Bug 101380 - [BDW] Freezes, whiteflash and flickering on kernel 4.11 and 4.12 Intel HD Graphics 5500
Summary: [BDW] Freezes, whiteflash and flickering on kernel 4.11 and 4.12 Intel HD Gra...
Status: CLOSED INVALID
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: 2017-06-11 13:47 UTC by Mark Spencer
Modified: 2017-09-18 08:53 UTC (History)
1 user (show)

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


Attachments
dmesg (1.17 MB, text/plain)
2017-06-11 13:47 UTC, Mark Spencer
no flags Details
dmidecode (12.53 KB, text/plain)
2017-06-11 13:49 UTC, Mark Spencer
no flags Details

Description Mark Spencer 2017-06-11 13:47:40 UTC
Created attachment 131875 [details]
dmesg

On 4.11 kernel I get very strange issues after start from cold boot.

When I log in and my graphic session is established after couple seconds my screen freezes and starts whitening then returns to normal and repeats freezes/whitening for few minutes. (I'm unable to record this behavior).

Meanwhile my screen starts HEAVY flickering (refresh rate is extremely low) which stops after about 10-15 minutes. This flickering doesn't stop even when I reboot my machine or reboot and start windows 10 session. (those issues don't' happen in windows independently). After 10-15 minutes everything is back to normal and I don't see anything suspicious.

I remember that I saw this weird thing in linux 4.7 (4.8 and 4.9 were ok, didn't test 4.10). I tested boot with following options:

i915.enable_psr=1
i915.enable_psr=0
i915.enable_fbc=1
i915.enable_fbc=0

and they didn't help. I also tested this with linux 4.12.rc4.

It happens only if I start system from cold. Restarting it after the time issues are gone doesn't trigger them again. Also it happens when I suspend my machine and resume it long after. 

My system:

HP Spectre x360 Convertible 13
Desktop: KDE Plasma 5.10.1
Display Server: x11 (X.Org 1.19.3)
driver: modesetting
GLX Renderer: Mesa DRI Intel HD Graphics 5500 (Broadwell GT2)
GLX Version: 3.0 Mesa 17.1.2
Comment 1 Mark Spencer 2017-06-11 13:49:03 UTC
Created attachment 131876 [details]
dmidecode
Comment 2 Mark Spencer 2017-06-12 12:33:27 UTC
I recorded example video with this issue.

https://www.dropbox.com/s/4ffyk8ddtht2w10/example.mp4
Comment 3 Elizabeth 2017-06-13 21:39:26 UTC
Hello Mark. 
Right now there are a lot of variables that could be causing this problem. To make sure that is possible to reproduce the issue, it'll be needed to apply the following steps: 
First, could you please add the parameter in the grub "drm.debug=0xe", and after that get the dmesg again.
Second, the best way to find out what is causing this problem is, if possible, to return your configuration to default, and try to reproduce de issue. If the issue is no present there, install the newest kernel and try again. The idea is to add every variable till the bug is present again, and please be sure to get a dmesg with the drm.debug=0xe parameter so it can be compared with the one where you have the issue. If you have any problem feel free to comment it here.
Thank you.
Comment 4 Elizabeth 2017-06-14 19:21:29 UTC
I forget to mention once you provide more information, please change the tag from "NEEDINFO" to "REOPEN". Thank you.
Comment 5 Ricardo 2017-07-21 16:24:06 UTC
based on the lack of response from the submitter the bug will be closed. If the bug is reproduced with the latest configuration and adding the parameters make sure you add the logs create a new bug
Comment 6 Nick Coghlan 2017-09-14 01:13:10 UTC
I've been seeing problems with my HP Spectre x360 that sound very similar to those reported by the OP. They've been intermittent over the past couple of years, and are very much still present with the current Fedora 26 kernel (kernel-4.12.11-300.fc26.x86_64)

The only custom boot settings I currently have configured are the ones needed to work around a problem with the device otherwise not emitting any sound (from https://bugzilla.redhat.com/show_bug.cgi?id=1220709#c10):

    acpi_backlight=vendor acpi_osi='!Windows 2013' acpi_osi='!Windows 2012'

I've added the drm.debug=0xe setting to my boot configuration, and the problem fortunately doesn't affect attached external monitors, so next time it happens I'll be able to capture the dmesg output.

On that front:

1. Is there a useful filter I can apply to the dmesg output to limit it to output that's likely to be relevant? Or should I just upload the whole thing?

2. Should I open a new issue with those logs, or reopen this one?
Comment 7 Elizabeth 2017-09-14 16:11:59 UTC
(In reply to Nick Coghlan from comment #6)
> ...
> On that front:
> 
> 1. Is there a useful filter I can apply to the dmesg output to limit it to
> output that's likely to be relevant? Or should I just upload the whole thing?
> 
> 2. Should I open a new issue with those logs, or reopen this one?
Hello Nick, 
drm.debug=0xe would be enough for now, you may want to add log_bug_len=2M (or bigger) in case more space for dmesg is needed. And yes, please file a new bug with HW and SW information, logs, and any other relevant information.
Thank you.
Comment 8 Nick Coghlan 2017-09-14 22:17:35 UTC
Thanks, new bug filed as BZ#102765
Comment 9 Jani Nikula 2017-09-18 08:53:32 UTC
(In reply to Nick Coghlan from comment #8)
> Thanks, new bug filed as BZ#102765

Bug #102765.


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.