Linux 3.8.x 64b: screen remains off after resume

R_Rios · 2013-04-09 10:19:43

marzoul wrote:

Tested yesterday: no it is not solved for me!

I can confirm that. Seems that I'll have to stick with 3.7 for a little more.

fincheti · 2013-04-09 14:48:49

marzoul wrote:

Tested yesterday: no it is not solved for me!
Additionally the Linux changelog says nothing remotely related to it.
Actually the Nouveau developpers advised me to try Linux 3.9rc, but I still not tried due to compilation time.

I did not see anything in the changelog but there is a git commit 'Revert "drm/i915: write backlight harder"' on Mar 28. That seems to solve my problem.

But I only have the integrated graphics. Perhaps it's not just the i915?

nLinked · 2013-04-10 09:11:56

Can confirm same issue in 3.8.6.1, Intel onboard graphics, desktop PC, i915. Monitor remains in standby mode after resume in 90% resumes. Have to force PC off and startup again.

JAORMX · 2013-04-10 20:01:03

I have the same problem.

$ uname -a
Linux coffeelap 3.8.6-1-ARCH #1 SMP PREEMPT Sat Apr 6 07:27:01 CEST 2013 x86_64 GNU/Linux

A Dell laptop with Intel onboard graphics

marzoul · 2013-04-11 07:56:51

I'm using an Acer laptop with nVidia GeForce 8600M GT, under Nouveau driver, x86-64.
The screen stays off when resuming from RAM suspend, under any 3.8.x kernel.

So I've tried to compile a custom kernel from Linus' GIT repo (linux-git AUR package). Currently it's version 3.9 rc6.
And my screen works perfectly (I played a lot with config so it's not the default config I'm using).

Maybe the problem lies in the config used to build the core repo package... but I see no reason why this would have been broken on multiple types of machines.

Warnaud · 2013-04-11 08:13:42

Hi there,
Just my 2 cents, I had this problem in early version of 3.8 but it works now 3.8.6-3-ck / stock kernel arch

00:02.0 VGA compatible controller: Intel Corporation 3rd Gen Core processor Graphics Controller (rev 09)

Workaround at this time was:

echo on > /sys/bus/pci/devices/0000\:00\:19.0/power/control

Of course I was doing so remotely ...

marzoul · 2013-04-11 08:24:25

Warnaud wrote:

echo on > /sys/bus/pci/devices/0000\:00\:19.0/power/control 
Of course I was doing so remotely ...

If this is reliable, affected users could simply define a keyboard shortcut to launch this directly from the faulty machine, when the screen refuses to turn on.

Last edited by marzoul (2013-04-11 08:24:54)

Warnaud · 2013-04-11 08:27:04

marzoul wrote:

Warnaud wrote:
echo on > /sys/bus/pci/devices/0000\:00\:19.0/power/control 
Of course I was doing so remotely ...
If this is reliable, affected users could simply define a keyboard shortcut to launch this directly from the faulty machine, when the screen refuses to turn on.

Or putting this line in the script executed when resuming?

nLinked · 2013-04-11 08:30:39

Warnaud wrote:

Workaround at this time was:
echo on > /sys/bus/pci/devices/0000\:00\:19.0/power/control 
Of course I was doing so remotely ...

When I try this I get

bash: /sys/bus/pci/devices/0000:00:19.0/power/control: No such file or directory

My kernal:

uname -a
Linux HOSTNAME 3.8.6-1-ARCH #1 SMP PREEMPT Sat Apr 6 07:27:01 CEST 2013 x86_64 GNU/Linux

Available dirs:

uname -a
0000:00:00.0/ 0000:00:1a.0/ 0000:00:1c.1/ 0000:00:1f.0/ 
0000:00:02.0/ 0000:00:1b.0/ 0000:00:1d.0/ 0000:00:1f.2/ 
0000:00:16.0/ 0000:00:1c.0/ 0000:00:1e.0/ 0000:00:1f.3/

Last edited by nLinked (2013-04-11 10:38:51)

nLinked · 2013-04-13 15:57:58

Is it easy to upgrade to the 3.9 kernal? I'm new to Arch Linux but this problem is really bugging me. If I do upgrade to 3.9, will future pacman updates automatically update me to future kernals or will it then be stuck on 3.9?

