Bug 11698

Summary: 2.6.27-rc7, freezes with more than 1 s2ram cycle
Product: Power Management Reporter: Rafael J. Wysocki (rjw)
Component: Hibernation/SuspendAssignee: Rafael J. Wysocki (rjw)
Status: CLOSED INSUFFICIENT_DATA    
Severity: normal CC: alan, kernel, rui.zhang
Priority: P1    
Hardware: All   
OS: Linux   
Kernel Version: 2.6.31 Subsystem:
Regression: Yes Bisected commit-id:
Bug Depends on:    
Bug Blocks: 7216, 11167    
Attachments: dmesg before suspend
dmesg after the first suspend
dmesg after the pm_debug session

Description Rafael J. Wysocki 2008-10-04 11:28:54 UTC
Subject    : 2.6.27-rc7, freezes with > 1 s2ram cycle
Submitter  : Soeren Sonnenburg <kernel@nn7.de>
Date       : 2008-09-29 11:29
References : http://marc.info/?l=linux-kernel&m=122268780926859&w=4
Handled-By : Rafael J. Wysocki <rjw@sisk.pl>

This entry is being used for tracking a regression from 2.6.26.  Please don't
close it until the problem is fixed in the mainline.
Comment 1 Soeren Sonnenburg 2008-10-08 07:45:53 UTC
I finally found the time to test your patch 
http://marc.info/?l=linux-kernel&m=122307130419753&w=4
on 2.6.27-rc9. Same thing. I actually managed to trigger the hang by issuing find / >/dev/null .
Comment 2 Rafael J. Wysocki 2008-10-08 10:05:21 UTC
It's a different problem, then.

I'm observing something similar on a MSI Wind, but I suspect in my case it's caused by the out-of-the-tree wireless driver, which is buggy.

It may be related to a low memory corruption too (bug #11237).
Comment 3 Soeren Sonnenburg 2008-10-27 04:24:58 UTC
2.6.28-rc2 hangs completely on resume, so cannot tell if this is related/whether this bug is still there.
Comment 4 Rafael J. Wysocki 2008-10-27 04:37:43 UTC
Can you please try with this patch from bug #11827:

http://bugzilla.kernel.org/attachment.cgi?id=18453&action=view

?
Comment 5 Rafael J. Wysocki 2008-11-16 11:05:14 UTC
Have you tried the patch from comment #4?
Comment 6 Zhang Rui 2008-11-20 21:44:17 UTC
ping soeren. :)
Comment 7 Rafael J. Wysocki 2008-11-24 13:16:38 UTC
On Monday, 24 of November 2008, Soeren Sonnenburg wrote:
> On Sat, 2008-11-22 at 21:59 +0100, Rafael J. Wysocki wrote:
> > This message has been generated automatically as a part of a report
> > of regressions introduced between 2.6.26 and 2.6.27.
> > 
> > The following bug entry is on the current list of known regressions
> > introduced between 2.6.26 and 2.6.27.  Please verify if it still should
> > be listed and let me know (either way).
> 
> current mainstream hangs on *every* resume (Carlos observed the same
> hangs on wakeup from s2ram but while the cause for him were the wireless
> drivers it must be something else here)...
> 
> so this bug and the bug I am having is again hidden under something
> else. one thing I noticed: The machine resumes and numlock still works
> when using an external usb keyboard. still no display and also no second
> s2ram possible :(
Comment 8 Zhang Rui 2008-12-08 19:45:02 UTC
Please attach the dmesg output after the first s2ram.

