Bug 93987 - [SKL] Screens fail to go into standby
Summary: [SKL] Screens fail to go into standby
Status: CLOSED FIXED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: medium major
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-02-03 15:29 UTC by Norbert Varzariu
Modified: 2016-09-20 12:35 UTC (History)
1 user (show)

See Also:
i915 platform: SKL
i915 features:


Attachments
dmesg 4.4.1 (241.68 KB, text/plain)
2016-02-03 15:30 UTC, Norbert Varzariu
no flags Details
dmesg 4.3.5 without i915.disable_power_well=0 - crashing (4.03 MB, text/plain)
2016-02-03 16:34 UTC, Norbert Varzariu
no flags Details
dmesg 4.3.5 with disable_power_well - working (251.94 KB, text/plain)
2016-02-03 16:35 UTC, Norbert Varzariu
no flags Details

Description Norbert Varzariu 2016-02-03 15:29:34 UTC

    
Comment 1 Norbert Varzariu 2016-02-03 15:30:32 UTC
Created attachment 121492 [details]
dmesg 4.4.1
Comment 2 Norbert Varzariu 2016-02-03 15:32:08 UTC
Whenever my screen try and go into standby, they get switched back on directly, and the windows of my second screen are moved to my first screen.

This worked fine with 4.3.X kernels. On an earlier run of this kernel, I also got an Error about an atomic update failure. I will try and see if it pops back up again.
Comment 3 Norbert Varzariu 2016-02-03 16:33:00 UTC
Excuse me as my post was a little hasty. I'm running skylake.

xorg-x11-drv-intel-2.99.917-16.20150729.fc23.x86_64
mesa-dri-drivers-11.1.0-2.20151218.fc23.x86_64

I also noticed the 4.3.5 kernel only works if I set i915.disable_power_well=0. If I remove it, the screen go into standby, but won't wake up again.
Much like the bug I reported https://bugs.freedesktop.org/show_bug.cgi?id=92926 there.
I will attach a dmesg from a 4.3.5 kernel without power_well setting, which went into standby and didn't wake up, and a dmesg from a 4.3.5 kernel with power_well, which works as expected.
Comment 4 Norbert Varzariu 2016-02-03 16:34:07 UTC
Created attachment 121497 [details]
dmesg 4.3.5 without i915.disable_power_well=0 - crashing
Comment 5 Norbert Varzariu 2016-02-03 16:35:08 UTC
Created attachment 121498 [details]
dmesg 4.3.5 with disable_power_well - working
Comment 6 Jani Nikula 2016-04-25 09:55:46 UTC
Please try v4.6-rc5 or later.
Comment 7 Norbert Varzariu 2016-04-25 10:57:40 UTC
Seems to be working fine now. Do you need a dmesg anyway?

Thank you very much!
Comment 8 Jani Nikula 2016-04-25 12:11:01 UTC
(In reply to Norbert Varzariu from comment #7)
> Seems to be working fine now. Do you need a dmesg anyway?
> 
> Thank you very much!

Great, thanks for testing! No need for the dmesg.
Comment 9 Norbert Varzariu 2016-05-24 19:23:01 UTC
OK, the quirks are back. It worked for about a month now with 4.6.0-rc6, rc7 and the final release, but since yesterday the quirks are back. 
When the screen try and go into standby, they get switched right back on again. Also, sometimes the windows from my second screen all are moved to the first screen.

Please let me know what information you need or what steps I can take to help resolve this once and for all, hopefully.
Comment 10 Norbert Varzariu 2016-05-25 18:53:21 UTC
OK, even more weird: 
- with former working 4.3.5 kernel I get complete lockups when the screens have been in standby for a while:
May 25 14:15:29 desktop kernel: [drm:check_wm_state [i915]] *ERROR* mismatch in DDB state pipe B plane 1 (expected (0,0), found (446,884))
May 25 14:15:29 desktop kernel: [drm:check_wm_state [i915]] *ERROR* mismatch in DDB state pipe B cursor (expected (0,0), found (884,892))

- with kernels 4.6-rc6 and later, which used to work kind of fine until just a couple of days ago, it now fails consistently to go into standby.

- standby with those kernels as well as the latest fedora kernel works in GDM when not logged in, but not after login.
Comment 11 Norbert Varzariu 2016-05-26 14:29:52 UTC
4.6.0 works if only one screen is connected, I just noticed when I switched off my second screen from its power button.
Comment 12 Norbert Varzariu 2016-05-30 18:19:50 UTC
And it suddenly started working again today, out of the blue. 4.6.0.
Comment 13 Jani Nikula 2016-09-20 12:32:05 UTC
(In reply to Norbert Varzariu from comment #12)
> And it suddenly started working again today, out of the blue. 4.6.0.

Okay, we've also fixed a bunch of skl stuff for v4.7 and v4.8-rc. Please reopen if the problem comes back.


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.