Bug 37490 - texture corruption in r600/r600g when using DRI2, no texture corruption in r600 with dri1
Summary: texture corruption in r600/r600g when using DRI2, no texture corruption in r6...
Status: RESOLVED WONTFIX
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/R600 (show other bugs)
Version: 7.10
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Default DRI bug account
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-05-23 01:58 UTC by William Pitcock
Modified: 2014-07-07 16:05 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
dmesg (13.97 KB, text/plain)
2011-05-23 21:29 UTC, William Pitcock
Details
glxinfo (13.35 KB, text/plain)
2011-05-23 21:31 UTC, William Pitcock
Details
actual dmesg (copy and paste fail) (13.97 KB, text/plain)
2011-05-23 21:40 UTC, William Pitcock
Details
Xorg.0.log (59.27 KB, text/plain)
2011-05-23 21:41 UTC, William Pitcock
Details

Description William Pitcock 2011-05-23 01:58:30 UTC
hi,

i get texture corruption as seen in the attached screenshot when using dri2.  when running more 3D intensive applications (like say, etracer), the texture corruption increases resulting in some graphics primitives not being rendered.

this does not happen with DRI1, with DRI1, there is no corruption and all primitives are rendered as expected.

this is on linux 2.6.38.6 with alpine patches, mesa 7.10.2 and xorg 1.10 with xf86-video-ati 6.14.1, libdrm 2.4.25 and uclibc 0.9.32-rc3 (also with alpine patches).

there are some bugs that look similar to what i am seeing, but they are all marked RESOLVED, and none of them mention DRI2.
Comment 1 William Pitcock 2011-05-23 02:00:48 UTC
it seems that attachment sizes are limited.

see http://nenolod.net/~nenolod/radeon-dri2-gltexture-corruption.png for an example of what i am seeing.

i think this is due to incomplete writes to the card or perhaps an off-by-one of some kind.
Comment 2 William Pitcock 2011-05-23 02:36:47 UTC
oh, i forgot to mention, this is on a dual radeon hd 4870 setup.
Comment 3 Alex Deucher 2011-05-23 05:49:22 UTC
Please attach your xorg log, dmesg output, and glxinfo output.
Comment 4 Fredrik Höglund 2011-05-23 10:22:09 UTC
Judging by the screenshot this may be a duplicate of bug 33824.

The logout effect in KWin uses automatically generated mipmaps + LOD bias to blur all windows except the dialog in the center of the screenshot.

Does the problem go away if you set ColorTiling to False in your xorg.conf file?
Comment 5 Tobias Jakobi 2011-05-23 10:31:38 UTC
This is most likely _no_ duplicate, since tiling still has to be force on with an envvar. And I don't see this happening from the explanations in the initial post.
Comment 6 William Pitcock 2011-05-23 12:45:35 UTC
re: color tiling, i believe it is disabled by default, but i will try and see if that is the problem.

i will also get dmesg / Xorg logs when i switch back over to DRI2.
Comment 7 William Pitcock 2011-05-23 21:29:45 UTC
Created attachment 47086 [details]
dmesg
Comment 8 William Pitcock 2011-05-23 21:31:43 UTC
Created attachment 47087 [details]
glxinfo
Comment 9 William Pitcock 2011-05-23 21:40:50 UTC
Created attachment 47088 [details]
actual dmesg (copy and paste fail)
Comment 10 William Pitcock 2011-05-23 21:41:53 UTC
Created attachment 47089 [details]
Xorg.0.log
Comment 11 William Pitcock 2011-05-23 21:44:01 UTC
glxinfo, dmesg and xorg log have been attached.

KMS color tiling is disabled.  color tiling is also disabled without KMS.
enabling color tiling on DRI1 does not trigger the bug.
Comment 12 Michel Dänzer 2011-05-24 08:43:15 UTC
Could be a mipmap generation issue then. Is it still broken with r600g from upstream Git master?
Comment 13 William Pitcock 2011-05-24 13:34:59 UTC
i don't know, and my concern is mostly with the non-gallium driver right now.

i can test the gallium driver, but i need the normal driver to work as i have 32-bit applications i am running in a debian chroot.  thusly, i do not really use the gallium driver yet... i only tested it to see if it did the same thing, which it does.

the result is not the same everytime you trigger that kwin effect though; sometimes i get part of a console framebuffer (e.g. radeondrmfb).

so would it make more sense for me to test non-gallium driver first?
Comment 14 Alex Deucher 2011-05-24 14:19:34 UTC
(In reply to comment #13)
> i don't know, and my concern is mostly with the non-gallium driver right now.
> 
> i can test the gallium driver, but i need the normal driver to work as i have
> 32-bit applications i am running in a debian chroot.  thusly, i do not really
> use the gallium driver yet... i only tested it to see if it did the same thing,
> which it does.
> 
> the result is not the same everytime you trigger that kwin effect though;
> sometimes i get part of a console framebuffer (e.g. radeondrmfb).
> 
> so would it make more sense for me to test non-gallium driver first?

No one is really working on the non gallium driver any more, so it's not likely to get much more attention.
Comment 15 William Pitcock 2011-05-24 14:32:50 UTC
(In reply to comment #14)
> (In reply to comment #13)
> > i don't know, and my concern is mostly with the non-gallium driver right now.
> > 
> > i can test the gallium driver, but i need the normal driver to work as i have
> > 32-bit applications i am running in a debian chroot.  thusly, i do not really
> > use the gallium driver yet... i only tested it to see if it did the same thing,
> > which it does.
> > 
> > the result is not the same everytime you trigger that kwin effect though;
> > sometimes i get part of a console framebuffer (e.g. radeondrmfb).
> > 
> > so would it make more sense for me to test non-gallium driver first?
> 
> No one is really working on the non gallium driver any more, so it's not likely
> to get much more attention.

well, i will try and see if r600g git master solves the problem.  unfortunately, i think this means i'm screwed when it comes to running proprietary apps in chroots.
Comment 16 Andreas Boll 2012-08-09 15:03:34 UTC
(In reply to comment #15)
> (In reply to comment #14)
> > (In reply to comment #13)
> > > i don't know, and my concern is mostly with the non-gallium driver right now.
> > > 
> > > i can test the gallium driver, but i need the normal driver to work as i have
> > > 32-bit applications i am running in a debian chroot.  thusly, i do not really
> > > use the gallium driver yet... i only tested it to see if it did the same thing,
> > > which it does.
> > > 
> > > the result is not the same everytime you trigger that kwin effect though;
> > > sometimes i get part of a console framebuffer (e.g. radeondrmfb).
> > > 
> > > so would it make more sense for me to test non-gallium driver first?
> > 
> > No one is really working on the non gallium driver any more, so it's not likely
> > to get much more attention.
> 
> well, i will try and see if r600g git master solves the problem. 
> unfortunately, i think this means i'm screwed when it comes to running
> proprietary apps in chroots.

hi William,

did you have any chance to test r600g git master in the mean time?
Comment 17 Andreas Boll 2014-07-07 16:05:28 UTC
The classic r600 driver has been abandoned long ago.
It was replaced by the Gallium driver r600g.

If you have issues with r600g please file a new bug report with component Drivers/Gallium/r600

Thanks.


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.