Bug 768937

Summary: suspend hangs on hp elitebook 8460p
Product: [openSUSE] openSUSE 12.1 Reporter: Rihards Olups <richlv>
Component: KernelAssignee: E-mail List <kernel-maintainers>
Status: RESOLVED DUPLICATE QA Contact: E-mail List <qa-bugs>
Severity: Major    
Priority: P5 - None CC: jeffm
Version: Final   
Target Milestone: ---   
Hardware: HP   
OS: openSUSE 12.1   
Whiteboard:
Found By: --- Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: dmesg of a clean boot
lsmod after booting
lspci output
lsusb output

Description Rihards Olups 2012-06-26 20:34:13 UTC 
Created attachment 496500 [details]
dmesg of a clean boot

User-Agent:       Opera/9.80 (X11; Linux x86_64; U; en) Presto/2.10.229 Version/11.64

opensuse 12.1, recently patched.

Linux 3.1.10-1.13-desktop #1 SMP PREEMPT Tue Jun 12 13:11:31 UTC 2012 (44a489e) x86_64 x86_64 x86_64 GNU/Linux

attempts to suspend mostly hang when suspending, although sometimes suspend/resume cycle succeeds. i haven't been able to determine what causes this difference (success is pretty rare).
i seem to recall both suspend and hibernate working w/o problems when i installed 12.1 initially, but i did not to lots of repeated tests.

-------------------------------- similar issues

this seems to be different than https://bugzilla.novell.com/show_bug.cgi?id=648742 as i can run all kinds of 3d things w/o any issues, and not running them does not make suspend succeed.

this seems to be different than https://bugzilla.novell.com/show_bug.cgi?id=725945 as this isn't asus and suspend does succeed every now and then

this is different than https://bugzilla.novell.com/show_bug.cgi?id=746580 as this system doesn't have realtek card

this doesn't seem to be http://askaralikhan.blogspot.de/2011/03/hp-elitebook-8540p-hibernatesuspend.html , as removing xhci_hcd did not help. while at that, everything else matching "hci" was removed (firewire_ohci, firewire_core, sdhci_pci, sdhci, mmc_core), which had no impact on the hanging.


-------------------------------- other things attempted

after trying the steps at https://bugzilla.novell.com/show_bug.cgi?id=749626 , system hung. some keys made it beep. single power button press made screen flash quickly. after a forced reboot screen brightness was stuck at 100% (but that seems to be a known upstream issue https://bugzilla.kernel.org/process_bug.cgi)

after reading http://www.kernel.org/doc/Documentation/power/basic-pm-debugging.txt , tried to follow http://en.opensuse.org/SDB:Suspend_to_RAM (namely, running s2ram -f -a 1/2/3), but that was probably almost useless - before suspending s2ram said "KMS graphics driver is in use, skipping quirks."

nevertheless, quick results : out of 5 attempts it hung in 4 of them. once it successfully suspended & resumed.

# s2ram -n
Machine unknown
This machine can be identified by:
    sys_vendor   = "Hewlett-Packard"
    sys_product  = "HP EliteBook 8460p"
    sys_version  = "A0001D02"
    bios_version = "68SCF Ver. F.22"

doublechecked that disconnecting all usb devices before a suspend attempt does not help.

unloaded iwlagn module before suspending, which did not help.

to make sure software state is ok,
echo freezer > /sys/power/pm_test
echo disk > /sys/power/state
were tried, which worked without any issues.

Reproducible: Always
Comment 1 Rihards Olups 2012-06-26 20:35:57 UTC 
Created attachment 496501 [details]
lsmod after booting
Comment 2 Rihards Olups 2012-06-26 20:36:15 UTC 
Created attachment 496502 [details]
lspci output
Comment 3 Rihards Olups 2012-06-26 20:36:34 UTC 
Created attachment 496503 [details]
lsusb output
Comment 4 Rihards Olups 2012-06-26 20:57:23 UTC 
downgraded to 
Linux 3.1.0-1.2-desktop #1 SMP PREEMPT Thu Nov 3 14:45:45 UTC 2011 (187dde0)
x86_64 x86_64 x86_64 GNU/Linux
from main 12.1 repo.

suspended 3 times in a row successfully, third time resume hung :>
so either it is harder to trigger in 3.1.0, or it's just some random chance
playing here. did the
echo devices > /sys/power/pm_test
echo mem > /sys/power/state
sequence again and ended up with flashing screen upon attempt to 'resume'
Comment 5 Rihards Olups 2012-06-26 21:00:24 UTC 
added /etc/pm/config.d/radeon-suspend with contents of :
SUSPEND_MODULES="radeon"

5 successful suspends & resumes with "echo mem > /sys/power/state" and pm_test set to devices

so i'll point a finger at radeon kms driver, although i'd need more successful suspend/resume with this hack to be sure.

which makes it very similar to a 2 year old bug : https://bugzilla.novell.com/show_bug.cgi?id=610514
Comment 6 Rihards Olups 2012-06-27 12:05:57 UTC 
another 2-year-old bug about this : https://bugzilla.novell.com/show_bug.cgi?id=626936
Comment 7 Jeff Mahoney 2012-06-27 16:32:12 UTC 
There were suspend issues in the kernel update recently released. We're re-releasing the update with the issue fixed. I'd like to keep your recent suspend issues and historical ones separate until you can try to reproduce with the new update.

In order to expedite testing, you can grab the kernel from the Kernel:openSUSE-12.1 repository at: http://download.opensuse.org/repositories/Kernel:/openSUSE-12.1/

Please let us know if that fixes the issues you've been experiencing lately.
Comment 8 Rihards Olups 2012-06-29 09:56:33 UTC 
upgraded to 3.1.10 from that repo and it indeed does seem to solve the problems - even the radeon nuking doesn't seem to be necessary. i've only done a few suspend/hibernate cycles, so i'd need more time to figure out whether problems are gone for good (leaving needinfo for now).

btw, what other problems could the previous version have caused ? could it resulted in severely degraded (at least wireless) network performance ?
Comment 9 Rihards Olups 2012-07-13 18:21:47 UTC 
because of issues with the opensource radeon driver i went with fglrx, and in the process moved over to kernel 3.1.10-1.16-desktop

does this version contain the suspend problem mentioned ?

(yes, i do see occasional hangs again :) )
Comment 10 Jeff Mahoney 2012-07-26 21:59:41 UTC 
3.1.10-1.16 contains the updated kernel with the suspend issues fixed. Any issues with suspend with the fglrx differ loaded will have to be referred to ATI or reproduced without the fglrx driver loaded.

Until then, with comment #8 in mind, I'm closing as a duplicate.

*** This bug has been marked as a duplicate of bug 764864 ***