Ask Your Question
1

Hibernate/sleep broken after kernel update - recurring issue

asked 2014-02-04 10:25:46 -0600

jemc gravatar image

updated 2014-02-05 11:22:29 -0600

I have recurring problems with hibernation/sleep for my laptop breaking on kernel updates. This has been an intermittently recurring problem for me over the last few years with fedora, and usually leads me to a clean install of the OS. Even if I clean-install the same exact kernel version that broke my sleep/hibernate in update, it works just fine on the clean install. This leads me to think something in the kernel-update-process is breaking my sleep/hibernate (not the actual changes to the kernel). Maybe something with grub gets out of sync? Specifically, my problem with sleep/hibernate is that my laptop will go to a black screen, but the fan remains running at full speed, and I am unable to regain control of my computer until I do a 'hard' reboot by holding the power button.

I'm currently running Fedora 20 XFCE, which I recently installed over Fedora 19 for this same reason when all recent kernels after a certain one were breaking this feature. Kernel 3.11.10-301.fc20.x86_64 was installed from the live image, but when I updated to 3.12.8-300.fc20.x86_64 (my first kernel update since the reinstall), hibernate and sleep were broken again. This problem has been with me since I started with Fedora (version 16), but has never come back after a reinstall so quickly.

EDIT: I should note that at this moment, with the two kernels installed, I have the option to choose between them at my grub menu. Choosing the 3.12 kernel causes me to have this problem. Choosing the 3.11 kernel yields a system with working sleep/hibernate (as long as I select the same kernel as it's coming out of hibernation, and don't let it auto-choose the newer one). This has been typical behavior every time I've had this problem with past versions of fedora and past kernels. For a time, while the old kernel is still in my boot menu, I can select it for a working system; but when it gets pushed out of that grub "queue" I'm screwed.

Below is a portion of my log from journalctl, where I press the power button and select 'suspend' from the xfce menu that appears. The '-- Reboot' at the end must have been from me doing the 'hard' reboot the next morning. Please let me know if I can provide any other helpful information to help figure this out.

Feb 03 22:19:34 localhost.localdomain systemd-logind[687]: Power key pressed.
Feb 03 22:19:37 localhost.localdomain NetworkManager[785]: <info> sleep requested (sleeping: no  enabled: yes)
Feb 03 22:19:37 localhost.localdomain NetworkManager[785]: <info> sleeping or disabling...
Feb 03 22:19:37 localhost.localdomain NetworkManager[785]: <info> (p3p1): device state change: unavailable -> unmanaged (reason 'sleeping') [20 10 37]
Feb 03 22:19:37 localhost.localdomain NetworkManager[785]: <info> (p3p1): taking down device.
Feb 03 22:19:37 localhost.localdomain NetworkManager[785]: <info> NetworkManager state ...
(more)
edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted
0

answered 2014-02-04 15:32:52 -0600

updated 2014-02-05 05:35:31 -0600

I'm afraid this is a bug with your BIOS, see bugzilla.kernel.org; the manufacturer has to update the BIOS in order to resolve the issue permanently, since it is not an issue just with Fedora but other Linux distros as well. Keep looking around for other/related bugs and/or some kernel parameter that you may be able to append prior to booting; it might be possible to resolve it like that if it mostly occurs after a kernel update.

As a point of reference, albeit a biased one, this is a "Customer Review" for your laptop from Samsung's own website:

Not a good value No, I would not recommend this to a friend. July 1, 2011 this computer performs well but has a lot of flaws. Samsung software is not UAC compatible so you have to run as an administrator to keep from getting a thousand prompts at login. Does not have bluetooth. Poor display quality. Drivers are out of date. Appears to have a charging glitch coming out of sleep mode. I really hate the dual graphics chips from different manufacturers being stuffed on to one motherboard. Come on Nvidia can't you just make one chip that throttles according to demand? Touchpad has really poor multi-touch support. 2 finger scroll is terrible. On a good note this device looks professional. Love the keyboard. Core I7 cpu is really quick. Laptop runs pretty cool while doing some serious number crunching or 3d graphics rendering. USB 3.0 is a great feature. Pros: Bright display, Lightweight, performance, battery life Cons: Not high def, Did not meet expectations

Try manually suspending the PC from the terminal # /usr/sbin/pm-suspend as root or, as advised in the bug-report, # echo mem > /sys/power/state

Finally, take a look at this FC20 bug-report; it sounds very similar as well.

edit flag offensive delete link more

Comments

the bug-report mentions people had very limited success with performing a "Suspend" once after installing new firmware and/or BIOS reset; though, that was shortlived and the issue comes back. Since they have similar hardware as you, same manufacturer, and similar problem, it makes sense to look at it and to try to learn from their experiences.

I've had issues with my own laptop, just as you described until very recently when Lenovo (Thinkpad W520) decided to work with kernel developers to try to resolve the issue.

ILMostro gravatar imageILMostro ( 2014-02-04 16:53:19 -0600 )edit

I don't really think it's a similar problem. As I mentioned above, in all of the experiences I read of theirs, they could sleep fine from a software command, just not from the lid switch.

And I saw the posts about it working for one power cycle after some magic reboot sequence, but that isn't at all like my issue either. Mine will work flawlessly for months after a clean install, and break at some point after yum messes with my kernel in some way. I think it might have something to do with not having run yum update in a long time, then running it. Is it possible that the kernel upgrade "path" is broken for large leaps in kernel version. If so, is it possible to mimic how a clean install would set up the kernel?

jemc gravatar imagejemc ( 2014-02-04 18:18:41 -0600 )edit

Question Tools

Stats

Asked: 2014-02-04 10:25:46 -0600

Seen: 2,657 times

Last updated: Feb 05 '14