Bug 89631 - [BSW] DP screen is black after resuming from S3/S4, if only DP connected
Summary: [BSW] DP screen is black after resuming from S3/S4, if only DP connected
Status: CLOSED NOTOURBUG
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: unspecified
Hardware: Other All
: highest major
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-03-18 06:40 UTC by Jeff Zheng
Modified: 2016-10-13 07:46 UTC (History)
4 users (show)

See Also:
i915 platform:
i915 features:


Attachments
dmesg log (286.22 KB, text/plain)
2015-03-18 06:40 UTC, Jeff Zheng
no flags Details

Description Jeff Zheng 2015-03-18 06:40:39 UTC
Created attachment 114421 [details]
dmesg log

==System Environment==
--------------------------
HDMI monitor: ASUS PA238Q
BIOS: V59 
Regression: No. Tried on drm-intel-testing-2015-01-30 and still saw this issue. This might be related with processor and BIOS upgrade, since I tested before and didn't find this issue at that time.

Non-working platforms: BSW

==kernel==
--------------------------
-testing: drm-intel-testing-2015-03-13 (fails)

==Bug detailed description==
-----------------------------
When connected ASUS PA238Q monitor with DP cable and un-plug eDP cable to keep DP to be the only connection, DP screen will be black screen after resuming from S3/S4.



==Reproduce steps==
---------------------------- 
1. Connected ASUS PA238Q with DP cable, un-plug eDP cable and boot
2. echo mem > /sys/power/state
3. Resume by pressing power button
Comment 1 Jeff Zheng 2015-03-31 06:21:09 UTC
exists on drm-intel-testing-2015-03-27
Comment 2 Gavin Hindman 2015-04-13 15:29:57 UTC
Is this still present on drm-intel-testing-2015-04-10?
Comment 3 Jeff Zheng 2015-04-14 02:21:18 UTC
Yes this issue is still present on drm-intel-testing-2015-04-10. 

This time I used Dell U2311Hb DP monitor to workaround bug 90006. After resuming, DP screen is black. I tried to hotplug, power off/on DP monitor, and run "testdisplay -i", but the screen is still black. "testdisplay -a" will make the screen recover.

I also tried S3 after X is started, the screen is black too after resuming. I can recover the screen by running 'xrandr --output DP1 --set "Broadcast RGB" Automatic/Full' to switch "Broadcast RGB" property.
Comment 4 Jeff Zheng 2015-04-14 03:15:05 UTC
I tried eDP + DP connected, S3 works on both console/X mode.

Under X mode, I also tried disabling eDP by "xrandr --output eDP1 --off", and DP screen works after system resumes from S3
Comment 5 Gordon Jin 2015-04-14 06:35:09 UTC
(In reply to Jeff Zheng from comment #4)
> Under X mode, I also tried disabling eDP by "xrandr --output eDP1 --off",
> and DP screen works after system resumes from S3

As this works, I'm wondering if we still want to treat this bug (with so special config -- no eDP) as high priority, or even won't fix?
Comment 6 David Weinehall 2015-04-14 14:51:19 UTC
I've tried to reproduce this, but failed; I've tested with eDP, DP, eDP + DP, HDMI->DVI, and eDP + HDMI->DVI. All of these successfully resume *if* they manage to suspend (suspend fails if attempted again right after resume).

I'm using BIOS v58 for this though (and another monitor -- if you think that matters I can try to track down the same model); would it be possible for you to test with a downgraded BIOS -- either 56 or 58 should work, though 56 has some performance issues.
Comment 7 Gavin Hindman 2015-04-14 15:04:57 UTC
You need to have at least BIOS 61.3 and latest stepping
Comment 8 Jeff Zheng 2015-04-15 06:43:16 UTC
Yes. I used latest stepping. I tried another Dell U2311Hb DP monitor and saw the same issue. You might need to make sure:
1. Only DP connection, I unplugged the eDP cable.
2. Resume after the monitor power led color changes from white to orange. 

My BIOS version was V59.

I have upgraded BIOS to V65, but my DUT is unable to get latest kernel because of local server room clear up. I will test and comment after local server recovers.
Comment 9 Jeff Zheng 2015-04-15 07:12:37 UTC
Unfortunately with V65 BIOS, system hang during resuming from S3. So V59 BIOS looks to be a better choice.
Comment 10 David Weinehall 2015-04-15 11:22:49 UTC
I've now tested the following setups:

* eDP cable disconnected from motherboard

Display:
1. Dell U2412M
2. HP ZR24w

(both behaved similarly)

Bios: v56
Bios: v58

Wakeup source:
1. power button
2. USB
3. rtcwake

Environment:
1. X
2. Command line

I'll try to install Bios v59 to see if that makes a difference.
Comment 11 David Weinehall 2015-04-15 12:36:41 UTC
I've now tested also with BIOS 59; no different behaviour observed -- I still cannot reproduce this bug.

Do I understand correctly that you both tried from command line and X? If so I cannot think of any differences apart from the displays in our setup, and seeing as we both tried two different models (one being the same brand) it seems unlikely to be the cause of this.

Your HW is a C0, right?
Comment 12 Gavin Hindman 2015-04-15 14:10:48 UTC
Once again - you need to have at least BIOS 61.3, or you don't have the right PUNIT FW in BIOS.  Testing on anything else is wasting cycles.
Comment 13 David Weinehall 2015-04-16 07:56:53 UTC
I've tested with v61.4 too, without issues.

One thing: the bug title mentions S3/S4, but the bug reproduction steps only involves passing "mem" to /sys/power/state, which means only S3 is tested.

Is the bug title wrong, or is there a missing step somewhere?

Also, I'd still need info on whether you can reproduce this from the command line, or if it's only in X (or, for that matter, vice versa)?
Comment 14 Jeff Zheng 2015-04-16 07:59:50 UTC
This morning, I tested on another machine, and didn't see this issue.

I upgraded to V65 on original machine, but S3 will hang system. 

So I upgrade my test machine from V59 to V61.3, this issue does not exist.

I think we can close this bug.
Comment 15 Jeff Zheng 2015-04-16 08:13:59 UTC
(In reply to David Weinehall from comment #13)
> I've tested with v61.4 too, without issues.
> 
> One thing: the bug title mentions S3/S4, but the bug reproduction steps only
> involves passing "mem" to /sys/power/state, which means only S3 is tested.
> 
> Is the bug title wrong, or is there a missing step somewhere?

If using V59, both S3 and S4 will cause this issue.
> 
> Also, I'd still need info on whether you can reproduce this from the command
> line, or if it's only in X (or, for that matter, vice versa)?
If using V59, both X and command line will cause this issue.

This is just FYI. with V61.3, I don't see this issue.
Comment 16 Gordon Jin 2015-04-16 11:08:59 UTC
closing, since this issue gone with BIOS v61.3.
Comment 17 Jari Tahvanainen 2016-10-13 07:46:47 UTC
Closing verified+notourbug.


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.