Bug 91694

Summary: [SKL] Low performance of rendering on UXA
Product: DRI Reporter: Janusz <januszmk6>
Component: DRM/IntelAssignee: Intel GFX Bugs mailing list <intel-gfx-bugs>
Status: CLOSED FIXED QA Contact: Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity: normal    
Priority: medium CC: intel-gfx-bugs
Version: XOrg git   
Hardware: Other   
OS: All   
Whiteboard:
i915 platform: SKL i915 features: GEM/Other
Attachments:
Description Flags
dmesg
none
xorg log
none
image of corruption
none
image of corruption 2 none

Description Janusz 2015-08-20 00:15:27 UTC
Created attachment 117796 [details]
dmesg

HD Graphics 530, when I am switching to full screen window on i3wm and back, it takes long time, like its on heavy load. When I am watching movie and doing someting else on other monitor, it skips frames until operation is done - mostly when I am switching to full screen, also on slim, also . It doesn't happen on SNA or BLT but  when I switch into SNA or BLT, it doesn't render properly my movie, when I am switching for example to full screen on my player, only player controls goes into full screen, the view of the movie is staying at the same resolution it was. Also on SNA and BLT  I get some flickers on when I am typing login into slim manager. 
On UXA also switching tabs on firefox takes lot of time. 
The similar problems I am experiencing on KDE.

kernel version drm-intel-next from 2015-08-18, xf86-video-intel-2.99.917-r2
Comment 1 Janusz 2015-08-20 00:16:47 UTC
Created attachment 117797 [details]
xorg log
Comment 2 Chris Wilson 2015-08-20 06:01:14 UTC
Yes, UXA will be slow and SNA runs the risk of bugs in with greater complexity to avoid that slowness. Your hardware doesn't support scaling Xv in hardware (that requires a GPU shader instead). As to the question of why the glyph rendering is corrupt is either in the reading back or writing of the results into the scanout, try AccelMethod "none".
Comment 3 Janusz 2015-08-20 15:55:41 UTC
(In reply to Chris Wilson from comment #2)
> Yes, UXA will be slow and SNA runs the risk of bugs in with greater
> complexity to avoid that slowness. Your hardware doesn't support scaling Xv
> in hardware (that requires a GPU shader instead). As to the question of why
> the glyph rendering is corrupt is either in the reading back or writing of
> the results into the scanout, try AccelMethod "none".

Thanks for comment, but AccelMethod "none" didn't helped, with "none" its getting even worse with slowness, and corruption, I will add some photos now.
If occasionally I get black screen on my monitor 2560x1440 for few seconds, should I report another bug?
Comment 4 Janusz 2015-08-20 15:56:21 UTC
Created attachment 117816 [details]
image of corruption
Comment 5 Janusz 2015-08-20 15:56:41 UTC
Created attachment 117817 [details]
image of corruption 2
Comment 6 Jari Tahvanainen 2017-03-08 08:49:07 UTC
Janusz, sorry about this long lead time for getting to you. Is this still problem for you and if yes then can you please get us the latest data with newest drm-tip?
Comment 7 Jari Tahvanainen 2017-04-10 11:44:00 UTC
Timeout - assuming resolved+fixed. If problem still persist with the latest kernels (preferable drm-tip from git://anongit.freedesktop.org/git/drm-tip), reopen this bug with latest logs as attachments.

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.