This is printed in mass quantities: EE r600_pipe.c:80 r600_create_fence - r600: too many concurrent fences They don't cause any visible rendering error though.
If this is a regression can you bisect? Please also attach your xorg log and dmesg output.
Yes, it is a regression. I'll bisect later.
Created attachment 72396 [details] dmesg
Created attachment 72397 [details] Xorg.0.log
There are only 'skip'ped commits left to test. The first bad commit could be any of: 4003961fbf7e59de4b95448eada569adb91e5852 4a486f8bf2ca3d88228f8313282289abe78bc2f8 3039addf9363dbae1bbad1cd75d32e0013038c08 We cannot bisect more! With 4003961f and 4a486f8b it segfaults immediately, since 3039addf9 it prints the errors.
Marek, any ideas? I'm also seeing this with radeonsi.
The error messages disappeared with a fresh checkout, so I believe commit 4f0563a6584dd83462d5cf35a7a1ebc4a4d6378f works. However, neverball still flashes garbage when I start it with __GL_FSAA_MODE=4 or any other number, so that was not caused by this problem.
Now this problem seems to be gone. Closing.
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.