Bug 101343 - [SKL][PSR] Intermittent screen freezes with PSR enabled on XPS 9350
Summary: [SKL][PSR] Intermittent screen freezes with PSR enabled on XPS 9350
Status: CLOSED FIXED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: XOrg git
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Dhinakaran Pandiyan
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard: ReadyForDev
Keywords:
Depends on:
Blocks:
 
Reported: 2017-06-08 07:49 UTC by François Guerraz
Modified: 2018-08-15 16:32 UTC (History)
2 users (show)

See Also:
i915 platform: SKL
i915 features: display/PSR


Attachments
dmesg with drm-intel-nightly (740.35 KB, text/plain)
2017-06-12 22:46 UTC, François Guerraz
no flags Details
xrand output (682 bytes, text/plain)
2017-06-15 11:19 UTC, François Guerraz
no flags Details
intel_reg dump --all (27.79 KB, text/plain)
2017-06-15 11:19 UTC, François Guerraz
no flags Details

Description François Guerraz 2017-06-08 07:49:14 UTC
On a Dell XPS 13 9350 (i7-6560U, Iris Graphics 540), i915.enable_psr=1 causes intermittent randomly spaced screen freezes that last for a couple of seconds.

Setting PSR=2 or 0 fixes the problem. Disabling FBC and running wayland both reduce the frequency of theses freezes.
Comment 1 François Guerraz 2017-06-08 07:51:32 UTC
I forgot to mention I'm running 4.12.0-rc4-mainline on arch. Problem has been present in all kernel versions I've tried before that.
Comment 2 Elizabeth 2017-06-08 19:50:11 UTC
In order to provide more assistance please include the logs.

For further information please refer to:
https://01.org/linuxgraphics/documentation/bugs-and-debugging

Once you provide the information needed please change the status of the bug to REOPEN.
Comment 3 François Guerraz 2017-06-08 22:23:32 UTC
I have tried to find relevant log entries but couldn't find anything, so if you don't tell me more precisely what you want, I will not reopen this bug.
Comment 4 Elizabeth 2017-06-09 17:24:02 UTC
(In reply to François Guerraz from comment #1)
> I forgot to mention I'm running 4.12.0-rc4-mainline on arch. Problem has
> been present in all kernel versions I've tried before that.

Could you add your hardware and software information?
You mentioned you had tried in others kernels. Could you provide the kernel versions? Also could you attach the dmesg and kern.log, and try with this kernel version https://cgit.freedesktop.org/drm-tip
Comment 5 François Guerraz 2017-06-12 22:46:36 UTC
Created attachment 131908 [details]
dmesg with drm-intel-nightly

Yes, the issue is present with the latest drm-intel git version.

I've had the issue will all kernel version since I use this laptop (kernel 4.7 to now).

As mentioned in previous comment, the hardware is a Dell XPS 13 9350 (i7-6560U, Iris Graphics 540) with QHD screen, software is up to date Arch Linux with Gnome under wayland or X11.
Comment 6 François Guerraz 2017-06-15 11:19:28 UTC
Created attachment 131974 [details]
xrand output
Comment 7 François Guerraz 2017-06-15 11:19:51 UTC
Created attachment 131975 [details]
intel_reg dump --all
Comment 8 François Guerraz 2017-06-15 11:20:45 UTC
Is there any other information you need?
Comment 9 Elizabeth 2017-06-15 15:40:08 UTC
(In reply to François Guerraz from comment #8)
> Is there any other information you need?

Hello and thanks for adding the information.

Adding tag into "Whiteboard" field - ReadyForDev
*Status is correct
*Platform is included
*Feature is included
*Priority and Severity correctly set
*Logs included
Comment 10 François Guerraz 2017-06-20 13:14:12 UTC
Also, interestingly, the bug doesn't appear immediately, it takes a few minutes of usage to trigger and then it becomes more and more frequent and then unusable (due graphics memory usage increase?).
Comment 11 Jani Nikula 2017-06-30 11:20:33 UTC
So what happens if you leave out i915.enable_psr parameter completely, and use defaults?
Comment 12 François Guerraz 2017-06-30 13:04:37 UTC
(In reply to Jani Nikula from comment #11)
> So what happens if you leave out i915.enable_psr parameter completely, and
> use defaults?

Power consumption goes up by 0.5/0.7W but no screen freeze.
In general, power usage is quite high on this system (almost twice as high as the model with the FHD panel) so I've tried to enable as many power saving features as I could.
Comment 13 Jani Saarinen 2018-03-29 07:10:32 UTC
First of all. Sorry about spam.
This is mass update for our bugs. 

Sorry if you feel this annoying but with this trying to understand if bug still valid or not.
If bug investigation still in progress, please ignore this and I apologize!

