Bug 108266 - Screen flashing and dmesg flooded with *ERROR* [CONNECTOR:65:eDP-1] Link Training failed at link rate = 270000, lane count = 2 when psr enabled
Summary: Screen flashing and dmesg flooded with *ERROR* [CONNECTOR:65:eDP-1] Link Trai...
Status: RESOLVED MOVED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: low enhancement
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard: Triaged, ReadyForDev
Keywords:
Depends on:
Blocks:
 
Reported: 2018-10-07 16:21 UTC by kitestramuort
Modified: 2019-11-29 17:55 UTC (History)
3 users (show)

See Also:
i915 platform: BDW
i915 features: display/backlight, display/eDP, display/PSR, power/runtime PM, power/suspend-resume


Attachments
dmesg output (503.45 KB, text/x-log)
2018-10-07 16:21 UTC, kitestramuort
no flags Details
dmesg log (656.90 KB, text/x-log)
2018-10-07 18:51 UTC, kitestramuort
no flags Details
screen edid (826 bytes, text/plain)
2018-10-08 19:54 UTC, kitestramuort
no flags Details
psr-status (169 bytes, text/plain)
2018-10-08 19:54 UTC, kitestramuort
no flags Details
dmesg log (drm-tip + patches) (877.21 KB, text/x-log)
2018-10-09 20:55 UTC, kitestramuort
no flags Details
log of screen locking after blanking (1.28 MB, text/x-log)
2018-10-09 21:31 UTC, kitestramuort
no flags Details

Description kitestramuort 2018-10-07 16:21:45 UTC
Created attachment 141929 [details]
dmesg output

If I enable psr with i915.enable_psr=1 the syslog is flooded by the message below and the screen blanks intermittently for a second or so.
 
*ERROR* [CONNECTOR:65:eDP-1] Link Training failed at link rate = 270000, lane count = 2

No such message is logged when psr is disabled and no flashing occurs.

This is on a Thinkpad T450s, Core i5-5300U, kernel 4.19.0-rc6, Arch Linux.
Comment 1 kitestramuort 2018-10-07 18:51:49 UTC
Created attachment 141930 [details]
dmesg log
Comment 2 Dhinakaran Pandiyan 2018-10-08 18:02:00 UTC
PSR on BDW isn't supported currently.

But, here's something you can try on top of drm-tip (https://cgit.freedesktop.org/drm-tip) https://patchwork.freedesktop.org/series/50654/


As documentation for future BDW enablement, it'd be great if you also attach the output of 
edid-decode /sys/class/drm/card0-eDP-1/edid and /sys/kernel/debug/dri/0/i915_edp_psr_status

-DK
Comment 3 kitestramuort 2018-10-08 19:54:07 UTC
Created attachment 141941 [details]
screen edid
Comment 4 kitestramuort 2018-10-08 19:54:32 UTC
Created attachment 141942 [details]
psr-status
Comment 5 kitestramuort 2018-10-09 07:52:51 UTC
With drm-tip and your patches the errors are gone, but the flashing still occurs albeit very occasionally. I guess PSR is still inactive on Broadwell though. 

Do you know if this platform is supposed to reach PC10? All I could get to is PC7, maybe due to lack of PSR support?
Comment 6 Jose Roberto de Souza 2018-10-09 17:28:44 UTC
Hi
Could you attach dmesg logs when running with the drm-tip & patches? Please don't forget to increase the log level(drm.debug=0x1e).

Thanks
Comment 7 kitestramuort 2018-10-09 20:55:59 UTC
Created attachment 141964 [details]
dmesg log (drm-tip + patches)
Comment 8 Jose Roberto de Souza 2018-10-09 21:05:10 UTC
There is a lot of short pulses from your eDP but no PSR or other errors detected.
Before you dump this log did you got any flashing?
Comment 9 kitestramuort 2018-10-09 21:14:54 UTC
no flashing before this dump. There is some flashing after the first time the display comes back from blanking for inactivity.  I'll try and capture a new dmesg when flashing occurs
Comment 10 kitestramuort 2018-10-09 21:30:04 UTC
So, this time the screen didn't come back from blanking. I had to force reboot. journalctl -b -1 attached. There's an *ERROR* [CRTC:39:pipe A] flip_done timed out
Comment 11 kitestramuort 2018-10-09 21:31:00 UTC
Created attachment 141965 [details]
log of screen locking after blanking
Comment 12 Jose Roberto de Souza 2018-10-16 18:01:58 UTC
The problem is '*ERROR* [CRTC:39:pipe A] flip_done timed out' but there is no clue of what caused that in your logs.

Try to reproduce with this kernel parameters: drm.debug=0x1e log_buf_len=4M
Comment 13 Jose Roberto de Souza 2018-10-16 19:31:16 UTC
Hum this may not be 100% related to PSR and there is reports of flip_done timeout with PSR disabled too in BDW:

https://bugs.freedesktop.org/show_bug.cgi?id=107732
Comment 14 Jani Saarinen 2019-11-26 15:53:16 UTC
You are reporter of the issue currently having low priority. Do you still see issue. If so, please spesify clearly what is impact to you.
Comment 15 Martin Peres 2019-11-29 17:55:31 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/drm/intel/issues/170.


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.