Bug 202915

Summary: 5.0 black screen on xps 15 9570
Product: Drivers Reporter: Mark Weiman (mark.weiman)
Component: Video(Other)Assignee: drivers_video-other
Status: RESOLVED INVALID    
Severity: normal CC: ilpanich, kayu.guo, mark.weiman
Priority: P1    
Hardware: All   
OS: Linux   
Kernel Version: 5.0.1 Subsystem:
Regression: No Bisected commit-id:
Attachments: kernel log
patch

Description Mark Weiman 2019-03-13 23:19:17 UTC
Created attachment 281815 [details]
kernel log

On Arch Linux, when upgrading to 5.0 from 4.20, when the machine boots, the computer starts normally until a certain point where the screen goes black and never reappears. I've tried setting nomodeset with no success.

Attached is a kernel log where loglevel=6 and drm.debug=6 are set.
Comment 1 Mark Weiman 2019-03-14 14:09:14 UTC
Applying the patch mentioned here solves the issue.

https://bugs.archlinux.org/task/61964
Comment 2 Mark Weiman 2019-03-15 14:27:29 UTC
Not sure if anyone's looking at this, but what about adding something to the kernel to allow a parameter to force the slow and wide eDP configuration on machines with this issue?
Comment 3 Emanuele Panigati 2019-03-15 16:34:06 UTC
Same issue for me, both lts kernel and 4.20.13 works fine.
Comment 4 Kayu Guo 2019-03-16 01:41:08 UTC
Have same issue too
Comment 5 Emanuele Panigati 2019-03-16 14:14:27 UTC
(In reply to Mark Weiman from comment #1)
> Applying the patch mentioned here solves the issue.
> 
> https://bugs.archlinux.org/task/61964

I confirm that applying the patch solves the issue.

But since the patch mentioned seems to ignore the result of intel_dp_is_edp call... maybe a better solution can be found.
Comment 6 Mark Weiman 2019-03-16 14:20:36 UTC
The patch restores old behavior from 4.20 and before. I don't like the patch as is myself, but it does solve this problem. I am thinking of writing up a patch that changes this behavior to a parameter that can be passed on boot.
Comment 7 Emanuele Panigati 2019-03-16 14:38:21 UTC
(In reply to Mark Weiman from comment #6)
> The patch restores old behavior from 4.20 and before. I don't like the patch
> as is myself, but it does solve this problem. I am thinking of writing up a
> patch that changes this behavior to a parameter that can be passed on boot.

Sorry, I did not express myself well.
I totally agree with you: the patch solves the issue but I don't like the patch at all.

I hope that someone with a better understanding of the problematic function can solve it in a better way since I am not able to.
Comment 8 Mark Weiman 2019-03-18 14:37:19 UTC
Created attachment 281889 [details]
patch

Here is an untested patch that allows this to be fixed with a kernel parameter.
Comment 9 Mark Weiman 2019-03-18 18:10:38 UTC
Apparently, this is not the right place to open a bug for i915.