[IBM 1834RTG] late resume failure

I can confirm this on same laptop.

This also affects two other systems of which I am aware: HP Elitebook 2530P and a legacy Gateway desktop (originally desktop OS was with Windows Me). This seems to be quite a prolific issue across a number of systems. Particularly bothersome on laptops due to the likelihood of entering suspend mode while running on batteries.

Hi kwizk, I'm quite happy to know more systems are affected. It means this issue is not isolated and hardware-specific, and hopefully will be tackled more efficiently. Suspend is an important mode, I for one use it extensively on my laptop, so I agree with you when you say it's particularly bothersome.

Also, people in the thread mentioned in my original report indicate the following systems as being affected: Sony Vaio VGN-FW139E, Inspiron 600m, Aspice 3690. With regards to the latter, he or she suspects it's the new kernel, because "in Juanty the suspend/hibernation function works fine with 2.6.28, but in other distributions with kernel 2.6.31 (like Mandriva 2010 or Fedora 12) the system freeze too". So, can we get a confirmation from someone up above on this, please? Any sign would be greatly appreciated! Many thanks in advance.

Same, or at least similar, problems on my five year old Chembook (asus) notebook. Hard freeze sometimes on waking up, sometimes in the middle of graphic operations, sometimes returning from hibernate. On one occasion still had mouse movement but everything else frozen; all other occasions complete freeze. Happens more frequently running Ubuntu 9.10 than it did in 9.04.

Hi rlichter, your comment drew my attention to the frequent freezes you seem to experience, especially to the part where you describe freezes "in the middle of graphic operations", or where mouse movements were still possible but everything else froze. The freezes you describe have been experienced by other users, and are reported in Bug #479296, here https://bugs.launchpad.net/ubuntu/+bug/479296. It never occurred to me to draw a connection between these two bugs, but after your comment I wonder: can the two issues be triggered by a common cause? Can anyone confirm this might be the case, please?

Hi also Sony Vaio VGN-FZ31S affected

Hi, I suspect this issue is caused by whatever is causing the issue reported at Bug #488673. Apport indicates a serious kernel problem.

Further to my previous comment, Bug #493505 is joining the list of related bugs. This is getting frustrating. Since the kernel problem reported is serious, can anyone change at least the importance of the bug to medium or higher, please?

In my dell mini 9 worked fine beafore today, now it freeze.

My e-machine desktop frequently freezes while the screen is asleep. Since I upgraded to Karmic this has been a real problem. Only way to get the system back is to hard boot. I get lots of these before things lock up:

Dec 9 18:00:46 localhost kernel: [ 6600.564034] i915/0 D c07ff1b0 0 295 2 0x00000000
Dec 9 18:00:46 localhost kernel: [ 6600.564042] f6463f04 00000046 f6bacacc c07ff1b0 f6bacd38 c07ff1b0 c07ff1b0 c07ff1b0
Dec 9 18:00:46 localhost kernel: [ 6600.564054] f6bacd38 442269e0 00000581 c07ff1b0 c07ff1b0 ea1c7e00 00000581 f6bacaa0
Dec 9 18:00:46 localhost kernel: [ 6600.564066] f6a8a014 f6a8a018 ffffffff f6463f30 c05620d6 f6bacaa0 f6a8a01c f6bacaa0
Dec 9 18:00:46 localhost kernel: [ 6600.564078] Call Trace:
Dec 9 18:00:46 localhost kernel: [ 6600.564091] [<c05620d6>] __mutex_lock_slowpath+0xc6/0x130
Dec 9 18:00:46 localhost kernel: [ 6600.564098] [<c0561ff0>] mutex_lock+0x20/0x40
Dec 9 18:00:46 localhost kernel: [ 6600.564130] [<f82d4c0a>] i915_gem_retire_work_handler+0x2a/0x70 [i915]
Dec 9 18:00:46 localhost kernel: [ 6600.564138] [<c015459e>] run_workqueue+0x6e/0x140
Dec 9 18:00:46 localhost kernel: [ 6600.564158] [<f82d4be0>] ? i915_gem_retire_work_handler+0x0/0x70 [i915]
Dec 9 18:00:46 localhost kernel: [ 6600.564165] [<c01546f8>] worker_thread+0x88/0xe0
Dec 9 18:00:46 localhost kernel: [ 6600.564171] [<c0158da0>] ? autoremove_wake_function+0x0/0x40
Dec 9 18:00:46 localhost kernel: [ 6600.564177] [<c0154670>] ? worker_thread+0x0/0xe0
Dec 9 18:00:46 localhost kernel: [ 6600.564182] [<c0158aac>] kthread+0x7c/0x90
Dec 9 18:00:46 localhost kernel: [ 6600.564187] [<c0158a30>] ? kthread+0x0/0x90
Dec 9 18:00:46 localhost kernel: [ 6600.564194] [<c0103fc7>] kernel_thread_helper+0x7/0x10