If you think this is not anymore valid, please comment to the bug that can be closed.
If you haven't tested with our latest pre-upstream tree(drm-tip), can you do that also to see if issue is valid there still and if you cannot see issue there, please comment to the bug.
Comment 14 Jani Saarinen 2018-04-23 10:09:28 UTC
DK, any commments from you on this? 
Note on CI fi-skl-6600u is the same model. 
https://intel-gfx-ci.01.org/tree/drm-tip/igt@kms_psr_sink_crc@psr_basic.html
Comment 15 Dhinakaran Pandiyan 2018-04-23 17:23:59 UTC
This is a good candidate for retesting, we've had several PSR patches to address screen freezes since 2017-06-30.

François,

Can you please retest with drm-tip and report back the results?
Comment 16 Jani Saarinen 2018-04-23 19:03:57 UTC
Lets mark this resolved and if no response in a month then we just closed as works for me.
Comment 17 François Guerraz 2018-04-24 08:02:59 UTC
I certainly will test.
I'm still experiencing problems with the latest stable kernel (4.16.3) but I think it only happens with IOMMU enabled and appears more frequently on X than Wayland.

I'll give the drm-tip a shot.
Comment 18 Jani Saarinen 2018-05-04 07:49:56 UTC
Any updates on testing this?
Comment 19 François Guerraz 2018-05-04 10:23:16 UTC
Yes, I think the problem is fixed.

However I'm still having issues when IOMMU is enabled, but this is another bug.
(getting "*ERROR* uncleared fifo underrun on pipe A" and screen flashes and graphic corruption occurs for a few seconds, randomly)
Comment 20 Jani Saarinen 2018-05-04 12:04:12 UTC
OK, There are already some mapped to Meta bug for same, can you find yours from these bugs under https://bugs.freedesktop.org/show_bug.cgi?id=105980?
Comment 21 Dhinakaran Pandiyan 2018-05-08 20:01:26 UTC
(In reply to François Guerraz from comment #19)
> Yes, I think the problem is fixed.
> 

Great. Please feel free to report any new issues you notice with PSR.
Comment 22 Baybal 2018-07-26 12:18:58 UTC
Dell 9370 - KBL-R

Linux xps 4.17.9-gentoo #7 SMP PREEMPT Wed Jul 25 14:02:26 +06 2018 x86_64 Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz GenuineIntel GNU/Linux

psr=1 periodic screen freezes

When the last batch of screen freeze patches landed in kernel? Did they make it to 4.17?
Comment 23 Dhinakaran Pandiyan 2018-07-26 16:28:35 UTC
Most of the work is going to land in 4.18.

Would you be able to test drm-tip (https://anongit.freedesktop.org/git/drm-tip.git)?

Can you also attach dmesg booting the kernel with drm.debug=14?
Do you use a compositor?
Can you also post the output of /sys/kernel/debug/dri/0/i915_edp_psr_status?
How about edid-decode /sys/class/drm/card0-<eDP>/edid ?
Comment 24 François Guerraz 2018-07-26 18:49:16 UTC
FYI, the issue doesn't happen any more for me on recent drm-tip but is still present with mainline kernels.
Comment 25 Baybal 2018-07-31 11:59:59 UTC
@(In reply to Dhinakaran Pandiyan from comment #23)
> Most of the work is going to land in 4.18.
> 
> Would you be able to test drm-tip
> (https://anongit.freedesktop.org/git/drm-tip.git)?
> 
> Can you also attach dmesg booting the kernel with drm.debug=14?
> Do you use a compositor?
> Can you also post the output of /sys/kernel/debug/dri/0/i915_edp_psr_status?
> How about edid-decode /sys/class/drm/card0-<eDP>/edid ?

I will test it this weekend
Comment 26 Baybal 2018-08-15 15:34:01 UTC
(In reply to Dhinakaran Pandiyan from comment #23)
> Most of the work is going to land in 4.18.
> 
> Would you be able to test drm-tip
> (https://anongit.freedesktop.org/git/drm-tip.git)?
> 
> Can you also attach dmesg booting the kernel with drm.debug=14?
> Do you use a compositor?
> Can you also post the output of /sys/kernel/debug/dri/0/i915_edp_psr_status?
> How about edid-decode /sys/class/drm/card0-<eDP>/edid ?

I confirm, PSR issues are gone on 4.18. Though I had issues booting 4.18 due to unrelated NVME bug.
Comment 27 Dhinakaran Pandiyan 2018-08-15 16:32:42 UTC
(In reply to Baybal from comment #26)
> (In reply to Dhinakaran Pandiyan from comment #23)
> > Most of the work is going to land in 4.18.
> > 
> > Would you be able to test drm-tip
> > (https://anongit.freedesktop.org/git/drm-tip.git)?
> > 
> > Can you also attach dmesg booting the kernel with drm.debug=14?
> > Do you use a compositor?
> > Can you also post the output of /sys/kernel/debug/dri/0/i915_edp_psr_status?
> > How about edid-decode /sys/class/drm/card0-<eDP>/edid ?
> 
> I confirm, PSR issues are gone on 4.18. Though I had issues booting 4.18 due
> to unrelated NVME bug.

Thanks for the update.

-DK


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.