ThinkPad X60 cannot Resume from Suspend

[mroth]

IBM ThinkPad X60s ( for more info https://wiki.ubuntu.com/LaptopTestingTeam/ThinkpadX60s wiki page ) will not resume from suspend to ram. Button presses have no affect on system. Power button appears to rev up the fan for a few seconds, but system remains off and sleep LED on hardware remains lit.

[Paul Sladen (sladen)]

Does the caps-lock key change the state of the caps-lock LED?

Does:

  Alt-SysRq-s Alt-SysRq-u Alt-SysRq-b

Sync, Unmount and Reboot the machine?

[mroth (mrothy)]

It appears either the "Fn" key or the "Power" key will turn on the fan (pseudo waking the machine?)

In this state, caps-lock DOES change the state of the LED.

The Alt-SysRq-s Alt-SysRq-u Alt-SysRq-b combos appear to have no effect. (However, is it a possibility this may be a complexity due to the Fn key being required to get a SysRq press?)

[Paul Sladen (sladen)]

Yes, Fn or Power will wake the machine, as may opening the lid if it has been closed. Okay, the caps-lock business indicates that the machine is coming back, to a point; but the display (and probably more) aren't.

Regarding the Fn needed to get SysRq... what have they done to the keyboard; are you able to take a photograph as I think Fn-F2 and Fn-F3 have been swapped too.

[mroth (mrothy)]

Found an existing image of the keyboard here:
http://www.notebookreview.com/assets/8714.jpg

[Paul Sladen (sladen)]

Oh. Wonderful(!). Go Lenovo++. Please could you run:

  sudo showkey -u

(from the console, Ctrl-Alt-F1) and file a separate bug report against `hotkey-setup` with the results. And of course, Fn-F8 doesn't even make sense on this laptop since it doesn't have a touchpad.

Could you also file a separate bug against `acpi-support` and attach the dmidecode data suggested in:

  https://wiki.ubuntu.com/LaptopTestingTeam/HotkeyResearch

so that we deal with the Fn-F3-is-now-battery-not-lock issue.

Many Thanks,

[Matthew Garrett (mjg59)]

Hibernate should work with the latest kernel (2.6.15-20)

[mroth (mrothy)]

Unfortunately, in latest kernel, sytem crashes when attempting to wake from hibernate.

Process is normal boot, ubuntu logo boot until "mounting root file system" stage, past this stage it goes to a corrupted graphics screen (alternating brown and black vertical lines) and freezes completely (ctrl-alt-f1 will not drop to a text console).

[Matthew Garrett (mjg59)]

Can you try editing /etc/default/acpi-support and changing ACPI_SLEEP to true?

[mroth (mrothy)]

Behavior is identical after editing ACPI_SLEEP.

[mroth (mrothy)]

Attached screenshot of graphics corruption during freeze upon resume from hibernation.

[Andreas Jobs (andreas-jobs)]

I had the same problem with my acer. After changing
  MODULES=""
to
  MODULES="ehci_hcd uhci_hcd"
in /etc/default/acpi-support sleep and hibernate are working perfectly.

[mroth (mrothy)]

Tried that modification on suggestion of Andreas, still same behavior exhibited.

[Paul Sladen (sladen)]

Andreas, you have a different type of machine and a different issue. Please file a new bug against 'acpi-support' with the details of your laptop.

[Paul Sladen (sladen)]

mroth: is hibernate/suspend still an issue on your laptop?

[KeithCu (keithcu)]

I have a T43 which is quite similar and I'm running fresh flight 6 install + latest fixes and both sleep and hibernate do not work for me. Note I'm not even using the 3-d drivers yet.

sleep: goes to sleep okay, but when it wakes up the screen is hosed and machine is dead

hibernate: hibernates ok (no visual status which is annoying!) and on wakeup it hangs.

[Paul Sladen (sladen)]

Hi Keith, this is specifically about the new batch of Lenovo Pads (which have new hardware that we don't completely support yet). Could you file a separate bug for your T43 not hibernating.

[mroth (mrothy)]

sladen: Yes, bug still exists as of Apr7 nightlies on my hardware.

[Paul Sladen (sladen)]

mroth: could you try editing '/etc/default/acpi-support' and uncommenting:

  # DOUBLE_CONSOLE_SWITCH=true

and try with various combinations of 'SAVE_VBE_STATE' and 'POST_VIDEO'

and if you also have a chance, the effect of booting with the kernel command line set with:

  acpi_sleep=s3_bios
  acpi_sleep=s3_mode
  acpi_sleep=s3_bios,s3_mode

and let us know the results.

[mroth (mrothy)]

All combinations of DOUBLE_CONSOLE_SWITCH, SAVE_VBE_STATE, and POST_VIDEO tested, with no apparent change in behavior.

[Matthew Garrett (mjg59)]

With luck, this will be improved (but not completely fixed) in the next
kernel update. The problem looks like a SATA suspend/resume issue.

--
Matthew Garrett | <email address hidden>

[mroth (mrothy)]

As of kernel -21, hibernate is now mostly functional (sleep still broken with same status as above)

[Ari (ari-reads)]

Is there a chance of a fix for this in Dapper? I hope I can turn my laptop into linux soon! i wouldn't like to wait for edgy for this to work

[Matthew Garrett (mjg59)]

The next kernel release may help improve suspend to RAM support, with a bit of luck.

[Scott Norris (scottie-z)]

I also have an X60. It appears that the devs commenting on this bug are aware of the fact that this is a SATA problem, but I just thought i would post a relevant links. The problem resuming from sleep/suspend, as well as links to various patches, is described here:

http://www.thinkwiki.org/wiki/Problems_with_SATA_and_Linux

Apparently the problem is mostly fixed in the 2.6.16 kernel, but the X60 needs an additional patch, which is also described in the article.

[Matthew Garrett (mjg59)]

I believe we have the appropriate patches in the forthcoming kernel update (should be within a couple of days) - if you could test that and let us know, that would be very helpful.

[Littleiffel (thierry-nicola)]

I have Hibernate/Sleep not working on a x60 Thinkpad. I get the following error message and then jumping back to Ubuntu. The error message is the following

[4299816.456000] swsusp : cannot find swap device : try swapon -a

Coming back to Ubuntu I get a message Hibernate Problem HAL fail to hibernate. Check FAQ page...

