Radeon r600g and WebGL

[89c51]

With the last mesa update my system can't use WebGL anymore. Tested with both my main browser (FF) and Chromium and the result is the same. They weird thing is that people with the same card (6850) are using it fine.

Any ideas on how to track whats wrong.

glxinfo | grep -i opengl
OpenGL vendor string: X.Org
OpenGL renderer string: Gallium 0.4 on AMD BARTS
OpenGL core profile version string: 3.3 (Core Profile) Mesa 10.2.1
OpenGL core profile shading language version string: 3.30
OpenGL core profile context flags: (none)
OpenGL core profile profile mask: core profile
OpenGL core profile extensions:
OpenGL version string: 3.0 Mesa 10.2.1
OpenGL shading language version string: 1.30
OpenGL context flags: (none)
OpenGL extensions:

A message in the journal (no idea if it is related)

Jun 22 11:02:49 mainland kernel: traps: firefox[896] general protection ip:7f6a6a5dcb3f sp:7fff043936e0 error:0 in r600_dri.so[7f6a6a2d0000+4f2000]
Jun 22 11:38:28 mainland kernel: traps: firefox[1066] general protection ip:7fd9a73dcb3f sp:7fffab2b4ed0 error:0 in r600_dri.so[7fd9a70d0000+4f2000]
Jun 22 12:24:37 mainland kernel: traps: firefox[1423] general protection ip:7f9de4adcb3f sp:7fff5accea30 error:0 in r600_dri.so[7f9de47d0000+4f2000]
Jun 22 13:38:35 mainland kernel: traps: firefox[2282] general protection ip:7f7b968dcb3f sp:7fff2c2786a0 error:0 in r600_dri.so[7f7b965d0000+4f2000]
Jun 22 15:39:09 mainland kernel: traps: firefox[3411] general protection ip:7f53614dcb3f sp:7fff6ad13c80 error:0 in r600_dri.so[7f53611d0000+4f2000]

Linux mainland 3.15.1-1-ARCH #1 SMP PREEMPT Tue Jun 17 09:32:20 CEST 2014 x86_64 GNU/Linux

[89c51]

It's probably a firefox issue. If you force it to use WebGL it works. Something with blacklisting i suppose.

[89c51]

And to give an update just in case a dev wants to look into it the damn thing crushes with many of the chrome experiments when you force webgl and msaa.

On a slightly unrelated topic i won't even try to get HW acceleration to work with gstreamer. Not nice AMD. Not nice. /end rant.

[bridgman]

And to give an update just in case a dev wants to look into it the damn thing crushes with many of the chrome experiments when you force webgl and msaa.

On a slightly unrelated topic i won't even try to get HW acceleration to work with gstreamer. Not nice AMD. Not nice. /end rant.

Is there a connection between WebGL and gstreamer I'm missing ?

[89c51]

Is there a connection between WebGL and gstreamer I'm missing ?

No. As i said its an unrelated topic (i admit not slightly). It was a small rant. Don't take it seriously or something.

[BlackStar]

And to give an update just in case a dev wants to look into it the damn thing crushes with many of the chrome experiments when you force webgl and msaa.

And now you know why R600g is blacklisted in Firefox.

[89c51]

And now you know why R600g is blacklisted in Firefox.

Yes. At least i was expecting a note like "we blacklisted R600g" or something. It was working and it stopped just like that. I know i am using an "unsupported OS" (as in having a corporate entity to take care of things like that) and i can't expect everything to work but its 2014 and at least i'd like GL and HW accelerated video to work properly. Anyway.

[89c51]

However i have to say that the video problem is not an AMD one and i am an idiot.

[BlackStar]

Yes. At least i was expecting a note like "we blacklisted R600g" or something. It was working and it stopped just like that. I know i am using an "unsupported OS" (as in having a corporate entity to take care of things like that) and i can't expect everything to work but its 2014 and at least i'd like GL and HW accelerated video to work properly. Anyway.

Totally agreed.

But you still have it easy with your R600g and its fancy automatic performance switching and its video decoding and vsyncing. Try getting stuck with Nouveau for a year or two and you'll start to appreciate the small things more

Now, if only GPU switching worked on my rMBP then I could at least use the Intel drivers...

[curaga]

Is this the same system that was showing broken hw in another thread? If so, not mesa bug.