use s3ram for the first suspend/resume, and then
# cat /sys/power/pm_test 
[none] core processors platform devices freezer
#echo {core,processors,platform,devices,freezer} > /sys/power/pm_test
#echo mem > /sys/power/state
wait and see if the laptop can resume back in about 10 seconds and tell me in which mode the laptop freezes
Note that you don't need to press the power button.
Comment 9 Rafael J. Wysocki 2008-12-09 12:20:46 UTC
On Tuesday, 9 of December 2008, Soeren Sonnenburg wrote:
> On Mon, 2008-12-08 at 19:40 +0100, Rafael J. Wysocki wrote:
> > On Monday, 8 of December 2008, Soeren Sonnenburg wrote:
> > > On Sun, 2008-12-07 at 21:16 +0100, Rafael J. Wysocki wrote:
> > > > On Thursday, 4 of December 2008, Soeren Sonnenburg wrote:
> > > > > On Wed, 2008-12-03 at 23:40 +0100, Rafael J. Wysocki wrote:
> > > > > > This message has been generated automatically as a part of a report
> > > > > > of regressions introduced between 2.6.26 and 2.6.27.
> > > > > > 
> > > > > > The following bug entry is on the current list of known regressions
> > > > > > introduced between 2.6.26 and 2.6.27.  Please verify if it still
> should
> > > > > > be listed and let me know (either way).
> > > > > > 
> > > > > > 
> > > > > > Bug-Entry   : http://bugzilla.kernel.org/show_bug.cgi?id=11698
> > > > > > Subject             : 2.6.27-rc7, freezes with &gt; 1 s2ram cycle
> > > > > > Submitter   : Soeren Sonnenburg <kernel@nn7.de>
> > > > > > Date                : 2008-09-29 11:29 (66 days old)
> > > > > > References  :
> http://marc.info/?l=linux-kernel&m=122268780926859&w=4
> > > > > > Handled-By  : Rafael J. Wysocki <rjw@sisk.pl>
> > > > > 
> > > > > well it freezes at any s2ram resume attempt, even without X just from
> > > > > console.
> > > > 
> > > > There are quite some things you can try.
> > > > 
> > > > First, it would be helpful to verify if it also fails in the minimal
> > > > configuration, ie. when booted with init=/bin/bash, as described
> > > > in Documentation/power/basic-pm-debugging.txt under
> > > > "Testing minimal configuration".
> > > 
> > > OK, I've found the problem:
> > > 
> > > 1. it does work with init=/bin/bash
> > > 2. it did not work after doing a binary search in the modules, even
> > > removing all modules did not help
> > > 3. I figured out that it only happens when I update the CPU's microcode,
> > > (booting with init=/bin/bash and /etc/init.d/microcode.ctl start
> > > triggers it)
> > > 
> > > Even better: s2ram always gave me my display back when s2ram under
> > > console.
> > > 
> > > Now I wonder if this is Bug #12100 ?
> > 
> > Well, that's quite likely.
> > 
> > > Unfortunately I need the microcode update, as I am suffering from
> > > a) PAT WC disabled due to known CPU erratum.
> > > b) coretemp coretemp.0: Errata AE18 not fixed, update BIOS or microcode
> of the CPU!
> > > 
> > > Anyway I now have a chance to test whether #11698 still occurs.
> > 
> > Hm, #11698 is the $subject bug.  Did you mean some other bug?
> 
> No, I meant this one (but it says freezes with > 1 s2ram cycle, which
> was supposed to say: when working for days or hours but having s2ram'd
> at least once the machine might freeze at some point)
Comment 10 Soeren Sonnenburg 2008-12-09 13:55:53 UTC
Regarding comment #8: I ran all the tests in the order core ... freezer, but I did not see any hang. Anyway I am attaching the dmesg before and after s2ram and all of these attempts. 

The problem is, that it is not reliably freezing on the >=2nd resume, but just sometimes and sometimes also in the middle of doing some work...
Comment 11 Soeren Sonnenburg 2008-12-09 13:56:55 UTC
Created attachment 19226 [details]
dmesg before suspend

dmesg before suspend (directly after booting up)
Comment 12 Soeren Sonnenburg 2008-12-09 13:57:20 UTC
Created attachment 19227 [details]
dmesg after the first suspend
Comment 13 Soeren Sonnenburg 2008-12-09 13:58:33 UTC
Created attachment 19228 [details]
dmesg after the pm_debug session
Comment 14 Soeren Sonnenburg 2008-12-13 22:58:41 UTC
yes, still happens with commit 6c34bc2976b30dc8b56392c020e25bae1f363cab
Comment 15 Shaohua 2008-12-23 22:51:27 UTC
how about boot option 'nohz=off'"
Comment 16 Soeren Sonnenburg 2009-01-01 12:00:39 UTC
nope, same thing. I am admittedly seeing hangs every now and then already on resume, so I cannot be sure if this is all the same bug. As the console display is left blank after resume, I will have to wait for kms for ati-radeons to be in-kernel to see any potential clues on the console.
Comment 17 Zhang Rui 2009-10-15 07:52:11 UTC
does the problem still exists in the latest kernel, say 2.6.31?
Comment 18 Anonymous Emailer 2009-10-15 17:46:48 UTC
Reply-To: sonne@debian.org

On Thu, 2009-10-15 at 07:52 +0000, bugzilla-daemon@bugzilla.kernel.org
wrote:
> http://bugzilla.kernel.org/show_bug.cgi?id=11698
> 
> 
> Zhang Rui <rui.zhang@intel.com> changed:
> 
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>                  CC|                            |rui.zhang@intel.com
> 
> 
> 
> 
> --- Comment #17 from Zhang Rui <rui.zhang@intel.com>  2009-10-15 07:52:11 ---
> does the problem still exists in the latest kernel, say 2.6.31?

Well the problem got worse, s2ram works a couple of times and causes
disk corruptions at some point... anyway I no longer own this machine...

Soeren
-- 
For the one fact about the future of which we can be certain is that it
will be utterly fantastic. -- Arthur C. Clarke, 1962