The problem is the same if I use hibnate function key Fn + F12 or the Ubuntu Hibernate button(in shutdown menu).

[Matthew Garrett (mjg59)]

You have no swap partition

[Littleiffel (thierry-nicola)]

Well yes I hab no swap partition. But now hibernation works, however I cannot resume.
When trying to hibernate a second time, after a reboot, it is impossible to hibernate. I just come back to Ubuntu

[Matthew Garrett (mjg59)]

Matthew, linux-image-2.6.15-22 is in the archive now - any chance you can give it a go with that?

[redbeard (erictree)]

I just tried using sleep with linux-image-2.6.15-22, and the system
still fails to "wake up".

[Littleiffel (thierry-nicola)]

Hi everyone

Well for me suspend-to-disk works now. It was that i had to install the package 'hibernate' via aptitude. But know it works great(I only had to change the grub menu.lst manually)

Well sleep however still does not work, Computer does not wake up

[Littleiffel (thierry-nicola)]

Wait,....

Sleep is working with me. Just to come back to life, I don't know what I have done but sleep now works for me on an thinkpad x60.

Sorry, the message above I was little too fast..

[Scott Norris (scottie-z)]

I installed -22, and no love here. Will provide any requested message outputs.

[jukka holm (jukka-holm-gmail)]

I am having some success with my thinkpad x41 using -22 kernel from repository and added the ata_piix in /etc/default/acpi-support MODULES -section

[Matthew Garrett (mjg59)]

Jukka,

ata_piix in MODULES should make no difference - the driver will be in use and so can't be unloaded. If you remove that, does it still work?

[jukka holm (jukka-holm-gmail)]

I am having some success with my thinkpad x41 using -22 kernel from repository and added the ata_piix in /etc/default/acpi-support MODULES -section

[Ari (ari-reads)]

What is the current status after the last kernel update? wake up after hibernation is still producing error messages? how about suspend to ram?

[Scott Norris (scottie-z)]

I installed the most recent kernel stuff as of 10 May, 10:30 CDT. I'm
not sure what it all does, but I installed that everything that
gnome-update-manager offered with "linux" in the name -- i.e.,

  linux-386
  linux-image-386
  linux-image-2.6.15-22-386 (.34 version)
  linux-restricted-modules-386
  linux-restricted-modules-common
  linux-restricted-modules-2.6.15-386-22

I have *NOT* kept the rest of the system up to date, as we have a slow
internet connection. I will be happy to apply specific updates if
necessary, and report back.

Results
---------
Fn-F4 now works to put the computer to sleep. However, the computer
still fails to wake up. (Fn-F4 produces HD LED action, but no wake.)

Fn-F12 works to put the computer into hibernate. Upon thawing, we get
to "mounting root file system...", then some vertical brown-and-black
stripes as reported elsewhere, then a shell with a blinking cursor, and
then the successfully reloaded desktop.

Computer can be successfully hibernated/thawed multiple times (well, two
at least).

This may require a separate bug report, but the wired NIC fails to work
after thawing. To reproduce: start computer without internet
connection, then hibernate, then thaw, then 'ifdown eth0'. After this,
'ifup eth0' appears to send out DHCP requests, but appears not to hear
any responses. After a 'real' restart, however, the network can be
successfully set up in this way. I've reported elsewhere (42572) that
the wired NIC failed to be auto-detected on my machine, and required me
to remove and then re-install the e1000 driver. I'll add this comment
to that bug.

[Scott Norris (scottie-z)]

Sorry for the bad formatting -- was trying a direct reply from evolution.

[Matthew Garrett (mjg59)]

Can you try booting with the

pci=nomsi

kernel parameter and see if things work?

[Scott Norris (scottie-z)]

Still fails to wake from sleep with pci=nomsi.

[Chris Howells (chris-chrishowells)]

Kernel -23 breaks resume from hibernate on the X60s. I've gone back to -22.

[Matthew Garrett (mjg59)]

Chris, any chance of you being able to do a git bisect to find out what broke it?

[Matthew Garrett (mjg59)]

(And any idea why it worked for you in -22 and failed for the original reporter?)

[Chris Howells (chris-chrishowells)]

Hi Matthew,

Yes, I'll have a go at working at what caused it.

No idea why it worked for me in -22, other than possibly I'm using the 686 kernel build -- broken in the 386 build for some weird reason? I also went through the BIOS when I got the machine and disabled everything I wasn't using (like parallel port) which I guess can only help to improve things. I also upgraded the machine to have 1.5GB RAM rather than the standard 512MB. I can't imagine that would make any difference, though considering that disabling the modem in the BIOS breaks sound (see bug 42361), I don't think I'd be completely sure.

Oh, I also have BIOS 1.04, it came with 1.03 but I upgraded it when I was having problems with the onboard NIC.

[Ari (ari-reads)]

BIOS 1.05 is available for download. I installed it already.
I just received the machine yesterday. I will try the latest dapper snapshot
as soon as i find the way to install it (i don`t have a cdrom, just a usb
boot disk)

On 5/20/06, Chris Howells <email address hidden> wrote:
>
> Hi Matthew,
>
> Yes, I'll have a go at working at what caused it.
>
> No idea why it worked for me in -22, other than possibly I'm using the
> 686 kernel build -- broken in the 386 build for some weird reason? I
> also went through the BIOS when I got the machine and disabled
> everything I wasn't using (like parallel port) which I guess can only
> help to improve things. I also upgraded the machine to have 1.5GB RAM
> rather than the standard 512MB. I can't imagine that would make any
> difference, though considering that disabling the modem in the BIOS
> breaks sound (see bug 42361), I don't think I'd be completely sure.
>
> Oh, I also have BIOS 1.04, it came with 1.03 but I upgraded it when I
> was having problems with the onboard NIC.
>
> --
> ThinkPad X60 cannot resume from suspend to RAM
> https://launchpad.net/bugs/35174
>

[Korny Sietsma (korny)]

Just to confirm the above - Hibernate was working most of the time with kernel -22, not working at all with -23

It occasionally failed with the previous kernel, though, so I wasn't really relying on it - quite often it would get most of the way out of hibernate, and then stop. Sometimes you could get it to come alive by hitting ctrl-alt-f1 ctrl-alt-f7 and somehow it reloaded the X session and everything worked.

