Bug 27018

Summary: swrast not working correctly - samples/stencil
Product: Mesa Reporter: Andrew Randrianasulu <randrik>
Component: OtherAssignee: mesa-dev
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: medium    
Version: git   
Hardware: x86 (IA32)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments: correct rendering from rv280
wrong rendering from swrast
same glxinfo -l as in bug #27017

Description Andrew Randrianasulu 2010-03-11 06:44:37 UTC
Using mesa's swrast i have incorrect rendering in progs/samples/stencil, see picture

I have mesa master up to commit a0ae2ca033ec2024da1e01d1c11c0437837c031b ( st/dri: move extension initilization to st/mesa completely )
Comment 1 Andrew Randrianasulu 2010-03-11 06:45:33 UTC
Created attachment 33957 [details]
correct rendering from rv280
Comment 2 Andrew Randrianasulu 2010-03-11 06:46:30 UTC
Created attachment 33958 [details]
wrong rendering from swrast
Comment 3 Andrew Randrianasulu 2010-03-11 06:47:45 UTC
Created attachment 33959 [details]
same glxinfo -l as in bug #27017
Comment 4 Andrew Randrianasulu 2010-03-11 06:51:20 UTC
Note - rendering directly to fb/vram is  currently way too slow on rv280/KMS

I can see for minute as demo draws itself. Composite manager speed it up .... 
Comment 5 Brian Paul 2010-03-11 07:13:51 UTC
This demo (and the others you've filed bugs against) work fine here with swrast built with 'make linux'.  With this configuration, glxinfo reports:

OpenGL vendor string: Brian Paul
OpenGL renderer string: Mesa X11
OpenGL version string: 2.1 Mesa 7.9-devel


It looks like the issue/bugs are in the libGL GLX code or something about the swrast_dri.so module.  I see the same thing here using a DRI build from a couple months ago.  I tried updating my DRI stuff just now but ran into a bunch of broken dependency issues.  I don't have right time to investigate any further.

If you need a reference sw driver, I suggest 'make linux'.
Comment 6 Brian Paul 2010-03-11 07:22:39 UTC
Looks like the issue is related to single buffer rendering.

Some of the tests take a -db flag to turn on double buffering.  point-wide looks OK with that flag.

Hopefully someone else can look into this.
Comment 7 Andrew Randrianasulu 2010-03-27 02:05:23 UTC
(In reply to comment #6)
> Looks like the issue is related to single buffer rendering.
> 
> Some of the tests take a -db flag to turn on double buffering.  point-wide
> looks OK with that flag.
> 
> Hopefully someone else can look into this.
> 

I think  this one is fixed, probably by 

commit 7996f0fc2c167c84552701be7a48d20a897e7978
    dri/swrast: Fix frontbuffer rendering.

-------

currently tested on x86, with mesa git master up to 
commit c8844c5549dec7e9c47ab49d6e229980f62da74d
    r300g: Make SWTCL clear work again.

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.