Hi,
I experience the same issues as lelamal on my Thinkpad R50e. Same issue, the system freezes every now and then and I can't suspend to disk or ram anymore. Used to work perfectly before Karmic Koala. I would greatly appreciate any help to solve thsi problem.
Olf

Yes, I could suspend and resume with no problem under 9.04. Now I can suspend but not resume. System powers up but screen is blank. HP Compaq nx6125 running 9.10 i386.

I've this problem with HP Compaq nx 5000. Closing the lid lets the computer freeze.

I upgraded from 9.04 to 9.10.

Same problem here with an R50e. Worked fine in 9.04 and also in previous versions (some earlier releases required a bit more tweaks), but now it cannot enter the suspended state -- half-moon keeps blinking, fan is running and the PC cannot wake up again.

-- Per.

Just tested with an older kernel. 2.6.31 (Karmic version) has this error, but booting up on a 2.6.28 from Jaunty makes suspend work fine again ;-)

However, this makes the sound completely non-functional, so it even if it is a workaround for this present bug, it renders the snd intel 8xx driver non-operational (aplay -l shows no devices)... so it isn't really a useable workaround for most, I guess.

-- Per.

Been doing a few more experiments. Definitely looks like a kernel problem to me. If I do

   # echo mem > /sys/power/state

on the 2.6.28 kernel, the machine goes to suspend mode immediately and wakes up as expected afterwards when the power button is pressed.

However, doing the same thing on the 2.6.31 kernel causes the disk to spin down but the machine just becomes non-responsive after that. The fan keeps running (and the screen remains on).

Same thing happens on 2.6.31 with

   # echo -n 3 > /proc/acpi/sleep

Let me know if any further experiments are needed on a failing machine.

-- Per.

I have the same problem as Lelamal with a Fujitsu Lifebook T5010: when I try to suspend or hibernate under Karmic, the machine never actually turns off. The machine seems to go through the suspend procedure and the screen goes black, but the power stays on. At that point, I can do nothing but turn it off by holding the power switch.

Suspend and hibernate were working just fine under Jaunty.

I depend on this laptop for work, and this is a *major* annoyance.

It looks like, on my Fujitsu T5010 at least, the suspend and hibernate work (almost) perfectly if I unmount or remove the SD card from the SD card reader slot prior to suspending/hibernating. I found that trick on some ubuntuforum thread. I always had an SD card in the slot (it's the best way to not loose it!), and never thought that could be a problem (it was working fine on Jaunty).

There is still a problem though: the wireless card doesn't wake up. But that's outside the scope of the present bug report.

I have the same problem with my HP Compaq. I will be happy to provide more information. But in the mean time can someone tell me how to make suspend and hibernate a noop. I don't want to disable or remove the menu options, I want to make it impossible for any app to trigger them. If disabling is not possible and this issue is not fixed soon I think I will switch to gentoo.

Sorry, I didn't read the problem description close enough. My problem is that it suspends fine but I can't resume. I will look for a more appropriate bug report or file a new bug.

I have a Thinkpad X61 tablet and have this same symptom regarding "sleep" or "suspend"
(this is not "hibernate")

** press Fn+F4 to request suspend
(alternately battery discharge or idle state asks for suspend before battery death)
** X11 desktop (I run gnome without 3D) and applications wind down visually
** screen goes dark
** half-moon LED blinks
(under v8.04.3 it was on solid during suspend)
...
When I open the lid or try to wake the laptop nothing happens. The only way to wake things up requires that I press and hold the power button until full power off. When I then power-on, the system starts reasonably.
I have no idea how to discover if there were any left-over artifacts from the failed suspend-->resume.

Someone tell me what to instrument and what to report and I'll do what I can.