With kernel 23 it seems to freeze at a much earlier stage, and no key combinations seem to help.

[Korny Sietsma (korny)]

oops, I was going to mention - I'm still using the original bios that came with the laptop, and I haven't done any fiddling or disabling settings in the bios.

[Littleiffel (thierry-nicola)]

Hi for me updating to kernel -23 changed nothing with hibernate. "In the most cases" hibernate succeds, but sometimes fails(just as before). Before, using kernel -22, I had different problem to hibernate. With kernel -22 it would often not go to "shut down", and now with kernel -23 it fails to resume, maschine gets blocked with strange screen(See message above "screenshot of graphics corruption Posted by mroth at 2006-04-05 20:34:12 UTC").

[Littleiffel (thierry-nicola)]

Well sorry, I was little too fast yesterday(message above). Hibernate does not work for me on my thinkpad x60. It successfully goes down. When restarting it also resumes(Ubuntu boot sequence is interrupted at Mount Root System). However, I return to new desktop. So finally it is just a reboot.

I really need this to work, so as I am no expert, I don't know what to do. I am willing to help to resolve the problem, just need some instructions

[Luka Renko (lure)]

I would just like to point to bug 45759 and bug 42814 where users also report to get login screen after suspend to RAM and disk.
Not sure if this may be related...

[Littleiffel (thierry-nicola)]

Well sorry, I was little too fast yesterday(message above). Hibernate does not work for me on my thinkpad x60. It successfully goes down. When restarting it also resumes(Ubuntu boot sequence is interrupted at Mount Root System). However, I return to new desktop. So finally it is just a reboot.

I really need this to work, so as I am no expert, I don't know what to do. I am willing to help to resolve the problem, just need some instructions

[James Henstridge (jamesh)]

Matthew: I tried the pci=nomsi option on my X60s using the 2.6.15-23.39 686 kernel. That gives the following output in the dmesg output:

[4294667.296000] Kernel command line: root=/dev/sda3 ro quiet splash pci=nomsi
[4294667.296000] PCI: Unknown option `nomsi'

Later on in the dmesg output I see the following:

[4294668.285000] PCI: Setting latency timer of device 0000:00:1c.0 to 64
[4294668.285000] assign_interrupt_mode Found MSI capability
[4294668.285000] Allocate Port Service[pcie00]
[4294668.285000] Allocate Port Service[pcie02]
[4294668.285000] Allocate Port Service[pcie03]
[4294668.285000] ACPI: PCI Interrupt 0000:00:1c.1[B] -> GSI 21 (level, low) -> IRQ 177
...

It looks like MSI is still enabled with the kernel boot command line option. So MSI might still be the culprit for this bug.

[Matthew Garrett (mjg59)]

Yeah, the information I found about that seems to be slightly bogus. Ok. Can you boot with

acpi_sleep=s3_bios

and see if you get any screen output during the failed resume?

[James Henstridge (jamesh)]

The symptoms I'm seeing are the same both with and without acpi_sleep=s3_bios set:

Screen goes blank, with a bit of disk activity and the sleep light starts flashing maybe twice a second (when running windows, the sleep light would go on solid). In this mode, the fan is still on and I can toggle the capslock and numlock LEDs. I can't resume the system though.

[Matthew Garrett (mjg59)]

Chris - any joy?

[James Henstridge (jamesh)]

Just tried the acpi_sleep=s3_mode kernel boot option, which seems to be a bit more successful.

If I choose suspend from the gnome-power-manager icon, the sleep light flashed as before and finally went on solid.

I was able to resume from sleep here, which brought up the gnome-screensaver unlock dialog and then the logout dialog was displayed after I'd unlocked the screen.

[James Henstridge (jamesh)]

What is weirder is that I seem to be able to suspend without any additional kernel arguments now.

I powered down my system fully for a few minutes and started it again, and things still seemed to be working correctly. I wonder what changed.

[Paul Sladen (sladen)]

James: is this _Suspend_ or _Hibernate_ that you're referring to as now successfully working?

[Paul Sladen (sladen)]

James: Can you file a separate bug for "I was able to resume from sleep here, which brought up the gnome-screensaver unlock dialog and then the logout dialog was displayed after I'd unlocked the screen." against 'gnome-power-manager' and I'll dup it against the right one when I find it.

[James Henstridge (jamesh)]

Paul: I got suspend to ram working, which wasn't working before.

I had another go at hibernate (which I had tried before without success) without success. It would look like it was about to hibernate and then go back to the screensaver. When I looked at the logs, it was complaining that I had no swap partition.

This seemed odd since I had certainly configured one during the install. I tried a few things:
    $ sudo swapon -a
    swapon: /dev/sda2: Invalid argument
    $ sudo file -s /dev/sda2
    /dev/sda2: Linux/i386 swap file (new style) with SWSUSP1 image

I guess this is the left over from my previous attempts to suspend. I was able to enable the swap partition after reformatting it with mkswap and doing "swapon -a".

This time when I tried to hibernate, it wrote stuff to disk and turned the system off. However, when I rebooted, it didn't resume from the hibernation image. The following was in /var/log/messages:

    May 29 20:27:54 morbo kernel: [4294695.451000] Unable to find swap-space signature

At this point, I needed to rerun mkswap before I could enable the swap space again.

As for the logout dialog being displayed on resume from suspend issue, I haven't been able to reproduce it after updating. Do you still want me to report the bug, given that I probably won't be able to help debug the issue?

[James Henstridge (jamesh)]

I did a little more investigation, and I think I've got hibernation working too.

I looked into the initramfs config files in /etc/mkinitramfs, and I noticed that my old laptop had a /etc/mkinitramfs/conf.d/resume file identifying the swap partition while my new one didn't.

After creating the file with the contents "RESUME=/dev/sda2", and recreating the initramfs with "update-initramfs -c -k `uname -r`", I tried rebooting, logging in, and hibernating.

When I booted the machine again, it appears to have resumed from the hibernation image correctly.

So it appears that both suspend and hibernate are working for me now. That still leaves the question of why /etc/mkinitramfs/conf.d/resume didn't get created during the install though (or by any of the updates).

[James Henstridge (jamesh)]

So it looks like the problem I was having with hibernate was bug 42299. From reading that bug, if I did a fresh install from the latest dapper install CDs, the resume partition stuff would be configured correctly.

