Discussion:
Bug#944350: systemd: screen remains off after waking up from suspend
Add Reply
Michael Biebl
2019-11-08 14:00:01 UTC
Reply
Permalink
Control: tags -1 + moreinfo
Package: systemd
Version: 242-7
Severity: normal
Dear Maintainer,
Thinkpad X1 Extreme gen2 with OLED.
I am not able to find anything in logs. Can you perhaps let me know how I can
try to debug this problem?
This sounds like a kernel issue to me.
What made you decide to file this against systemd?
--
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?
Michael Biebl
2019-11-08 15:00:02 UTC
Reply
Permalink
I am not sure whether this is being handled by systemd or kernel. Please
change it to kernel is you think its kernel problem.
I am trying to use pm-suspend to debug it, so i might provide more info,
but so far I dont see any errors. The machine resume from suspend fine,
just the screen is black. External monitor works fine.
So you have the same issue with pm-suspend or not?

If the machine itself is running, can you reach it via SSH?

Have you tried changing the brightness after resume? Maybe it get's
reset to 0.
--
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?
Michael Biebl
2019-11-08 15:30:02 UTC
Reply
Permalink
Control: reassign -1 src:linux
Control: found -1 5.3.7-1
Hi Michael.
Thanks a lot getting back to me and for your interest in this issue.
I would like to confirm that when I resume from suspend the machine
seems to be running fine. I can SSH, if I connect external monitor the
external monitor comes up and I can login and work as normal, however
the internal OLED is turned off and I was not able to figure out how to
turn it on.
I tried pm-suspend and experience the same issue. The screen stays black
after resume.
Ok, thanks for the feedback. Given this, I'm reassigning this to the kernel.
I tried xrandr with brightness 1, i tried display setting in KDE to turn
it on, etc from the external monitor, but nothing seems to be helping.
When I reboot, the OLED works as normal.
Resume from hibernate also does not cause any issue.
If you have access to the system via SSH, it might help to provide a
dmesg output.

Regards,
Michael
--
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?
Jiri Kanicky
2019-11-25 02:20:01 UTC
Reply
Permalink
Further to the issue, when I set BIOS to discrete graphics only, I am
experiencing the same issue when booting the system from shutdown state.
The screen stays off, but the OS is booted and working.
Mark Pearson
2019-11-25 16:50:02 UTC
Reply
Permalink
Hi,

I don't know if this is helpful but I recently debugged a similar issue on the Lenovo P1Gen2. There the problem is only with the integrated graphics and an OLED screen - we tracked it down to an issue in the i915 driver where it wasn't giving enough time for eDP link training. It turned out this was due to the driver using the wrong clocks after a suspend and resume .

Intel recently up-streamed a fix (commit 2f216a85 - it went into 5.4-rc8)

I'm working on doing a patch that I'm hoping to get into Debian to backport - just not done yet :) The patch is pretty small (I've attached it) so you might want to give it a go and see if it helps.

Note - we did also test X1 extreme with OLED panel and didn't see a problem there....but it's a really subtle timing issue so some units might be more susceptible than others. Maybe worth a go?

If it does make a difference let me know.
Mark
-----Original Message-----
Sent: Sunday, November 24, 2019 9:13 PM
Subject: [External] Bug#944350: systemd: screen remains off after waking up
from suspend
Further to the issue, when I set BIOS to discrete graphics only, I am
experiencing the same issue when booting the system from shutdown state.
The screen stays off, but the OS is booted and working.
Jiri Kanicky
2019-11-30 12:10:02 UTC
Reply
Permalink
Hi.

It seems that Debian 5.3.0-2-amd64 includes the code in the patch
already and I still have the issue.

Jiri
Post by Mark Pearson
Hi,
I don't know if this is helpful but I recently debugged a similar
issue on the Lenovo P1Gen2. There the problem is only with the
integrated graphics and an OLED screen - we tracked it down to an issue
in the i915 driver where it wasn't giving enough time for eDP link
training. It turned out this was due to the driver using the wrong
clocks after a suspend and resume .
Post by Mark Pearson
Intel recently up-streamed a fix (commit 2f216a85 - it went into 5.4-rc8)
I'm working on doing a patch that I'm hoping to get into Debian to
backport - just not done yet :) The patch is pretty small (I've attached
it) so you might want to give it a go and see if it helps.
Post by Mark Pearson
Note - we did also test X1 extreme with OLED panel and didn't see a
problem there....but it's a really subtle timing issue so some units
might be more susceptible than others. Maybe worth a go?
Post by Mark Pearson
If it does make a difference let me know.
Mark
Mark Pearson
2019-12-06 19:00:01 UTC
Reply
Permalink
Hi Jiri,

I'm pretty sure it's not in 5.3.0-2. Are you sure? I can look again but I recently did put in a merge request to get it pulled into sid: https://salsa.debian.org/kernel-team/linux/merge_requests/188

Not sure if/when it will be accepted (it's only my 2nd attempt at doing this) but hopefully someone looks at it and thinks it's useful)

Mark
-----Original Message-----
Sent: Saturday, November 30, 2019 6:56 AM
Subject: Bug#944350: RE: [External] Bug#944350: systemd: screen remains
off after waking up from suspend
Hi.
It seems that Debian 5.3.0-2-amd64 includes the code in the patch
already and I still have the issue.
Jiri
On Mon, 25 Nov 2019 16:45:55 +0000 Mark Pearson
Post by Mark Pearson
Hi,
I don't know if this is helpful but I recently debugged a similar
issue on the Lenovo P1Gen2. There the problem is only with the
integrated graphics and an OLED screen - we tracked it down to an issue
in the i915 driver where it wasn't giving enough time for eDP link
training. It turned out this was due to the driver using the wrong
clocks after a suspend and resume .
Post by Mark Pearson
Intel recently up-streamed a fix (commit 2f216a85 - it went into
5.4-rc8)
Post by Mark Pearson
I'm working on doing a patch that I'm hoping to get into Debian to
backport - just not done yet :) The patch is pretty small (I've attached
it) so you might want to give it a go and see if it helps.
Post by Mark Pearson
Note - we did also test X1 extreme with OLED panel and didn't see a
problem there....but it's a really subtle timing issue so some units
might be more susceptible than others. Maybe worth a go?
Post by Mark Pearson
If it does make a difference let me know.
Loading...