marzoul · 2013-04-14 18:03:52

Hi,
I read the recent 3.8.7 Linux changelog in search for anything related.
https://www.kernel.org/pub/linux/kernel … eLog-3.8.7

There might be two interesting commits, the first about
"PCI/PM: Disable runtime PM of PCIe ports"
and the second about
"PCI/ACPI: Always resume devices on ACPI wakeup notifications"

The Warnaud workaround for its machine suggest our general problem could be caused by cases of bad power management related to PCIe ports. So it's worth giving a try with this kernel version.

I will not test this version right now due to my current laptop poor state for other reasons (upgrade to gnome 3.8 and subsequent).
However I suggest those who experienced issues with the linux 3.8 series give a little try with this version ?

marzoul · 2013-04-14 18:13:32

nLinked wrote:

Is it easy to upgrade to the 3.9 kernal? I'm new to Arch Linux but this problem is really bugging me. If I do upgrade to 3.9, will future pacman updates automatically update me to future kernals or will it then be stuck on 3.9?

Right now, the 3.9 version is not released by the Linux Foundation as being stable (only release candidates). So right now, it is not in the official Archlinux repositories, but it will be in the [testing] repo as son as it's released as stable.

Then, after a period of test by users (such like us if we want to give it a try), it will switch to branch [core] and pacman will notify all users that a Linux update is available. So you don't have to worry.

If you have personal reasons to do adopt the 3.9 version RIGHT NOW (+ compilation time), look in the AUR for the package linux-git. Be sure tu read the AUR page of the wiki to properly use AUR packages

nLinked · 2013-04-14 18:20:36

marzoul wrote:

nLinked wrote:
Is it easy to upgrade to the 3.9 kernal? I'm new to Arch Linux but this problem is really bugging me. If I do upgrade to 3.9, will future pacman updates automatically update me to future kernals or will it then be stuck on 3.9?
Right now, the 3.9 version is not released by the Linux Foundation as being stable (only release candidates). So right now, it is not in the official Archlinux repositories, but it will be in the [testing] repo as son as it's released as stable.
Then, after a period of test by users (such like us if we want to give it a try), it will switch to branch [core] and pacman will notify all users that a Linux update is available. So you don't have to worry.
If you have personal reasons to do adopt the 3.9 version RIGHT NOW (+ compilation time), look in the AUR for the package linux-git. Be sure tu read the AUR page of the wiki to properly use AUR packages

Thank you Is there a place to check the current progress of 3.9, such as a roadmap (sorry, new to Arch)?

Last edited by nLinked (2013-04-14 18:21:47)

marzoul · 2013-04-14 18:48:10

nLinked wrote:

Thank you Is there a place to check the current progress of 3.9, such as a roadmap (sorry, new to Arch)?

Well, it's not unusual for a Linux release to have 7 RCs, at the rhythm of one RC per week. And there is no way to predict when Mr Torvalds will decide that the next RC is the good one!
Currently it's RC6, so the release could be a matter of days, max 1-2 weeks I think.

Please do not continue this topic on this forum thread as it is not progressing in any way towards this original thread's subject, which an obscure problem on some machines at wake-up, for which the screen stays off, and only for 3.8 kernel series...

marzoul · 2013-04-17 20:18:33

The problem still exists on my laptop, with linux 3.8.8 !
But as I'm now using a custom compiled kernel from the Linus Torvalds GIT repo, I no longer suffer from this.
The 3.9 will soon be released... so it seems this 3.8 will have had this issue unsolved !!

R_Rios · 2013-05-01 20:00:01

Just informing that, at least for me, this seems to be fixed with linux 3.9.

Last edited by R_Rios (2013-05-01 20:00:14)

marzoul · 2013-05-02 06:57:18

Thnaks for the piece of info. I'll try to test that soon.

marzoul · 2013-05-20 22:45:44

I confirm my laptop works well under the core Linux 3.9.3 package.
It seems we'll have to conclude that the 3.8.x package was actually buggy (not knowing if it's from maintainers or the Linux guys), kinda a shame though.