If anyone else is having trouble with hibernate, they should check if /etc/mkinitramfs/conf.d/resume exists and has the right value.

I did not have to install the "hibernate" package that some other people mentioned earlier in this bug.

[Korny Sietsma (korny)]

Just a note - this doesn't help for me - my /etc/mkinitramfs/conf.d/resume does point to my swap partition (/dev/sda6) and running "update-initramfs -c -k `uname -r`" doesn't seem to help. I still can't resume from hibernate - after a few seconds all hard disk activity stops, the screen fills with vertical stripes, and the machine goes inactive, needing a full power cycle to restart.

However, I did notice that my swap partition /dev/sda6 is 1G but I have 1.5G of Ram, which may be a problem - I can't see how the system would swap out all my ram to swap. This doesn't explain why suspend usually worked with the previous kernel, and never works with this kernel, however.

I do intend to do a clean build when the official Dapper release comes out - if that improves things, I'll report it here.

[Littleiffel (thierry-nicola)]

No go here either... Followed the given procedure, but after resume my session is lost. Everything seems to work fine, but I end up without my session(normal reboot).

TO Korny: If your system is blocked try Crtl+Alt+(F1-F8), had the same problem some time ago.

My swap is larger then my ram, so no problem here. And with the build -22 or -21 hibernate was (sometimes) working.

[James Henstridge (jamesh)]

I actually did the following:
    sudo update-initramfs -d -k `uname -r`
    sudo update-initramfs -c -k `uname -r`

Maybe the delete step is necessary. You can make sure that the config item is in the initramfs with the following command:
    zcat /boot/initrd.img-`uname -r` | cpio -t | grep resume

which should print "conf/conf.d/resume".

[Littleiffel (thierry-nicola)]

zcat /boot/initrd.img-`uname -r` | cpio -t | grep resume

actually prints "conf/conf.d/resume"(+ 32319 blocks)

I tried both commands, but still the resume results in a normal reboot.

[Paul Sladen (sladen)]

LIttleiffel:

  echo /dev/xyz | sudo tee -a /etc/mkinitramfs/conf.d/resume
  sudo update-initramfs -u $(uname -r)
  sudo reboot
  ...
  sudo pmi action hibernate

James: if Suspend now works, we can white-list that. Please can you provide:

  cat /var/lib/acpi-support/*-*

[Littleiffel (thierry-nicola)]

Hi Paul,
Thanks a lot for your help:) I tried your suggestions but the command
sudo update-initramfs -u $(uname -r)

gives the following message:

/etc/mkinitramfs/conf.d/resume: line 2: /dev/sda5: Permission denied

and resuming now hangs on a screen with vertical lines.(I think there is a screenshot attached here)

Nothing happens anymore, no reaction to keypress(CapsLk, Ctrl-Alt-Del,...), but the thinklight is working.(Fn+PgUp)

[James Henstridge (jamesh)]

Littleiffel: the /etc/mkinitramfs/conf.d/resume file should read:
    RESUME=/dev/sda5
If you edit the file to read like that and redo Paul's steps from the second one on, it should work.

Paul: here's the output of that command on my system:
    7BET44WW (1.04 )
    LENOVO
    17024EM
    ThinkPad X60s

[Korny Sietsma (korny)]

I've tried the latest suggestions to no avail.
To summarise:
- I have a 1G swap partition on /dev/sda6 (and 1.5G of ram)
- /etc/mkinitramfs/conf.d/resume contains:
RESUME=/dev/sda6
- zcat /boot/initrd.img-`uname -r` | cpio -t | grep resume returns:
conf/conf.d/resume
32319 blocks
- running "sudo update-initramfs -d -k `uname -r`" followed by "sudo update-initramfs -c -k `uname -r`" does not fix the problem.
- The current symptoms (for me) are that trying to resume from hibernate results in brown streaks down the screen after a couple of seconds, and no other response. When in this mode, ctrl-alt-F1 to F8 make no difference, but NumLock (Fn-ScrLck) toggles the numlock indicator, so the keyboard is at least functional at the BIOS level. The power button only works if I hold it down for 10 seconds though.

in case it is of interest, cat /var/lib/acpi-support/*-* returns:
ZBET43WW (1.03 )
LENOVO
170686M
ThinkPad X60

Hmm - I wonder if this is somewhere where an X60 and an X60s differ? Littleiffel: do you have an X60 or X60s? I understood that the only real difference was the use of a low-powered chipset...

[James Henstridge (jamesh)]

Korny: It seems that you are running an older BIOS than me. It might be worth upgrading and seeing if that helps. There is a 1.06 release up on the website now.

During resume from hibernation, I see the vertical brown lines momentarily, but only momentarily.

[Korny Sietsma (korny)]

Ok, I updated my BIOS to 1.06, cat /var/lib/acpi-support/*-* now returns:
ZBET46WW (1.06 )
LENOVO
170686M
ThinkPad X60

it didn't seem to make any difference however - same symptoms result.
Just for completeness, I also tried with the laptop undocked, and I re-ran update-initramfs again just in case. Still no luck.

When I find the time I'll try to do a full reinstall, and see if that helps - this machine has been continually updated since flight 5, and something may have gone screwy somewhere. But I'd prefer to avoid that option!

[James Henstridge (jamesh)]

Korny: I'm not sure what the problem is. The other difference you've identified is that your swap partition is smaller than your available physical memory.

That could lead to hibernation failures if you had more anonymous pages (i.e. program data as opposed to mapped files or code) allocated and they didn't fit in the swap partition. I don't know whether that'd still be a problem if your system isn't heavily loaded.

[Littleiffel (thierry-nicola)]

I have a X60. My output of cat /var/lib/acpi-support/*-* returns:
7BET43WW (1.03)
LENOVO
170685G
ThinkPad X60

My file /etc/mkinitramfs/conf.d/resume reads:
    RESUME=/dev/sda5

To suspend still works but resuming hanks on the screen with the lines. I will try a bios update today and see it that helps.

