, we'll be able to forward your bug upstream.

(For more information on triaging X freeze bugs, see
 https://wiki.ubuntu.com/X/Troubleshooting/Freeze )

[Martin Schaaf (mascha)]

The tarball of the debugging results.

[Bryce Harrington (bryce)]

Thanks Martin, however in order to upstream this, we need a consistent set of data from a single machine, not a mix of some stuff from you, some from Steve. So would you mind running 'apport-collect 356850' to gather this data?

@Steve, it would be great if we could also get a dri dump from you (see comment #3) as well.

[Martin Schaaf (mascha)]

Architecture: i386
DistroRelease: Ubuntu 9.04
Package: xserver-xorg-video-intel 2:2.6.3-0ubuntu9
PackageArchitecture: i386
ProcEnviron:
 SHELL=/bin/bash
 PATH=(custom, user)
 LANG=de_DE.UTF-8
ProcVersion: Linux version 2.6.30-020630rc2-generic (root@zinc) (gcc version 4.2.3 (Ubuntu 4.2.3-2ubuntu7)) #020630rc2 SMP Wed Apr 15 14:00:27 UTC 2009
Uname: Linux 2.6.30-020630rc2-generic i686
UserGroups: adm admin audio cdrom dialout dip fax floppy fuse lpadmin netdev plugdev powerdev scanner tape vboxusers video

[Bryce Harrington (bryce)]

Forwarding this bug from a Ubuntu reporter:
https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/356850

[Problem]
A freeze occurred while opening a firefox tab.

Two users report seeing this same problem in the aforementioned bug, however this report focuses on Martin Schaaf's issue.

 distro: Ubuntu
 architecture: i686
 kernel: 2.6.30-020630rc2-generic
 xserver-xorg: Installiert: 1:7.4~5ubuntu18
 mesa: Installiert: 7.4-0ubuntu3
 libdrm: Installiert: 2.4.9-1build1
 -intel: Installiert: 2:2.6.3-0ubuntu9

[lspci]
00:00.0 Host bridge [0600]: Intel Corporation Mobile 945GM/PM/GMS, 943/940GML and 945GT Express Memory Controller Hub [8086:27a0] (rev 03)
 Subsystem: Dell Device [1028:01d7]
00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller [8086:27a2] (rev 03)
 Subsystem: Dell Device [1028:01d7]

[Bryce Harrington (bryce)]

Created an attachment (id=25486)
dri_debug.tgz

[Bryce Harrington (bryce)]

Created an attachment (id=25487)
XorgConf.txt

[Bryce Harrington (bryce)]

Created an attachment (id=25488)
XorgLog.txt

[Bryce Harrington (bryce)]

Created an attachment (id=25489)
XorgLogOld.txt

[Steve Beattie (sbeattie)]

Bryce, I tried to reproduce the freeze with UXA enabled while using the packages from https://edge.launchpad.net/~ubuntu-x-swat/+archive/x-freeze-test, the rc4 kernel from http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.30-rc4/ and the repro.sh script. Instead of freezing, something in the software stack leaked enough memory that compiz got killed by the oom-killer (machine has 4GB of RAM and 4GB of swap and was otherwise idle) after a few hours.

Sorry for the inconclusive result; I'll try to reproduce again when I can free that machine from other tasks.

[Gordon Jin (gordon-jin)]

dmesg says:

[ 95.353976] [drm:i915_get_vblank_counter] *ERROR* trying to get vblank count for disabled pipe 0

[ 3749.511387] [drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
[ 3749.511395] [drm:i915_gem_object_pin] *ERROR* Failure to bind: -12
[ 3749.511399] [drm:i915_gem_evict_something] *ERROR* inactive empty 1 request empty 1 flushing empty 1
[ 3749.511403] [drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
[ 3749.511406] [drm:i915_gem_object_pin] *ERROR* Failure to bind: -12
[ 3749.511409] [drm:i915_gem_execbuffer] *ERROR* Failed to pin buffers -12

[Bryce Harrington (bryce)]

Martin,

I've forwarded your issue upstream to https://bugs.freedesktop.org/show_bug.cgi?id=21572 - please subscribe to this bug report in case upstream needs further information or wishes you to test something. Thanks ahead of time.

Also, in case they wish you to test an even newer version of the driver, you can check here for git snapshots: https://edge.launchpad.net/~xorg-edgers/+archive/ppa

[Zack Evans (zevans23)]

@Steve, see bug 360319 for memory leak. The final comment on that one suggests Intel driver 2.7.0 fixes the leak - guessing you're on 2.6.3 given the repos you mention.

Re your hang, there are several later 2.6.28 kernels in circulation now, could try some of those and see if you can bisect if the latest one fixes it?

[Jesse Barnes (jbarnes-virtuousgeek)]

Adjusting severity: crashes & hangs should be marked critical.

[Eric Anholt (eric-anholt)]

This is probably related to the specific website being viewed. We need to know what that is.

[Jesse Barnes (jbarnes-virtuousgeek)]

Yeah the original backtrace indicates a legacy path is being taken... sounds like newer drivers fix the issue but introduce a memory leak. Can you confirm with the xorg edgers bits?

[Tobias-seriesdb (tobias-seriesdb)]

To #7, you said it was probably related to specific sites. I have found that this page makes the freeze consistently.

http://blogs.discovermagazine.com/badastronomy/2009/06/11/saturns-rings-do-the-wave/

[TobiasTheCommie (lp453452)]

Just a note, i get this consistently on

http://blogs.discovermagazine.com/badastronomy/2009/06/11/saturns-rings-do-the-wave/

Before i went to the karmic alpha it happened quite a bit, with karmic it rarely freezes anymore. But the above link has consistently made my computer freeze.

Thought it might help with debug.

[Martin Schaaf (mascha)]

Currently I use version 2.7.99.901+git20090619.534e73ad and can't any longer reproduce the freeze. For now the only problem I have is that after closing and opening the lid the background light isn't turned on again.

[Eric Anholt (eric-anholt)]

The original reporter says it's fixed, so assume that the fixes for firefox since 2.6.3 fixed it.

[Bryce Harrington (bryce)]

I notice the upstream bug here has been closed as fixed as of 2.7.99.901+git20090619.534e73ad... I guess we ought to close this bug as fixed, although I'm still not 100% sure that steve and martin had exactly the same issue. Let's be optimistic and assume so.

@Steve, if it does turn out your original issue still exists, feel free to either reopen this bug or (even better) file a new one. In either case we would still need that dump in order to get upstream's attention on it.