Bug 26372

Created attachment 32986 [details]
Xorg.conf showing error from 3192.857

http://intellinuxgraphics.org/how_to_report_bug.html 

It looks like this is a kernel issue

2.6.32 works perfectly on my system with my current user space

I git bisected back to 4e2ccdb0409146f8cf64a11b6ef82a9c928ced2a which doesn't really make sense but 3ef884b4c04e857c283cc77ca70ad8f638d94b0e seems more likely

git bisect start
# good: [22763c5cf3690a681551162c15d34d935308c8d7] Linux 2.6.32
git bisect good 22763c5cf3690a681551162c15d34d935308c8d7
# bad: [b9b4fd4b44a3959f59d3db6a9fb65d2f63aa9824] Merge branch 'drm-intel-next' of git://git.kernel.org/pub/scm/linux/kernel/git/anholt/drm-intel into drm-intel-next
git bisect bad b9b4fd4b44a3959f59d3db6a9fb65d2f63aa9824
# good: [d2fdb776e08d4231d7e86a879cc663a93913c202] dm snapshot: use merge origin if snapshot invalid
git bisect good d2fdb776e08d4231d7e86a879cc663a93913c202
# bad: [238ccbb050a243e935bb3fc679c2e4bbff7004aa] Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
git bisect bad 238ccbb050a243e935bb3fc679c2e4bbff7004aa
# bad: [9c3936cb694ffd559c80dc3eb75b61f769a39259] Merge branch 'omap-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap-2.6
git bisect bad 9c3936cb694ffd559c80dc3eb75b61f769a39259
# bad: [748e566b7e24541e05e3e70be311887a1262f2a1] Merge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6
git bisect bad 748e566b7e24541e05e3e70be311887a1262f2a1
# good: [3ef884b4c04e857c283cc77ca70ad8f638d94b0e] Merge branch 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6
git bisect good 3ef884b4c04e857c283cc77ca70ad8f638d94b0e
# bad: [11bd04f6f35621193311c32e0721142b073a7794] Merge branch 'linux-next' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci-2.6
git bisect bad 11bd04f6f35621193311c32e0721142b073a7794
# good: [56ddf4d3cf04e80254d3d721c6bea2f8ec44c41a] x86/PCI: MMCONFIG: add resource to struct pci_mmcfg_region
git bisect good 56ddf4d3cf04e80254d3d721c6bea2f8ec44c41a
# good: [2093abf9cbcec3cb1409a67d8bce51854595b1d5] nilfs2: change return type of nilfs_commit_chunk
git bisect good 2093abf9cbcec3cb1409a67d8bce51854595b1d5
# good: [fbb5de70bbe13ecbebb04226dd6d52b1258dc247] PCI: portdrv: add missing irq cleanup
git bisect good fbb5de70bbe13ecbebb04226dd6d52b1258dc247
# bad: [4e2ccdb0409146f8cf64a11b6ef82a9c928ced2a] Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ryusuke/nilfs2
git bisect bad 4e2ccdb0409146f8cf64a11b6ef82a9c928ced2a
# good: [5f1586d0dd8f6eeecf6c0d35cbca6291afd6f1cc] nilfs2: do not return io error for bio allocation failure
git bisect good 5f1586d0dd8f6eeecf6c0d35cbca6291afd6f1cc
# good: [9c965bac169f786cc6cca8ff81d3b636e923c960] nilfs2: hide nilfs_write_info struct in segment buffer code
git bisect good 9c965bac169f786cc6cca8ff81d3b636e923c960
# good: [a694291a6211537189c6080f77f63cdabfc9b63e] nilfs2: separate wait function from nilfs_segctor_write
git bisect good a694291a6211537189c6080f77f63cdabfc9b63e

So you say 3ef884b4c04e857c283cc77ca70ad8f638d94b0e is more likely.  Your log shows that you tested it and marked it good.  Was it really good?  If so, then graphics would pretty clearly be off the hook.  If it isn't good, then you need to bisect reset, edit the log, then replay it and continue bisecting.

This bug has been fixed - sorry I should have closed the bug as soon as I noticed

However the last week or so I've been experiencing lots of GPU hangs I think this may have something to do with using the latest cairo from git

However I'm seeing graphical corruptions even using the stable version of cairo in 2D and I can't start compositing on KDE

This is all working fine on a radeon setup with the latest git too, I think the problem maybe with in libdrm as the 2D artefacts appear in the latest testing kernel and a 2.6.33 kernel too

The gpu hangs were caused by using the drm backend for i915 & i965

But I'm still getting a lot of artifacts in 2D using an old kernel & new, using stable cairo and new

(In reply to comment #5)
> This bug has been fixed - sorry I should have closed the bug as soon as I
> noticed

then closing.
 
> However the last week or so I've been experiencing lots of GPU hangs I think
> this may have something to do with using the latest cairo from git
> 
> However I'm seeing graphical corruptions even using the stable version of cairo
> in 2D and I can't start compositing on KDE
> 
> This is all working fine on a radeon setup with the latest git too, I think the
> problem maybe with in libdrm as the 2D artefacts appear in the latest testing
> kernel and a 2.6.33 kernel too

Please file new bugs for the other issues.