[SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

yourself · 2012-07-02 19:25:04

Hello,

I'm running arch on a T420 bare (without optimus, just intel hd3000) and after upgrading to 3.4.x kernel I have found that some times, when I close the lid instead of propely going to sleep, my laptop enters a weird mode (the sleep 'moon' led is constantly flashing) and if I open the lid linux is unusable, I can only press the power button for a proper poweroff (most of the times it has dropped to my display manager). If I browse through my messages.log I can clearly see that on the weird times, there are no 'CPUx is now offline' messages and the laptop consumes battery quite fast (compared to sleeping).

I have searched for something resembling this but found nothing. Does this sound familiar to someone? Any ideas?

I will provide any info you think is necessary!!

Thanks!

Update: Most of our issues were due to misbehaving synaptics driver and specifically xf86-input-synaptics-1.6.2-1. After downgrading to 1.5.99.902-1, almost(*) all issues went away.

(*) There was one report that this did not entirely fix the issue, but made the symptoms appear more rarely.

Last edited by yourself (2012-09-20 15:57:46)

root · 2012-07-02 19:50:35

I had a similar issue also taking place occasionally when I tried to suspend my T23. Instead of actually suspending it just showed the slimlock window and the moon led flashing as you mention.

In my case I fixed it removing the 'autodetect' hook as I had also the 'Blank screen issue'.

Last edited by root (2012-07-06 21:24:17)

yourself · 2012-07-03 05:02:19

Thanks, I will try that and report back..!

yourself · 2012-07-04 16:15:52

No luck, same thing happened again.
Any other ideas?

yourself · 2012-07-05 11:09:12

So, I have tried to narrow the problem down and collected a bit of info. The same thing (exactly) happens with uswsusp.
A normal sleep procedure goes like this (dmesg output):

[ 3671.683523] PM: Syncing filesystems ... done.
[ 3671.685784] PM: Preparing system for mem sleep
[ 3671.754668] Freezing user space processes ... (elapsed 0.01 seconds) done.
[ 3671.768015] Freezing remaining freezable tasks ... (elapsed 0.01 seconds) done.
[ 3671.781334] PM: Entering mem sleep
[ 3671.781384] Suspending console(s) (use no_console_suspend to debug)
[ 3671.794663] ehci_hcd 0000:00:1a.0: power state changed by ACPI to D0
[ 3671.807842] ehci_hcd 0000:00:1a.0: power state changed by ACPI to D0
[ 3671.807900] ehci_hcd 0000:00:1a.0: setting latency timer to 64
[ 3671.807923] ehci_hcd 0000:00:1d.0: power state changed by ACPI to D0
[ 3671.821180] ehci_hcd 0000:00:1d.0: power state changed by ACPI to D0
[ 3671.821272] ehci_hcd 0000:00:1d.0: setting latency timer to 64
[ 3671.995975] sd 0:0:0:0: [sda] Synchronizing SCSI cache
[ 3671.996110] sd 0:0:0:0: [sda] Stopping disk
[ 3672.247087] i915 0000:00:02.0: power state changed by ACPI to D3
[ 3672.275044] e1000e 0000:00:19.0: wake-up capability enabled by ACPI
[ 3672.313866] PM: suspend of devices complete after 493.197 msecs
[ 3672.314001] PM: late suspend of devices complete after 0.131 msecs
[ 3672.327196] ehci_hcd 0000:00:1d.0: wake-up capability enabled by ACPI
[ 3672.340396] ehci_hcd 0000:00:1d.0: power state changed by ACPI to D3
[ 3672.340646] ehci_hcd 0000:00:1a.0: wake-up capability enabled by ACPI
[ 3672.353709] ehci_hcd 0000:00:1a.0: power state changed by ACPI to D3
[ 3672.367103] PM: noirq suspend of devices complete after 53.177 msecs
[ 3672.367289] ACPI: Preparing to enter system sleep state S3
[ 3672.510204] PM: Saving platform NVS memory
[ 3672.513482] Disabling non-boot CPUs ...
[ 3672.616612] CPU 1 is now offline
[ 3672.618229] CPU 2 is now offline
[ 3672.620042] CPU 3 is now offline

Whereas the borked sleep goes like this:

[ 5911.037381] PM: Syncing filesystems ... done.
[ 5911.039871] PM: Preparing system for mem sleep
[ 5911.415413] BUG: Bad rss-counter state mm:ffff880116bea300 idx:1 val:-1
[ 5911.415417] BUG: Bad rss-counter state mm:ffff880116bea300 idx:2 val:1
[ 5911.463208] BUG: Bad rss-counter state mm:ffff88010f46f380 idx:1 val:-3
[ 5911.463212] BUG: Bad rss-counter state mm:ffff88010f46f380 idx:2 val:3

and stays there (no more dmesg output)

root · 2012-07-06 21:26:51

Just to be sure, you built a new image after removing the hook, didn't you?

yourself · 2012-07-08 21:31:19

Yes, of course. And just to be sure,

# mkinitcpio -p linux

(Did I need to do something more?)

I'm going a bit crazy here. Maybe it's a kernel bug related to 3.4.x and will be fixed in 3.5?

yourself · 2012-07-08 22:01:20

Hm.....
My problem seems very much like this one here, only it is not solved for me

https://bbs.archlinux.org/viewtopic.php?pid=1127031

Hspasta · 2012-07-09 05:55:47

I have the exact same laptop and I have not ever had any suspending issues since linux 3.x. I'm not sure what could be the difference in our systems that causes yours to fail. I am using a very minimal mkinitcpio build.

MODULES="ahci sd_mod ext4"
BINARIES="fsck fsck.ext4"
HOOKS="base"

I don't know if its really relevant. (or helpful)

yourself · 2012-07-09 06:19:22

Well, I'm using a more or less standard mkinitcpio.conf:

MODULES="i915"
BINARIES=""
FILES=""
HOOKS="base udev autodetect pata scsi sata uresume filesystems usbinput fsck"

The only difference from the "reference" in Arch being that I added "i915" (as desribed here: https://wiki.archlinux.org/index.php/Intel ) and uresume in HOOKS (as described here: https://wiki.archlinux.org/index.php/Uswsusp )

The worst thing about this is that I don't have a clue on where to look..!

I'll try your bare minimum and see what happens....

TheFiddler · 2012-07-10 22:41:47

Are you using laptop-mode-tools? Try disabling them.

yourself · 2012-07-11 06:30:11

The bare minimum mkinitcpio.conf did not really help.... :-(

I'll try TheFiddler's suggestion and disable laptop-mode-tools and report back..

yourself · 2012-07-11 07:41:36

Furthermore, I tried adding thinkpad_acpi to SUSPEND_MODULES but this particular module cannot be unloaded because 'it is in use' so I have blacklisted it along with disabling laptop-mode-tools and will report my results....

msx · 2012-07-11 08:29:36

Hi, have you tried suspending from CLI with pm-suspend and playing with its quircks options? Also, did you have install laptop-mode-tools package and if so did you tried disabling the package and then suspend the system?

On the other hand, did have you already check your /var/log files for any clue? One way to check them quickly is to wipe'em out (or the smartet choice, just move them to another directory), then delete dmesg log ($ dmesg -c) and proceed to suspend your notebook and check if there's any new error log generated.

Last edited by msx (2012-07-11 10:54:58)

yourself · 2012-07-11 09:31:56

Hey msx!

No, I haven't tried running pm-suspend from command line but in theory this is what gets executed, isn't it?

It's true, I have had laptop-mode-tools installed (but never had any problems before 3.4.x kernel) but I have disabled them, for now.

I have checked /var/log/pm-suspend.log (nothing there), the last /var/log/messages state:

Jul 11 08:42:31 localhost logger: LID closed
Jul 11 08:42:31 localhost laptop-mode: Warning: Configuration file /etc/laptop-mode/conf.d/board-specific/*.conf is not 
readable, skipping.
Jul 11 08:42:31 localhost laptop-mode: Laptop mode 
Jul 11 08:42:31 localhost laptop-mode: enabled, 
Jul 11 08:42:31 localhost laptop-mode: active [unchanged]
Jul 11 08:42:32 localhost NetworkManager[580]: <info> sleep requested (sleeping: no  enabled: yes)
Jul 11 08:42:32 localhost NetworkManager[580]: <info> sleeping or disabling...
Jul 11 08:42:32 localhost NetworkManager[580]: <info> (eth0): now unmanaged
Jul 11 08:42:32 localhost NetworkManager[580]: <info> (eth0): device state change: unavailable -> unmanaged (reason 'sle
eping') [20 10 37]
Jul 11 08:42:32 localhost NetworkManager[580]: <info> (eth0): cleaning up...
Jul 11 08:42:32 localhost NetworkManager[580]: <info> (eth0): taking down device.
Jul 11 08:42:32 localhost NetworkManager[580]: <info> (wlan0): now unmanaged
Jul 11 08:42:32 localhost NetworkManager[580]: <info> (wlan0): device state change: activated -> unmanaged (reason 'sleeping') [100 10 37]
Jul 11 08:42:32 localhost NetworkManager[580]: <info> (wlan0): deactivating device (reason 'sleeping') [37]
Jul 11 08:42:32 localhost dhcpcd[12815]: received SIGTERM, stopping
Jul 11 08:42:32 localhost dhcpcd[12815]: wlan0: removing interface
Jul 11 08:42:32 localhost NetworkManager[580]: <info> (wlan0): canceled DHCP transaction, DHCP client pid 12815
Jul 11 08:42:32 localhost kernel: [ 8419.684144] cfg80211: Calling CRDA to update world regulatory domain
Jul 11 08:42:32 localhost NetworkManager[580]: <info> (wlan0): cleaning up...
Jul 11 08:42:32 localhost NetworkManager[580]: <info> (wlan0): taking down device.
Jul 11 08:42:32 localhost dbus[372]: [system] Activating service name='org.freedesktop.nm_dispatcher' (using servicehelper)
Jul 11 08:42:32 localhost NetworkManager[580]: <warn> (pid 12815) unhandled DHCP event for interface wlan0
Jul 11 08:42:32 localhost dbus[372]: [system] Successfully activated service 'org.freedesktop.nm_dispatcher'
Jul 11 08:42:33 localhost kernel: [ 8420.217456] EXT4-fs (sda1): re-mounted. Opts: discard,barrier=0,commit=0
Jul 11 08:42:33 localhost NetworkManager[580]: <warn> error requesting auth for org.freedesktop.NetworkManager.wifi.share.open: (3) GDBus.Error:org.freedesktop.DBus.Error.NameHasNoOwner: GDBus.Error:org.freedesktop.DBus.Error.NameHasNoOwner: Could not get UID of name ':1.19': no such name
Jul 11 08:42:50 localhost acpid: client 586[0:0] has disconnected
Jul 11 08:42:50 localhost logger: LID opened

(I had not yet disabled laptom-mode-tools) but as you can see, 20 seconds after closing the lid not a hint about suspending (of course, the sleep led was flashing).... Something is really fishy here..

On the other hand, /var/log/pm-suspend.log contains just the last RESUME messages for some reason. Something overwrites it. Who is responsible for writing this file?

disabled, not active

/usr/lib/pm-utils/sleep.d/01laptop-mode resume suspend: success.
Running hook /usr/lib/pm-utils/sleep.d/01grub resume suspend:

/usr/lib/pm-utils/sleep.d/01grub resume suspend: success.
Running hook /usr/lib/pm-utils/sleep.d/00powersave resume suspend:

/usr/lib/pm-utils/sleep.d/00powersave resume suspend: success.
Running hook /usr/lib/pm-utils/sleep.d/00logging resume suspend:

/usr/lib/pm-utils/sleep.d/00logging resume suspend: success.
Wed Jul 11 12:19:37 EEST 2012: Finished.

dwnomad · 2012-07-12 02:31:43

I've had the same problems with my Asus Eee 1001PX. After 4 - 7 suspends, it fails and usually kicks me back to the display manager. On occasion it just leaves me with a blank screen and I have to pull up another TTY and blindly login and reboot. Disabling laptop-mode-tools did not resolve this for me. One interesting thing is that when I reboot after either type of failed suspend, it will finally suspend midway through the shutdown process.

yourself · 2012-07-12 06:46:09

A very similar thing happens to me!

When I suspend and the system fails to properly suspend, when I open the lid, the computer has dropped to the login screen. From this point on the wireless driver for some reason has failed (network manager says that there aren't any wireless drivers available) and if I initiate a poweroff, a "normal" suspend actually executes, suspending the system midway the poweroff! If I then resume it, it completes the poweroff!