For the record, I tried hibernate-resume and received the same behavior as happened with suspend-resume: blinking "sleep" LED and hard power-off then power-on to restart.

Like I said, I have a Thinkpad X61 Tablet.

Sysinfo reports the following:
[code]
        Running Ubuntu Linux, the Ubuntu 9.10 (karmic) release.
        GNOME: 2.28.1 (Ubuntu 2009-11-03)
        Kernel version: 2.6.31-16-generic-pae (#53-Ubuntu SMP Tue Dec 8 05:20:21 UTC 2009)
...
        GenuineIntel, Intel(R) Core(TM)2 Duo CPU L7500 @ 1.60GHz
        Number of CPUs: 2
        CPU clock currently at 800.000 MHz with 4096 KB cache
        Numbering: family(6) model(15) stepping(11)
        Bogomips: 3191.85
[/code]

I'm also having trouble getting CPU frequency scaling to work, but that is way off this reservation...

  I've got an ancient (10-years-old this month [January]) Apple PowerBook G3 Pismo 500MHz [PPC] that suffers from this same problem. I've used Xubuntu in the past (Hardy and Jaunty) and decided to give it another go with Karmic. While my suspending was sporadic at worst before, it bricks up now.

  The Pismo will shut down everything except some electronic device that emits a faint, high-pitched whine, and won't come out of sleep (the sleep light doesn't start blinking, either, so I'm not sure it's really asleep).

  I'm glad I decided to dual-boot Xubuntu Karmic and Mac OS 10.4 Tiger. Otherwise, I'd be stuck.

  On a more positive note, I love Ubuntu Karmic Netbook Remix (GNOME with some KDE elements installed) on the family HP Pavilion desktop. With the netbook launcher and CairoDock, it's almost painfully easy to use. I was trying to get such an installation working on my laptop, but it hasn't worked yet.

UPDATE:

  I went into my software sources on a hunch and decided to start accepting the "Karmic Proposed" packages. After running software update, my laptop took a step in the right direction: I can now suspend (and quite quickly compared to Hardy and Jaunty), but when I come out of suspend, my laptop has an enlarged portion of the upper right corner of the screen in inverted colors over the whole monitor, and everything is frozen. Closer to functionality, but not quite there.

  The updates included Linux firmware, a few lib- packages, and some updates to Xserver. Either one of these (firmware or Xserver) sounds like a likely culprit to me.

  Just my two cents.

I've tested karmic alpha 2 and there is a little change. When you try to hibernate, the computer actually turns itself of, but when you restart it, it just starts normally, as when you shut down it normally. There is no change with suspending to ram.

Well, if I got the descriptions here right I got the same problem (->https://bugs.launchpad.net/ubuntu/+source/linux/+bug/516294).
Just tried what Austin L. said, enabling the Karmic Proposed packages, but absolutely nothing changed, same problem still exists, system goes almost off, screen is black but backlight is on, system is complete freezed and unaccessable and have to do a hard-reboot. But a kind of frightening this still exists since the kernel-version 2.6.30 and still exists in Lucid with 2.6.32 =/

hm, somehow I can't edit my last post, so here's the description for the attachment above:
setted the PM_DEBUG-Variable to true with 'export', switched to init-runlevel 1 and started suspending with 'pm-suspend'. Also added the line SUSPEND_MODULES="$SUSPEND_MODULES ehci-hcd uhci-hcd usbcore forcedeth" to my /etc/pm/config.d/00sleep_module, didn't have any effect...

...and running Karmic with an old 2.6.29-Vanilla-kernel leads to the result that our suspend-problem is a pure, non-ubuntu kernel fault I fear ('cause it works). Couldn't get any kernel above 2.6.30 working with jaunty but with 2.6.30 in jaunty suspend still worked. Well, if you I'm right with the opinion it's a vanilla-kernel-problem, please report it to the kernel team or let me know how to do that...

I have a similar problem on an Asus G1S. Sometimes it can resume from suspend just fine and other times it just cant resume. Seems to always fully suspend either way though...

@Everyone:

This being a suspend/resume bug, there is lots of interest from many people. To help diagnose and fix the issue we have to focus on one symptom at a time. This means that for this bug report, we need to only be concerned with symptoms matching what lelamal reported. Other symptoms, such as hangs on resume (not suspend) should be dealt with in another bug.