[Paul Sladen (sladen)]

Thanks for the DMI data; regarding hibernate, please check that:

  /etc/mkinitramfs/conf.d/resume

contains a string:

  RESUME=/dev/sdXYZ

that is the same device as running:

  $ sudo fdisk -l /dev/sda | grep swap

lists. Please make sure that these match. Then, run:

  $ sudo update-initramfs -u $(uname -r)

this still recreate the initramfs (startup files) for the running kernel and set that configuration option in it. To check that this has happened, please run:

  $ zcat /boot/initrd.img-$(uname -r) | strings | grep -a RESUME

[Littleiffel (thierry-nicola)]

Okay the command "zcat /boot/initrd.img-$(uname -r) | strings | grep -a RESUME" prints the following result:

RESUME=/dev/sda5
#RESUME=
<6>cdrom: entering CDROMPAUSE/CDROMRESUME
KEY_RESUME
export resume=${RESUME}

I tried then to hibernate with the command(after reboot) "sudo pmi action hibernate", and with the logUp Applet of Ubuntu.

The command "sudo pmi action hibernate" does now fail to suspend. This is strange then suspending was working all the time now?(I just made the updates from Ubuntu??) I waited for about 5 minutes and nothing happend. Maybe Is hould wait longer next time.
With the LogOut Applet the suspend works?!? Now I am really confused, and I will retry with the command.
As several times before resuming results in a normal reboot.

Are there some configuration files I can post(which are of interest)??

[julien (julien-ubuntu)]

I also have a ThinkPad X60

I added acpi_sleep=s3_bios to my kernel parameters
I changed /etc/default/acpi-support to set ACPI_SLEEP to true commented VBE and REPOST.

Now:

Suspend to ram
---------------------

- Fn+F4 does'nt work: according to /var/log/acpid, /etc/acpi/sleepbtn.sh should be launched but nothing seem to happen
- If I manually echo mem > /sys/power/state from gterm on gnome, the X60 suspends to mem. However when resuming, X restarts. If I do the same from a console (while X launched) I can go to X with alt-F7 properly.
When X crashed, I can see "Error in I830WaitLpRing() now is....." follow by "fatal server error: lockup"

Suspend to disk
---------------------

Suspend works either with Fn+F12 or echo disk > /sys/power/state
When resuming I get the strange screen documented by mroth above.

I recompiled a 2.6.17-rc5 manually, resuming from suspend to disk almost works but the SATA drive seems to sleep

[Ari (ari-reads)]

Hi,

    I have done a fresh new install of the new official 6.06 release on my thinkpad x60s

    suspend to disk (Fn12) worked out of the box. I have a 1GB swap, and 512MB of RAM.

    suspend to ram (Fn4) is not working at all, nothing happens when I press Fn4. I also edited acpi-support in the same way as Julien. Is there any extra step I am missing for Fn4 to work??

  Also, this is a bit off-topic, but since we are all x60 users, If anyone else is unhappy with the battery drain problem (cpu frequency scaler not going below 1Ghz) and has found a solution/thread... please send me an email :-)

Thanks

[julien (julien-ubuntu)]

Update to my previous post: Ari is right, suspend to disk actually works after default installation.

But I installed the SMP i386 kernel and now I have the strange screen after resume.

[James Henstridge (jamesh)]

If you can't get suspend (Fn+F4) working, go to System -> Preferences -> Power Management. On the "General" page of the notebook, make sure that "Sleep button action" is set to "Suspend".

[julien (julien-ubuntu)]

Thanks james. I had just figured out the same by looking at the sleep.sh script which obviously will let gnome-power-manager handle everything.

Btw runing all the scripts in /etc/acpi/suspend.d obviously helps because I don't have the X server crash anymore.

So my only remaining problem is with suspend to disk which only resumes properly sometimes with i386 kernel and never with i686 SMP kernel. When it does'nt work I get a screen similar to this: http://librarian.launchpad.net/1943585/x60s-wakefromhibernatecorruption.jpg

[Martin Aumueller (aumuell)]

Resuming from suspend to RAM works for me on my X60s with Hitachi HTE721010G9SA00 100GB/7200 rpm drive when I switch from "compatibility" to "ahci" mode in the BIOS settings. It's a 1702-55G with 2 GB and BIOS 1.06. However, power consumption has risen by 1 - 2 W since I changed to ahci.

[Korny Sietsma (korny)]

I have now done a clean install from the released CDs (alternate-install), with a 2G swap partition - to no avail.
With kernel 2.6.15.23-686 I still get the same symptoms as before - suspend seems to work, resume gets part way then gives the vertical bars described above.
Using the 386 boot doesn't seem to help - I also get the vertical lines of doom.
I'm going to try the instructions at https://wiki.ubuntu.com/LaptopTestingTeam/ThinkpadX60s for sleep mode, to see if they help; but this is definitely not fixed for me.

[Littleiffel (thierry-nicola)]

I made a fresh install of Dapper, with the official release on my X60.

And Hibernate does not work out of the box(nor does suspend-to-ram). I have read again through the whole topic here, and tried the suggestions, but still no go.

I added resume=/dev/sda4 to boot parameters in /boot/grub/menu.lst

I verified that /etc/mkinitramfs/conf.d/resume reads:
    RESUME=/dev/sda4

and the command "zcat /boot/initrd.img-$(uname -r) | strings | grep -a RESUME" prints the following result:

RESUME=/dev/sda4
#RESUME=
<6>cdrom: entering CDROMPAUSE/CDROMRESUME
KEY_RESUME
export resume=${RESUME}

I have not installed any additional package(concerning hibernate, suspend2,...). I know that hibernate for x60 works out of the box with suse 10.1. But I want to stay with Ubuntu.

[julien (julien-ubuntu)]

I have a regression here with kernel 2.6.15-25.

With kernel 2.6.15-23

- suspend to ram and resume worked while using acpi_sleep=s3_bios kernel boot option.
- suspend to disk seems to works, resume gives the strange screen described above

With kernel 2.6.15-25 I can't resume even from suspend to RAM. It looks like the X60 does'nt even sleep properly since the fan continues is still running.
Moreover if I try to wake it up by pressing the power button or Fn+F4 nothing happens

[Henrik Karlsson (hkarlsson)]

I confirm, I have the same regression on my T60.

[Korny Sietsma (korny)]

Same here - suspend to ram (which used to work with the s3_bios setting) no longer works.

Hibernate, which used to (apparently) hibernate OK, but failed to resume, now doesn't even hibernate properly - the fan doesn't turn off, and the moon-shaped "sleep" light flashes continually.

