Bugzilla – Bug 42175
RV730: Display errors in glxgears & WebGL
Last modified: 2011-10-26 06:07:50 UTC
Since Mesa 7.11-rc2 the gallium driver is used on my machine and produces display errors. Find attached a glxgears screenshot and a screendump of http://learningwebgl.com/lessons/lesson03/index.html produced with the current Firefox.
Before and including 7.11-rc1 seemed to use the old driver which does not produce rendering errors but which crashes Firefox in http://learningwebgl.com/lessons/lesson16/index.html or in the Aquarium of http://code.google.com/p/webglsamples/.
Created attachment 52710 [details]
Created attachment 52713 [details]
Lesson 3 screenshot
Can you bisect?
(In reply to comment #3)
> Can you bisect?
The change took place between 7.11-rc1 and -rc2:
a8907c6005d7935b4520255e12184c139471b5b9 is the first bad commit
Author: Benjamin Franzke <firstname.lastname@example.org>
Date: Sat Jul 2 13:41:35 2011 +0200
But: This is where the switch from the old driver to the Gallium takes place. The old driver got used in my configuration up to 7.11-rc1 because the udev-devel packet was not installed. Eventually after installing it 7.11-rc1 uses the new driver.
So currently I don't have a "good" reference for a bisection.
BTW: https://bugzilla.mozilla.org/show_bug.cgi?id=693056 (origin of the current problems).
Mesa 7.12-devel (git-faa16dc) Works. No crash, no picture errors.
(In reply to comment #5)
> Mesa 7.12-devel (git-faa16dc) Works. No crash, no picture errors.
What about the current 7.11 branch? If that still shows the display errors, you should be able to bisect the change that fixed it on the master branch.
(In reply to comment #6)
> What about the current 7.11 branch? If that still shows the display errors, you
> should be able to bisect the change that fixed it on the master branch.
No bisection required. The bug was fixed by this commit:
"r600g: set correct tiling flags in depth info"
which I just verified by re-compiling mesa with the reversed patch applied.
(In reply to comment #7)
> No bisection required. The bug was fixed by this commit:
> "r600g: set correct tiling flags in depth info"
Weird, as the commit log says, it's supposed to be a no-op ATM...
Marek, does this make sense to you? Should this change be backported to the 7.11 branch?
No, it doesn't make sense. :) He might have an older, more buggy kernel than me. That would be the only explanation. I was working with the latest drm-fixes branch (3.1 rc10). Anyway I think the fix can safely go into 7.11.
(In reply to comment #9)
> He might have an older, more buggy kernel than me.
I'm using the longterm kernel 184.108.40.206.
Marek backported the fix to the 7.11 branch, thanks.