So just to be clear, these are the people I saw who seem to have different issues:

ermacaz
rlichter
Oz Man

Please look for other bug reports that more closely match your issues. If you don't find any, please open new bugs.

@BlackVodka, @lelamal:

Since this seems to work on 2.6.29 vanilla but not 2.6.30, I suggest browsing the bugzilla.kernel.org database and filing a new bug if you don't find any matching your issues. Once you've done so, link the kernel.org bug report to this bug report. Because this is an upstream issue, we really need to get them involved in trying to fix it.

Thanks

I've browsed the bugzilla.kernel.org database as you suggested, and found this bug: http://bugzilla.kernel.org/show_bug.cgi?id=14640. It was reported on the same day as this one, and it already refers to this page. Since I'm not familiar with Bugzilla, when you say "link the kernel.org bug report to this bug report", do you actually mean to paste a link to this bug report into a comment box as Wayne already did, or is there another way to link it? Please let me know if there is anything else I can do.

@lelamal:

Basically you've done everything you need to here already! By mentioning an upstream bug url in your comment you've linked it to this bug. Please continue to follow the upstream bug report, and when there's something we can pull into Ubuntu we will take a look at it.

Thanks

I have the same issue on my Sony VGN-FW11ER with kernel from Ubuntu karmic repository. But in my case I can make suspend to work (but lost brightness in this case). I install kernel image from mainline (At present I use 2.6.31.6 version). In this case brightness doesn't work (there no device in /sys/class/backlight/), but if I patch DSDT table, brightness starts working, /sys/class/backlight/ has device, but suspend do not want to resume - laptop completely hangs. So I think the problem is situated in kernel - acpi. What else my test shown that in kernel 2.6.29 brightness and suspend worked, in 2.6.30 suspend with brightness did not want to work.

I upgraded to Lucid, and the suspend issue seemed to have been fixed. However, I cannot be sure, for I couldn't remain on Lucid for long, due to the problems with Intel chipsets best described here: http://www.ubuntugeek.com/ubuntu-10-04-lucid-lynx-and-intel-video-chipsets.html. I was suggested to upgrade to Maverick if I wanted to avoid booting from a regressed kernel, and so I did. That finally solved the black screen issue. But I'm here to report that the issue of the present bug I reported in November against Karmic is still affecting me on Kernel 2.6.35-2, using fully updated Maverick. We would need somebody who was affected on Karmic to report whether it is still occurring on Lucid, so as to understand if this is a regression of the new kernel. Thanks.

Oh, and by the way, hibernate mode seems to work properly.

Yes, in Lucid it's solved, even with activated KMS (Grub-Line i915.modeset=1 activated) standby and hibernation working proberly. But, like you said, with any newer kernel (Vanilla and built from the Ubuntu-git-tree) which has e.g. the possibility for video playback on an Intel 855GM-chip, video playback works again flawlessly but the old-known bug with the standby-mode occurs again. And, hibernation still works, hibernation seems to work on a different way than standby does.
I've got a little bit these two patches I read about a few times in context to the video-playback-bug in Lucid with activated KMS in suspicion, but I ain't a developer, this was just an idea what could be responsible for this bug's "rebirth" in the kernels 2.6.33 and above.
But in summary, like I said, I tried the Vanilla-Kernels *32,*33,*34 and *35 from http://kernel.ubuntu.com/~kernel-ppa/mainline/ and built for myself the same versions from the Ubuntu-git-tree, but with none of them I got a working standby-mode.
Maybe also useful for (finally) fixing this bug, with the vanilla kernel 2.6.32 standby also didn't (!) work, but with the Ubuntu-git-tree version it worked. This was the cause getting me to compile 2.6.33 and so on from the Ubuntu-git-trees, but it didn't make any change =/
And another point, I can't really remember where, but I read about a week before in some forums that in Fedora 13 this issue is not present AND video playback is working. I would be glad to know if that's correct so maybe the Ubuntu-developers could get a benefit form knowing that, at the moment I can't afford the time to repartitionizing my hard drive and to test Fedora.

And guys, a little bit critic with good intentions, as a 'noob' you can't see any progress been made to this topic, is there anyone working on this bug or is it just forgotten? :)

This bug was filed against a series that is no longer supported and so is being marked as Won't Fix. If this issue still exists in a supported series, please file a new bug.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.