Also, after power cycling during a failed hibernate, the machine no longer boots with that kernel (it stalls on "Mounting root file system...") - I assume it's looking for the hibernate files or something. I had to select an older kernel before I could boot.

[Korny Sietsma (korny)]

I now have a secondary problem - whenever I boot with the kernel 2.6.15-25 I get a very long (about 3 minutes) delay at the "Mounting root file system" message. Eventually it recovers and boots normally, but every time I boot I have this insanely long delay.
I've tried re-running "sudo update-initramfs -u $(uname -r)" to see if this helps, but to no avail.
Anyone have any other suggestions? This is slightly inconvenient, especially now that sleep mode doesn't work... :-}

[Matthew Garrett (mjg59)]

acpi-support-0.85 (in dapper updates) should fix the regression.

[Alex Bligh (ubuntu-alex-org)]

I've just upgraded to acpi-support-0.85 an it still hangs for a couple of minutes at "Mounting root file system"

[Matthew Garrett (mjg59)]

Can you boot in recovery mode and see what is printed on screen while
it's hung?

--
Matthew Garrett | <email address hidden>

[Alex Bligh (ubuntu-alex-org)]

Hope the following is sufficient (there is a short hang and a long hang). If not I have the complete dmesg.

Alex

File Edit Options Buffers Tools Help
[17179570.640000] ACPI: CPU1 (power states: C1[C1] C2[C2] C3[C3])
[17179570.640000] ACPI: Processor [CPU1] (supports 8 throttling states)
[17179570.644000] ACPI: Thermal Zone [THM0] (61 C)
[17179570.644000] ACPI: Thermal Zone [THM1] (61 C)
[17179570.956000] ICH7: IDE controller at PCI slot 0000:00:1f.1
[17179570.956000] ACPI: PCI Interrupt 0000:00:1f.1[C] -> GSI 16 (level, low) -> IRQ 169
[17179570.956000] ICH7: chipset revision 2
[17179570.956000] ICH7: not 100% native mode: will probe irqs later
[17179570.956000] ide0: BM-DMA at 0x1880-0x1887, BIOS settings: hda:DMA, hdb:pio
[17179570.956000] Probing IDE interface ide0...

SHORT HANG HERE

[17179576.424000] hda: HL-DT-ST DVDRAM GSA-4083N, ATAPI CD/DVD-ROM drive
[17179578.888000] ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
[17179578.896000] hda: ATAPI 24X DVD-ROM DVD-R-RAM CD-R/RW drive, 2048kB Cache, UDMA(33)
[17179578.896000] Uniform CD-ROM driver Revision: 3.20
[17179579.184000] SCSI subsystem initialized
[17179579.188000] ACPI: bus type scsi registered
[17179579.188000] libata version 1.20 loaded.
[17179579.188000] ahci 0000:00:1f.2: version 1.2
[17179579.188000] ACPI: PCI Interrupt 0000:00:1f.2[B] -> GSI 16 (level, low) -> IRQ 169

LONG HANG HERE

[17179666.844000] PCI: Setting latency timer of device 0000:00:1f.2 to 64
[17179666.844000] ahci 0000:00:1f.2: AHCI 0001.0100 32 slots 4 ports 1.5 Gbps 0x1 impl SATA mode
[17179666.844000] ahci 0000:00:1f.2: flags: 64bit ncq pm led clo pio slum part
[17179666.844000] ata1: SATA max UDMA/133 cmd 0xF884C500 ctl 0x0 bmdma 0x0 irq 66
[17179666.844000] ata2: SATA max UDMA/133 cmd 0xF884C580 ctl 0x0 bmdma 0x0 irq 66
[17179666.844000] ata3: SATA max UDMA/133 cmd 0xF884C600 ctl 0x0 bmdma 0x0 irq 66
[17179666.844000] ata4: SATA max UDMA/133 cmd 0xF884C680 ctl 0x0 bmdma 0x0 irq 66

[Alex Bligh (ubuntu-alex-org)]

And I should add that sound has mysteriously stopped working, and I'm using a T60p not an X60 (fair game I think though given you marked my T60p as a duplicate of this one).

[Paul Sladen (sladen)]

The IDE hang appears to be something relating to IDE probing; do you know approximately when this delay started appearing? I don't know what was changed in the kernel.

Sound maybe related to the Intel HDA that is in the new LenovoPads and the various workarounds to enable that to work. But that workarounds have been designed to /fix/ machines, rather than break them.

Can you keep the audio issue to a separate bug report (against 'linux-source-2.6.15' and assigned to the 'kernel-audio-team') and paste the DMI and lspci data there. On of the things to check is to fire up 'alsa-mixer' from the command-line and see if there are any 'jack-detect' flags that can be toggled (press the 'm' key to 'mute' the toggle).

Sorry if I've dup'ed one of those pair of laptops---they're both LenovoPads and basically the same inside (as far as we understand) but I may just have seen that they were both from the same person and dup'ed them. Please undup them if you think that they are more likely to be unrelated.

[Henrik Karlsson (hkarlsson)]

* After upgrading to acpi-support 0.85, my T60 now enters full sleep mode (fans, display and HDD turned off). However, when waking it up, everything seem to come back to life except the display. I can see the display backlight is turned on but the screen remains black.
* Still have the ~60s hang when "Mounting root file system..."
* Sound works fine (never had any problems with that).

[Alex Bligh (ubuntu-alex-org)]

Paul,

I'm not sure it was actually you that marked this as a dupe and I any case I don't think it's wrong so no need to apologize.

The boot hang problem appeared at some point within the last 2 weeks (sorry not to be more specific, I update very frequently but never reboot) - I am pretty sure it was after 31 May in any case.

I will report the sound problem separately.

Alex

[Matthew Garrett (mjg59)]

Both the hangs you've highlighted should come before the "Mounting the
root filesystem" line, so I'm a bit confused. I'll look into it.

--
Matthew Garrett | <email address hidden>

[Paul Sladen (sladen)]

Henrik: do you have Intel graphics, or something else? I think based on the reports we can generalise that all the LenovoPads are affected regarding the IDE scanning hang.

[Henrik Karlsson (hkarlsson)]

Paul: yes, Intel graphics.