If nobody has some new things to say about it, I'll close this thread in a few days...

nLinked · 2013-05-21 18:10:22

Also confirming fixed with 3.9

EDIT: It has started happening again! Display won't wake when resuming. Have to hard reset. Is there any way to troubleshoot this or debug? I am on kernal 3.9 and everything updated.

After a few days, is it still working for you?

Last edited by nLinked (2013-05-22 16:26:41)

marzoul · 2013-05-22 16:58:20

Yes I confirm the core Linux 3.9.x packages work like charm even after a dozen sleep/resume events... at least for my laptop. All packages up to date.

nLinked · 2013-05-22 17:34:14

Thank you. I have done some testing and can confirm my system doesn't resume when I use systemd for the suspend. Resume works perfectly when I use pm-suspend.

I guess then for the kernal side of things it must be solved - for me it's just a systemd issue now.

Edit: I have now solved my issue (if this helps anyone), by disabling the HPET timer option in my BIOS. Resume from any kind of suspend now works inc. systemd.

Last edited by nLinked (2013-05-22 22:06:48)

marzoul · 2013-05-23 08:04:33

So this may be of interest for systemd guys. I use systemd for power management and experienced no trouble (well at least Enlightenment is configured to use systemd), but I don't know what this BIOS option is.

Is there any chance this could be related to failures of linux 3.8.X ?

nLinked · 2013-05-23 08:24:29

It's started all over again! Today, resuming fails again regardless of that BIOS setting. It is a systemd or upower issue, but I'll keep updating in the other thread here as it may not be a kernal issue anymore? If I use pm-suspend to suspend, resume works fine. So I guess for me it's systemd-related.

https://bbs.archlinux.org/viewtopic.php … 3#p1276743

Arch Linux

#26 2013-04-09 10:19:43

Re: Linux 3.8.x 64b: screen remains off after resume

#27 2013-04-09 14:48:49

Re: Linux 3.8.x 64b: screen remains off after resume

#28 2013-04-10 09:11:56

Re: Linux 3.8.x 64b: screen remains off after resume

#29 2013-04-10 20:01:03

Re: Linux 3.8.x 64b: screen remains off after resume

#30 2013-04-11 07:56:51

Re: Linux 3.8.x 64b: screen remains off after resume

#31 2013-04-11 08:13:42

Re: Linux 3.8.x 64b: screen remains off after resume

#32 2013-04-11 08:24:25

Re: Linux 3.8.x 64b: screen remains off after resume

#33 2013-04-11 08:27:04

Re: Linux 3.8.x 64b: screen remains off after resume

#34 2013-04-11 08:30:39

Re: Linux 3.8.x 64b: screen remains off after resume

#35 2013-04-13 15:57:58

Re: Linux 3.8.x 64b: screen remains off after resume

#36 2013-04-14 18:03:52

Re: Linux 3.8.x 64b: screen remains off after resume

#37 2013-04-14 18:13:32

Re: Linux 3.8.x 64b: screen remains off after resume

#38 2013-04-14 18:20:36

Re: Linux 3.8.x 64b: screen remains off after resume

#39 2013-04-14 18:48:10

Re: Linux 3.8.x 64b: screen remains off after resume

#40 2013-04-17 20:18:33

Re: Linux 3.8.x 64b: screen remains off after resume

#41 2013-05-01 20:00:01

Re: Linux 3.8.x 64b: screen remains off after resume

#42 2013-05-02 06:57:18

Re: Linux 3.8.x 64b: screen remains off after resume

#43 2013-05-20 22:45:44

Re: Linux 3.8.x 64b: screen remains off after resume

#44 2013-05-21 18:10:22

Re: Linux 3.8.x 64b: screen remains off after resume

#45 2013-05-22 16:58:20

Re: Linux 3.8.x 64b: screen remains off after resume

#46 2013-05-22 17:34:14

Re: Linux 3.8.x 64b: screen remains off after resume

#47 2013-05-23 08:04:33

Re: Linux 3.8.x 64b: screen remains off after resume

#48 2013-05-23 08:24:29

Re: Linux 3.8.x 64b: screen remains off after resume

Board footer