[Dell Mini-9] suspend/resume problem

[Gavin McCullagh]

It's not clear why but after a recent jaunty dist-upgrade, I get this "kernel suspend problem" reported.

The machine is a dell mini-9

ProblemType: KernelOops
Architecture: i386
DistroRelease: Ubuntu 9.04
ExecutablePath: /usr/share/apport/apportcheckresume
FlagFile: suspend
InterpreterPath: /usr/bin/python2.5
OopsText: Suspend/Resume or hibernate problem

Originator: Steve Conklin <email address hidden>
Package: linux-image-2.6.28-7-generic 2.6.28-7.20
ProcAttrCurrent: unconfined
ProcCmdLine: User Name=UUID=ba8749d6-7cc6-42ef-9fb5-bf43900008f7 ro quiet splash
ProcCmdline: /usr/bin/python /usr/share/apport/apportcheckresume
ProcEnviron: PATH=(custom, no user)
ProcVersionSignature: Ubuntu 2.6.28-8.22-generic
SourcePackage: linux
StressLog: Error: [Errno 2] No such file or directory: '/var/lib/pm-utils/stress.log'
Tags: suspend resume hibernate
Title: Suspend/Resume or hibernate problem
UserGroups:

[Andy Whitcroft (apw)]

Did you recently have a suspend failure, this could have been some time before the report was triggered and offered to you. If you did, did the machine break while going to sleep, or while coming back from sleep? Is this reproducible say over 10 cycles. Has suspend ever worked in the past, if so which kernel release (cat /proc/version_signature)? Please include any information as to the circumstances leading up to this failure, for example did your battery run out? Did the machine lock up, did any messages come out? If it is reproducible could you try the suspend from VT1, press ctrl-alt-F1, login, and run pm-suspend there. If you could test and report back here that would be helpful.

[Gavin McCullagh (gmccullagh)]

> Did you recently have a suspend failure, this could have been some time before the report was triggered and offered to you.

Yes. Occasionally I have, possibly shortly before the message came on-screen. I think this failure happened occasionally prior to the jaunty upgrade too.

> If you did, did the machine break while going to sleep, or while coming back from sleep?

While coming back. One is left with a black screen with a cursor flashing in the top-left corner.

> Is this reproducible say over 10 cycles.

I'll try it out soon. From memory, over 10 cycles it probably does fail once, but I need to check.

> Has suspend ever worked in the past, if so which kernel release (cat /proc/version_signature)?

Yes it works most of the time. But now and then it fails.

> Please include any information as to the circumstances leading up to this failure, for example did your battery run out? Did the machine lock up, did any messages come out? If it is reproducible could you try the suspend from VT1, press ctrl-alt-F1, login, and run pm-suspend there. If you could test and report back here that would be helpful.

I'll hopefully do this in the next couple of days.

Gavin

[Gavin McCullagh (gmccullagh)]

Hi,

I'm in the middle of suspending & restarting over and over at the moment. I have yet to experience any problem and I'm well beyond ten retries now. I'll be using it on and off over the next few days so I'll keep an eye on it.

I do see one dmesg entry which says:

  WARNING: synaptics was reset on resume, see synaptics_resume_reset if you have trouble on resume

I'm not sure if this is relevant or not though.

Gavin

[Gavin McCullagh (gmccullagh)]

I noticed this failure again.

It's in bug 335431 which I've marked as duplicate of this one. This seems pretty rare and hard to reproduce I'm afraid

[Markus Schlager (m-slg)]

$ uname -a
Linux neuzing 2.6.28-11-generic #42-Ubuntu SMP Fri Apr 17 01:57:59 UTC 2009 i686 GNU/Linux

ubuntu jaunty-rc on dell mini 9, no swap-partition

This may be related:
http://ubuntuforums.org/showthread.php?t=1046727

I can confirm this behavior, it's precisely the same here: With SD-card inserted I can suspend exactly once. On resume the partitions on the SD-card seem to get remounted, at least nautilus-windows pop up. Trying to suspend for a second time freezes the computer with a black screen and a blinking cursor in the upper left corner. Attached is an excerpt of /var/log/syslog when this happened.

Without any SD-card inserted, suspend works better, although repeating suspend-resume sometimes brings up only the login-screen and looses the session. Attached is an excerpt of /var/log/syslog which is ending with such a failure.

[Markus Schlager (m-slg)]

attached the syslog-excerpt with sd-card inserted

[Markus Schlager (m-slg)]

attached output of

$ bash -x /etc/acpi/sleep.sh > sleepE5.log 2>&1
where suspend succeded for the last time, sd-card not mounted

and
$ bash -x /etc/acpi/sleep.sh > sleepE6.log 2>&1
where suspend failed (freeze), sd-card mounted

as well as two copies of /var/log/pm-suspend.log:

pm-suspend1.log: suspend successful (sd-card not mounted)
pm-suspend2.log: suspend failed (freeze) (sd-card mounted)

[Gary MicGee (worthing-jason)]

I am also having this issue. 100% of the time, if the SD card is mounted, then the system will lock when put to sleep. Best I can tell, it is trying to put the SD card to sleep, and cannot due to it being mounted. Unmounting the SD card by hand just prior to sleeping the system works, also 100%. However, while that is a good short-term solution, is it not a lasting one. If there is anything anyone needs done to show this issue in addition to what is mentioned, and the logs shown earlier, just ask.

Also, if it helps, I checked several logs, with no discernible difference in the logs between a successful suspend, and an unsuccessful one, Right up to before the statement that the CPU is being put into a sleep mode. (I don't remember the exact wording right now). When comparing them, there was that statement, then the one that it is waking.

Anything else I can contribute to assist in this effort, please let me know.

[Markus Schlager (m-slg)]

Unmounting the SD-card is what helps me too for the moment. But on resume
all partitions on the SD-card get mounted automatically and nautilus opens
windows for each of them.

On boot the SD-card doesn't get mounted, but whenever I start nautilus (on
whatever directory) all the partitions on the SD-card get mounted - even
if I configure nautilus not to mount media automatically.

[Peter Meyer (petermeyer69)]

Let's as the question another way. What made this work in the Dell Hardy release?

[Jeff Fortin Tam (kiddo)]

I have a different experience here. Suspend and resume work 99% of time (hell, I even ran the full https://wiki.ubuntu.com/KernelTeam/SuspendResumeTesting test without having the problem once, damnit), but once in a while, and for reasons that completely elude me, it will refuse suspending.

Indeed, sometimes when I close the lid or press the power button (both are set to make the computer suspend), I will see a black screen with a flashing "_" on the top left corner, and the computer will not go into powersave mode. To me, this is much more annoying than a computer failing to come back from suspend, because it will drain the battery.

This is regardless of the presence or absence of an SD card; I have a 8gb SDHC FAT32 card inserted 100% of time and the computer seems to not have any problems with it 99% of time. I checked if it made a difference whether it was mounted or not at suspend time, and it doesn't, in my case. I'm not sure if I need to stick to this bug report or report a new one.