[julien (julien-ubuntu)]

For your information on 2.6.15-25 suspend and hibernate will not work on a (older) thinkpad T40p either
However it works perfectly with 2.6.15-23

[Alex Bligh (ubuntu-alex-org)]

A T40p and a T60p/X60 are very different machines as far as I can tell.

[Matthew Garrett (mjg59)]

julien,

Have you updated to the latest acpi-support package (0.85)?

If people are still having problems with slow "mounting root filesystem" lines, please file another bug.

[Korny Sietsma (korny)]

Updating to acpi-support 0.85 seems to have mostly fixed both suspend and sleep, for me.

Also, having done a successful hibernate-resume seems to *sometimes* fix the slowdown at "mounting the root file system" - sometimes it seems to pause for 20 seconds or so at this point, sometimes it does the full 3 minute wait.

I say "mostly" because the first time I tried it, it didn't work - I'm now trying to narrow down what might have caused it. I suspect it's because I sometimes boot with the wifi/bluetooth hardware switch turned off, and then I turn it when I need it - this seems to sometimes confuse things.

But if I leave this switch off or leave it on, I seem to now consistently sleep and hibernate with no problems - thanks people! I hope it works for everyone else...

[Littleiffel (thierry-nicola)]

Okay, I still got no go for hibernate. And as you report your system suspends and hibernates, well I would be interested in the actions/steps you have done to get it working.

 - Have you change configurations manually?
 - Installed additional packages?
 - Changed the bootparameters?

Could you please write what you have done since the installation, this would be very helpful for me.

Thank you

[Littleiffel (thierry-nicola)]

Well I tried sleep an it seems to work. Before the screensaver comes back I get a message saying something like 'nsc-ircc, Wrong chip version ff'

But to that I report Sleep working on my x60:)

[Littleiffel (thierry-nicola)]

iSorry for the messages, this one is the last for today...

Hibernate is also working(suspending and getting session back works). However my wired network is down(after hibernate) and I am using the NetworkManager which does not work either after coming up again.

The NetworkManger is not working because hibernate changes the /etc/network/interfaces file.(To get NM working I had to comment out all the devices except lo)

And now the file /etc/network/interfaces contains an entry, and therefore NM is not working after coming up.

So does hibernate process alter the /etc/network/interfaces file???

[Matthew Garrett (mjg59)]

> So does hibernate process alter the /etc/network/interfaces file???

No. I don't know what's going on there...

[Littleiffel (thierry-nicola)]

Tested several times and it only happened once that the file /etc/network/interfaces was changed. However sometimes coming back from hibernate and sleep the NetworkManager has kind of problems.

However, I think these problems might be related to the NetworkManager and not to Hibernate/sleep.

And, all the network devices work properly when coming back to the session. I simply use the Network Menu to configure my wireless device.

So for me I am happy that Sleep and Hibernate are working(until now all the time). Except the first time, maybe one has to reboot after upgrading acpi-support?(As Korny reported a failure for his first try too)

So I repeat what I have done to get it working:
-upgrade to acpi-support 0.85
- edit /boot/grub/menu.lst and add resume=/dev/sda4 as a bootparameter.

/dev/sda4 should be your swap partition.

Thanks to all who helped (me) get this to work properly:)

[Stefan Wehr (mail-stefanwehr)]

I have the same problem that Alex with hangs during bootup (when coming from hibernation but also after normal reboot). I opened a new bug #50808.

[Littleiffel (thierry-nicola)]

I upgraded my kernel to 2.6.15-25-686 and now Resume/hibernate is not working anymore, (while it worked perfectly with the -23 kernel)

 -Now when trying to suspend, it is not going to sleep and it is impossible to wake it up again.
 - Hibernate, suspends without a problem however resuming hangs with the usual screen.

As I still have the old kernels, I tested them too, Hibernate does not work on any kernel (2.6.15-23-386, 2.6.15.-23-686, 2.6.15-25-386) Always the same hang.

However, Sleep is still working perfectly with the kernels (2.6.15-23-386, 2.6.15.-23-686).

So as before the update/upgrade hibernate and sleep was working I wonder why these kernels are not working anymore??

What I did, I install package linux-686-smp(apt-get install linux-686-smp), and then through the update-manager installed the -25 kernel(..and some other stuff, tetex,...)

And I also obtained the hang with kernel -25.

Bye

[Littleiffel (thierry-nicola)]

Sometimes hibernate is working and sometimes not(most cases)???

Now I have the problem that I will not go down when hibernating and not coming up from sleep.

My /var/log/syslog, gives the follwoing when hibernate does not go down:
Jun 28 09:46:38 localhost gnome-power-manager: Hibernating computer because the DBUS method Hibernate() was invoked
Jun 28 09:46:38 localhost NetworkManager: <information>^IGoing to sleep.
Jun 28 09:46:38 localhost dhclient: receive_packet failed on eth1: Network is down
Jun 28 09:46:38 localhost NetworkManager: <information>^Imatch
Jun 28 09:46:38 localhost last message repeated 3 times
Jun 28 09:46:39 localhost dhclient: DHCPRELEASE on eth0 to 134.21.198.4 port 67
Jun 28 09:46:39 localhost dhclient: send_packet: Network is unreachable
Jun 28 09:46:39 localhost dhclient: send_packet: please consult README file regarding broadcast address.
Jun 28 09:46:39 localhost kernel: [17183281.288000] ADDRCONF(NETDEV_UP): eth0: link is not ready
Jun 28 09:46:39 localhost kernel: [17183281.708000] ACPI: PCI interrupt for device 0000:03:00.0 disabled
Jun 28 09:46:39 localhost kernel: [17183281.712000] ieee80211_crypt: unregistered algorithm 'NULL'
Jun 28 09:46:39 localhost kernel: [17183281.716000] ieee80211_1_1_13_crypt: unregistered algorithm 'NULL'
Jun 28 09:46:39 localhost dhclient: Internet Systems Consortium DHCP Client V3.0.3
Jun 28 09:46:39 localhost dhclient: Copyright 2004-2005 Internet Systems Consortium.
Jun 28 09:46:39 localhost dhclient: All rights reserved.
Jun 28 09:46:39 localhost dhclient: For info, please visit http://www.isc.org/products/DHCP
Jun 28 09:46:39 localhost dhclient:
Jun 28 09:46:39 localhost kernel: [17183281.816000] ACPI: PCI interrupt for device 0000:02:00.0 disabled
Jun 28 09:46:39 localhost dhclient: Bind socket to interface: No such device
Jun 28 09:46:42 localhost kernel: [17183284.964000] Freezing cpus ...
Jun 28 09:46:43 localhost kernel: [17183285.080000] CPU 1 is now offline
Jun 28 09:46:43 localhost kernel: [17183285.080000] SMP alternatives: switching to UP code
Jun 28 09:47:04 localhost NetworkManager: <information>^IWaking up from sleep.
Jun 28 09:47:04 localhost NetworkManager: <information>^IDeactivating device eth1.
Jun 28 09:47:04 localhost NetworkManager: <information>^IDeactivating device eth0.
Jun 28 09:47:06 localhost NetworkManager: <WARNING>^I nm_device_802_11_wireless_get_mode (): error getting card mode on eth0: No such device

