Bug 88620 - [BSW full-ppgtt] testdisplay shows abnormal
Summary: [BSW full-ppgtt] testdisplay shows abnormal
Status: CLOSED NOTOURBUG
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: unspecified
Hardware: Other All
: high major
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-01-20 10:22 UTC by Jeff Zheng
Modified: 2017-10-06 14:32 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
testdisplay -a output (2.15 MB, image/jpeg)
2015-01-20 10:22 UTC, Jeff Zheng
no flags Details
dmesg log (659.12 KB, text/plain)
2015-01-20 10:25 UTC, Jeff Zheng
no flags Details

Description Jeff Zheng 2015-01-20 10:22:27 UTC
Created attachment 112531 [details]
testdisplay -a output

==System Environment==
--------------------------
BIOS version V53
Regression: Yes. I tried _drm-intel-testing 2014_12_06 commit and testdisplay shows normal. 

Non-working platforms: BSW

==kernel==
--------------------------
-testing: 2015_01_18 commit 6dd5bb25ce0ed3a0012b69d553d74e1a0d39c313(test fails)


==Bug detailed description==
-----------------------------
Boot BSW, login and run testdisplay, eDP screen shows abnormal sometimes
testdisplay -a, or -d 24 will show abnormal. HDMI and DP screen also some similar abnormal behavior.

The reproducibility > 90%

==Reproduce steps==
---------------------------- 
1. Boot BSW, login and run testdisplay

==Bisect==
I tried to bisect, but unfortunately because of bug 87671, some commits cannot be bisected, in which commits testdisplay will show black screen.

Bad commit: 9025452366e9a86f2e11292865d98430341005d1 : Revert "drm/i915: Parsing LFP brightness control from VBT"
Good commit: d9d8e6b3c01386e5b39b70dc07156ff0e407d984: the commit before "drm/i915: Parsing LFP brightness control from VBT"
testdisplay will show black screen with commits between these two commit, including commit 371abae844ede392066bfc21202b2e40f4a654d1: "drm/i915: Parsing LFP brightness control from VBT"
Comment 1 Jeff Zheng 2015-01-20 10:25:40 UTC
Created attachment 112533 [details]
dmesg log
Comment 2 Jeff Zheng 2015-01-20 10:29:23 UTC
The abnormal output also include the flashes
Comment 3 Rodrigo Vivi 2015-01-26 21:41:22 UTC
This bisect is invalid. You will probably need to apply the revert patch on all bisect rounds in order to find the offending commit since without this revert you just get black screens after first backlight off...
Comment 4 Jeff Zheng 2015-01-27 04:08:52 UTC
After workaround the black screen issue, bisect shows following:

2f82bbdf3d4f1361c3d713c516d8aa390102374d is the first bad commit
commit 2f82bbdf3d4f1361c3d713c516d8aa390102374d
Author: Michel Thierry <michel.thierry@intel.com>
Date:   Mon Dec 15 14:58:00 2014 +0000

    drm/i915: Use true PPGTT in Gen8+ when execlists are enabled
Comment 5 Ville Syrjala 2015-01-28 13:46:35 UTC
Have you tried to disable 'dynamic memory self refresh' in the BIOS? It's the usual suspect for all display related issue these days.
Comment 6 Jeff Zheng 2015-01-29 01:31:13 UTC
After upgrade BIOS from V53 to V55, this issue disappears whatever disable/enable 'dynamic memory self refresh'
Comment 7 Jani Nikula 2015-01-29 07:59:01 UTC
(In reply to Jeff Zheng from comment #6)
> After upgrade BIOS from V53 to V55, this issue disappears whatever
> disable/enable 'dynamic memory self refresh'

Closing based on this. Please reopen if the problem reappears with upgraded bios.
Comment 8 Elizabeth 2017-10-06 14:32:09 UTC
Closing old verified.


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.