Bug 29836

Summary:	bisected : MESA_FORMAT_S8_Z24' failed on some demos since e71a9042cb5e046072d2ea8fbd22a01de625bf8a Identify packed depth/stencil buffers using the Format field
Product:	Mesa	Reporter:	Andy Furniss <adf.lists>
Component:	Mesa core	Assignee:	mesa-dev
Status:	RESOLVED FIXED	QA Contact:
Severity:	normal
Priority:	medium	CC:	cedricv, nbowler, wolput
Version:	git
Hardware:	x86 (IA32)
OS:	Linux (All)
Whiteboard:
i915 platform:		i915 features:

Description Andy Furniss 2010-08-27 04:09:05 UTC

Running r600, but swrast is also affected.


There are only 'skip'ped commit left to test.
The first bad commit could be any of:
e71a9042cb5e046072d2ea8fbd22a01de625bf8a
7b07674667e8e34c936c417ab6f1c2e33093205b
We cannot bisect more!

Mesa demos engine, arbocclude and fogcoord fail since the above commits.

bash-3.2$ LIBGL_ALWAYS_SOFTWARE=1 ./engine
engine: main/depthstencil.c:229: put_mono_row_z24: Assertion `dsrb->Format == MESA_FORMAT_S8_Z24' failed.

bash-3.2$ ./engine 
engine: main/depthstencil.c:150: get_values_z24: Assertion `dsrb->Format == MESA_FORMAT_S8_Z24' failed.

Comment 1 Nick Bowler 2010-08-27 06:30:11 UTC

Can you try the patch posted here:

  http://lists.freedesktop.org/archives/mesa-dev/2010-August/002550.html

and see if it resolves your issue?

Comment 2 Nick Bowler 2010-08-27 07:08:39 UTC

*** Bug 29831 has been marked as a duplicate of this bug. ***

Comment 3 Andy Furniss 2010-08-27 07:55:10 UTC

(In reply to comment #1)
> Can you try the patch posted here:
> 
>   http://lists.freedesktop.org/archives/mesa-dev/2010-August/002550.html
> 
> and see if it resolves your issue?

Yes, but it regresses nexuiz (timedemo demos/demo1 with effects on normal) so that the floor/wall textures are mostly black.

Comment 4 Jos van Wolput 2010-08-27 08:30:23 UTC

(In reply to comment #1)
> Can you try the patch posted here:
> 
>   http://lists.freedesktop.org/archives/mesa-dev/2010-August/002550.html
> 
> and see if it resolves your issue?

I tried your patch on demos engine, arbocclude and fogcoord, they no longer
fail. The patch seems to resolves the issue.

Thanks!

Comment 5 Andy Furniss 2010-08-27 11:43:13 UTC

(In reply to comment #3)

> ... it regresses nexuiz (timedemo demos/demo1 with effects on normal) so
> that the floor/wall textures are mostly black.

Ignore that, I've managed to reproduce without the patch. It's an intermittent r600c issue that by chance only showed with the patch for the half a dozen runs I initially did.

Comment 6 Nick Bowler 2010-08-27 13:26:08 UTC

(In reply to comment #5)
> (In reply to comment #3)
> 
> > ... it regresses nexuiz (timedemo demos/demo1 with effects on normal) so
> > that the floor/wall textures are mostly black.
> 
> Ignore that, I've managed to reproduce without the patch. It's an intermittent
> r600c issue that by chance only showed with the patch for the half a dozen runs
> I initially did.

OK great, thanks for testing.

Comment 7 Brian Paul 2010-08-27 13:33:09 UTC

I've committed Nick's patch.
f57df7f2130dfe20a6f7e75cfc1f5d4e1717d098

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.