So my question is whether the problem may be related to the Network Manager??

[Matthew Garrett (mjg59)]

Upgrade to the acpi-support in dapper-updates

[Littleiffel (thierry-nicola)]

I think I have already the latest version installed of acpi-support. My version is actually 0.85 Is this the latest?

[Gregory Oschwald (osch0001)]

Suspend to RAM properly restores my session about once every eight tries on my new X60s. Most of the time it doesn't make it to the point where the LCD is turned on. I have tried both 2.6.15-25.4 from Dapper and 2.6.17-4.5 from Edgy. Hibernate is more reliable. I have acpi-support 0.85 installed.

My HD is a FUJITSU MHV2100B.

I will attach some relevant log snippets that include the time from when I suspend the machine to the point where I get sick of waiting for it to resume and force a power down.

I also experience the hang on boot, but I will comment on that bug separately.

[Gregory Oschwald (osch0001)]

See comment below.

[Gregory Oschwald (osch0001)]

Please contact me if more of these would be useful.

[Gregory Oschwald (osch0001)]

After testing it several times, resuming from suspend seems to work well for my X60s after upgrading to 2.6.15-26.

[ZZZ (kgad)]

But the current version of Ubuntu Dapper kernel is 2.6.15-23-686, still far from 2.6.15-26. So even if does work on 2.6.15-26, it doesn't help.

[Ben Collins (ben-collins)]

No, the current version in Ubuntu dapper-security is -26. Which will be included in the first point release.

I'll close the bug, thanks.

[ZZZ (kgad)]

I think you should wait before closing the bug so soon. Testing several times is not enough. I would wait for 2.6.15-26 to be updated, and then hear the users feedback. If changing to 2.6.15-26 helps, you'll know it.

[Alex Bligh (ubuntu-alex-org)]

This is NOT fixed at least on a T60p (my T60p bug was duped to this one), with Dapper security updates including 2.6.15-26.

On hitting suspend, the flashing "moon" indicator appears, and the system appears to suspend, with the LCD switching off. However, it soon switches back on, and goes to the Ubuntu login prompt (as if X has been restarted). Then doing anything (such as a VT switch) repeatedly restarts X in a similar way, so you always end up with a pause, the drum-roll, and the Ubuntu login screen.

[Alex Bligh (ubuntu-alex-org)]

For completeness, hibernate doesn't work either. The same thing happens, pretty much exactly.

[Gregory Oschwald (osch0001)]

Alex,

That sounds like a different bug. This bug was about failure to resume
from a successful suspend, while yours seems to be about a failure
suspending.

Greg

[ZZZ (kgad)]

I updated to the new kernel. I didn't try enough times to conclude wheather the bug is solved.

In any case, if we suppose it is solved, you basically switched this bug with another one: https://launchpad.net/bugs/50313 .

Now it takes me almost 10 minutes to open the laptop (when I show it to someone that has Windows, he laughs at me...). This bug did not exist in the old kernel version.

[ZZZ (kgad)]

The bug is NOT solved !
Once in a while my ThinkPad cannot resume from suspend.
It happens once every 4-5 times.

[Gregory Oschwald (osch0001)]

Although I have not fully diagnosed it, I am have some occasional
problems with suspend as well, but I am always able to get the laptop to
resume.

The problem seems to crop up when I change power states during the
suspend (e.g., I suspend it on AC and try to resume it on battery
power). Sometimes when doing this, it seems to get stuck resuming
before the LCD is turned on. This can be solved by cycling between AC
and battery. Often it starts up immediately during the first state
change, but sometimes it takes several tries and a bit of waiting.

Upon resume from these suspends, Network Manager seems to be broken, and
I need to restart dbus to get it working again, which isn't the case for
normal suspends but is the case when I hibernate the machine.

As the symptoms of this are significantly different than my earlier
resume problems, I think this should probably be filed as another bug.
ZZZ, can you get your machine to resume following this method?

[ZZZ (kgad)]

No, when it's stuck it's stuck.
It happens on both AC and battery, and changing this after it's stuck doesn't help.

[Gregory Oschwald (osch0001)]

ZZZ, I don't seem to experience that problem on my X60s. Given that I experienced bug 35174, perhaps it would be appropriate to file another bug with the details of your problem and hardware so that the information above does not confuse people.

Also, I want to note that my bug happens on both battery and AC, but only after I switch from one to another.

[ZZZ (kgad)]

Why open a new bug? It's the same bug.
I can copy the first comment of this bug:

IBM ThinkPad X60s
Sleep: Will not wake up. Button presses have no affect on system. Power button appears to rev up the fan for a few seconds, but system remains off and sleep LED on hardware remains lit.

This exactly what happens in my machine. Though it doesn't happen every time: sometimes it resumes and sometimes not.

Do you think I should copy these lines to a new bug?

[Loïc Minier (lool)]

I've just tested the newest acpi-support (under Debian though), and it indeed fixes the video for suspend to RAM under Xorg + fglrx.

However, I still a black screen with suspend to disk, which I don't when I s2both manually.

I also confirm that the freeze is libata related as I get warnings on console from libata, various timeouts occur, but sda doesn't come back up.

[ZZZ (kgad)]

I upgraded to ubuntu edgy (acpi-support Installed: 0.90). The suspend seems to be working better, but still once in a while system crushes. Although the frequency is lower - it still happens (once in every ~5 sleeps).