Bug 104820

Created attachment 137008 [details]
Xorg.0.log

Hello, thank you for this feedback. Your computer is using software rendering fallback instead of hardware acceleration. This is because no path could be found. This appears to be a common theme across many threads. Padoka is labeled in Ubuntu PPA as beta software. These guys have claimed to find a resolution procedure. https://www.phoronix.com/forums/forum/linux-graphics-x-org-drivers/open-source-amd-linux/913600-recommended-install-process-for-xubuntu-16-10-rx480-radv-vulkan 

This happens because the default xv acceleration is no longer available when using AIGLX (because everything is being drawn through textures on the 3D unit). I haven't seen this problem on NVIDIA cards using the more recent binary drivers (they seem to still have an accelerated xv) but the open source ATI drivers have a similar problem (very buggy xv when using AIGLX) as do the Intel drivers. AIGLX was only used in the short period of time when open drivers weren't capable of redirected direct rendering. DRI2 then made that possible, making AIGLX obsolete.

Therefore I am labeling this error as incomparable video drivers and marking this issue as resolved.

(In reply to Robert Wahlstedt from comment #2)
> Hello, thank you for this feedback. Your computer is using software
> rendering fallback instead of hardware acceleration. This is because no path
> could be found. This appears to be a common theme across many threads.
> Padoka is labeled in Ubuntu PPA as beta software. These guys have claimed to
> find a resolution procedure.
> https://www.phoronix.com/forums/forum/linux-graphics-x-org-drivers/open-
> source-amd-linux/913600-recommended-install-process-for-xubuntu-16-10-rx480-
> radv-vulkan 
> 
> This happens because the default xv acceleration is no longer available when
> using AIGLX (because everything is being drawn through textures on the 3D
> unit). I haven't seen this problem on NVIDIA cards using the more recent
> binary drivers (they seem to still have an accelerated xv) but the open
> source ATI drivers have a similar problem (very buggy xv when using AIGLX)
> as do the Intel drivers. AIGLX was only used in the short period of time
> when open drivers weren't capable of redirected direct rendering. DRI2 then
> made that possible, making AIGLX obsolete.
> 
> Therefore I am labeling this error as incomparable video drivers and marking
> this issue as resolved.

Robert, can you explain how did you come to that conclusion? Xorg.0.log attachment shows that glamor is used, DRI driver loads fine and according to glxinfo HW renderer is being used. AFAICT this bug has nothing to do with xv acceleration or with radv vulkan. In fact I cannot see anything that would prove that this is not our bug.

Blank white screen with SDDM mentioned also in bug #104806 but there the gpu used is Radeon while here it's Intel.

Fixed in git180129151100.0347a83~x~padoka0, there is still wrong color on taskbar, I will fill separate bug report.

Krzysztof

(In reply to krzysiek from comment #5)
> Fixed in git180129151100.0347a83~x~padoka0, there is still wrong color on
> taskbar, I will fill separate bug report.
> 
> Krzysztof

FYI If you mean 'too dark', this "wrong color" may be a duplicate for bug #103655 that has been fixed. There was other bug #102806 with similar symptoms as well.

If Padoka repos update also X server, there was X server bug 103000 that affected titlebar colors (and was recently fixed).

Hi,

Recent update of mesa broke thinks again,

git180129151100.0347a83 works fine
git180208124100.e843667 white screen again

Please let me know if you need more info.

Krzysztof

I'm also seeing this on a AMD Kaveri IGP box.

Reopening old bugs with "me too" doesn't help much, do file new ones.

Please work with your packager/distro tracking down the component you have based on the current suggestions.

Apparently this is a bug in Qt: https://bugreports.qt.io/browse/QTBUG-66348

Workaround is to delete $HOME/.cache/qtshadercache