Bug 11064 - resume failure on thinkpad x61s
resume failure on thinkpad x61s
Status: CLOSED CODE_FIX
Product: Power Management
Classification: Unclassified
Component: Hibernation/Suspend
All Linux
: P1 normal
Assigned To: Rafael J. Wysocki
:
Depends on:
Blocks: 7216 10492
  Show dependency treegraph
 
Reported: 2008-07-09 13:24 UTC by maximilian attems
Modified: 2008-07-17 01:16 UTC (History)
3 users (show)

See Also:
Kernel Version: 2.6.26-rc9
Tree: Mainline
Regression: Yes


Attachments
dmesg (44.76 KB, text/plain)
2008-07-09 13:24 UTC, maximilian attems
Details
Possible fix (1.24 KB, patch)
2008-07-13 21:19 UTC, H. Peter Anvin
Details | Diff

Description maximilian attems 2008-07-09 13:24:13 UTC
Latest working kernel version: 2.6.26-rc8
Earliest failing kernel version: 2.6.26-rc9
Distribution: Debian Lenny
Hardware Environment: Thinkpad x61s
Software Environment:
xserver-xorg-video-intel             2:2.3.2-2
linux-image-2.6.26-rc9-amd64                 2.6.26~rc9-1~experimental.1~snapshot.11810

Problem Description:
On wake up short blinking of the leds, but moon stays on and display is frozen

Steps to reproduce:
echo mem > /sys/power/state

Non longer resumes, haven't had time for git cherry-pick, will do tomorrow.
resume previously crashed xorg every tenth time, but now it always reliably freezes.
Comment 1 maximilian attems 2008-07-09 13:24:58 UTC
Created attachment 16777 [details]
dmesg
Comment 2 Rafael J. Wysocki 2008-07-09 15:53:04 UTC
Can you try to revert commit 4b4f7280d7fd1feeff134c2cf2db32fd583b6c29 and retest, please?
Comment 3 maximilian attems 2008-07-10 03:41:31 UTC
4b4f7280d7fd1feeff134c2cf2db32fd583b6c29 isn't easily revertable.

anyway bisecting and currently on second try 3ea9eed49346eb80f17f1c6539c47dc508be1173 is good as expected,
will post bisect result soonest.
Comment 4 Rafael J. Wysocki 2008-07-10 03:58:25 UTC
(In reply to comment #3)
> 4b4f7280d7fd1feeff134c2cf2db32fd583b6c29 isn't easily revertable.

Reverts for me 100% cleanly.  Can you please verify that you're testing -rc9 without any additional patches?
Comment 5 maximilian attems 2008-07-10 05:07:23 UTC
hmm bisecting on plain linux-2.6 linus branch, 537388bb6563bfde7c0307b95ecc1f7a1ae39b02 turned out to be good that includes 4b4f7280d7fd1feeff134c2cf2db32fd583b6c29.

not sure if i reproduced correctly immediately launching suspend to ram after gdm login. will recheck plain 2.6.26-rc9 on desktop usage soonest.
thanks
Comment 6 Rafael J. Wysocki 2008-07-12 14:03:52 UTC
The problem has been confirmed.
References : http://marc.info/?l=linux-kernel&m=121584459809675&w=4
Comment 7 Rafael J. Wysocki 2008-07-13 11:11:13 UTC
Handled-By : Rafael J. Wysocki <rjw@sisk.pl>
Comment 8 Rafael J. Wysocki 2008-07-13 11:13:30 UTC
The solution appears to be not to use the s3_bios user space resume quirk related to graphics.

Still, we need to verify if the quirk is generally broken by commit 4b4f7280d7fd1feeff134c2cf2db32fd583b6c29.
Comment 9 H. Peter Anvin 2008-07-13 21:19:13 UTC
Created attachment 16809 [details]
Possible fix

Please test out this patch.
Comment 10 maximilian attems 2008-07-14 08:13:53 UTC
Back, will continue bisect later today and test proposed patch.

could verify that 2.6.26 is affected.
thanks
Comment 11 Bruno Randolf 2008-07-14 08:34:03 UTC
got the same problem on a X61 thinkpad and kernel 2.6.26. peter anvin's patch solves the problem for me!
Comment 12 maximilian attems 2008-07-16 11:16:09 UTC
Tested-by: maximilian attems <max@stro.at>

patch solves resume problem. thanks hpa :)
Comment 13 Rafael J. Wysocki 2008-07-16 12:58:56 UTC
Fixed by: commit 3bf2e77453a87c22eb57ed4926760ac131c84459
Comment 14 maximilian attems 2008-07-17 01:16:12 UTC
cool, please send it over to stable@kernel.org too, thanks.

Note You need to log in before you can comment on or make changes to this bug.