[Bug 75357] Barts (HD6850): Failure in evergreen_surface_check_2d

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=75357

--- Comment #11 from Michel Dänzer <michel at daenzer.net> ---
(In reply to Nix from comment #10)
> FWIW, I have just started seeing behaviour very similar to this, also with an
> HD6850, though it is probably a distinct bug because all the software
> versions involved are very different.

So it would probably be better to file your own report.


> This happened after closely-adjoined upgrades of the kernel (from 3.19 to
> 4.0.3), the X server (from 1.15.1, yes, I know, it's because of the security
> holes I finally upgraded, to 1.17.1), the ATI driver (from 7.3.0-26 to
> 7.5.0.68, all "git describe" output from the unmodified trunk), Mesa (from
> 10.2.6 to 10.5.5), and LLVM (from 3.3 to 3.5.2), so the fault could, alas, in
> theory lie in any or all of these.

It would be great if you could narrow that down a bit. The r600g driver doesn't
use LLVM, so that's probably not it. I agree it's most likely between Mesa or
the kernel, or maybe xf86-video-ati.


> Glamor is not in use.

Have you tried glamor to see if it happens with that as well?

Does

    Option "ColorTiling2D" "off"

in /etc/X11/xorg.conf or the environment variable R600_DEBUG=no2d (or even
R600_DEBUG=notiling) for chromium work around the problem?


> dpm is, and has been working flawlessly thank you very much, cutting my
> desktop's power consumption from 130W to 90W :) over the last year that's
> saved at least the price of the beer I owe you for the dpm work, Michel :)

Sounds great, but I haven't done any DPM work, I think that beer would have to
go to Alex. :)

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/dri-devel/attachments/20150526/5a15f8ee/attachment.html>