Now that I think of it, maybe I sould have the wireless driver unload before suspension... I'll try and report what happens!

yourself · 2012-07-12 13:00:48

Furthermore, since I was 'plagued' by this: https://bbs.archlinux.org/viewtopic.php?id=139415 I disabled the security chip onboard my T420 and blacklisted the 'mei' module (ref. https://bbs.archlinux.org/viewtopic.php?id=133083 )

Fingers crossed.

ayr0 · 2012-07-15 01:56:52

Here's what I've done that seems to have solved the problem (at least on my Thinkpad T520 with Intel Wireless 6205)

I created /etc/pm/config.d/modules and put the following line it. So far, suspends have worked correctly everytime.

/etc/pm/config.d/modules

SUSPEND_MODULES="iwlwifi"

yourself · 2012-07-15 09:53:08

Yes, that's exactly what I have done when I said I'll try to have the wireless module unload before suspension but unfortunately I have permormed a couple of restarts from then for various reasons and have not left the system go with suspend/resume for more than a day.... But in any case, it hasn't happened since.

I'll report my findings after a couple of days with this setup!

yourself · 2012-07-16 10:25:01

Bummer!

It happened again, even with the

SUSPEND_MODULES="iwlwifi"

in /etc/pm/config.d/modules.

On a side note, I had re-enabled the loading of the 'mei' module, I'll try disabling 'mei' and see my mileage...

Essl · 2012-07-16 21:40:24

I seem to have a very similar/the same problem as you. On a Thinkpad W520.

thesequel · 2012-07-16 22:25:55

I have a T420 and suspend is working just fine. Did you install anything on your laptop (SSD, RAM, etc)? If you want a copy of any of my configuration files just let me know.

Last edited by thesequel (2012-07-16 22:30:37)

yourself · 2012-07-17 06:40:52

No, nothing works.

-> Tried blacklisting 'mei', no luck
-> Tried setting SUSPEND_MODULES="iwlwifi" in /etc/pm/config.d/modules, no luck
-> Tried disabling laptop-mode-tools, no luck
-> Tried removing 'autodetect' from HOOKS in mkinitcpio.conf (and furthermore stripping it to a bare minimum), no luck

Essl, from the looks of it, we have exactly the same problem! What I have noticed, though, is that when the problem happens I can log in the console (Ctrl+Alt+F1) as root and type 'reboot'. Then something really weird happens: the system starts shutting down normally and at some point, enters a normal sleep! When I resume it fom sleep, the reboot continues normally and the system reboots. Do you experience that as well? Also, do you have in your kernel log (/var/log/kernel.log.*) messages such as: "BUG: Bad rss-counter state mm:ffff880116400700 idx:1 val:-2"? My /var/log/kernel.log.* are full of them....

thesequel, yes, I have installed an 120GB intel SSD 320 instead of the original laptop's disk. Unfortunately I am 100% clueless on what could be causing the issue and don't know where to look. Furthermore, I have not meddled with many files in /etc, most should be at their defaults (and furthermore, I did not meddle with anything when the problem started in the first place because I have had arch on this laptop for more than a year now and never had any problems suspending/resuming before linux-3.4)... I am really in the dark, considering this.

This is driving me crazy!!! (Not to mention that this might as well fry my laptop: I'm used in suspending and throwing it in my backpack (without checking the moonlight 'sleep' led)! A couple of times I've thrown it in my backpack while *not* suspended and when I got it out it was really really hot!!!)

Last edited by yourself (2012-07-17 18:59:59)

Essl · 2012-07-17 14:11:18

I have no ssd, but I did swap the cd drive for a terabyte HDD. In addition to the inbuilt 500gb HDD.

I'll try the reboot thing once I experience this problem again and will let you know. Same for the kernel logs. Unfortunately I'm not home right now, so I don't have access to the laptop.

Last edited by Essl (2012-07-17 14:12:19)

Arch Linux

#1 2012-07-02 19:25:04

[SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#2 2012-07-02 19:50:35

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#3 2012-07-03 05:02:19

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#4 2012-07-04 16:15:52

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#5 2012-07-05 11:09:12

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#6 2012-07-06 21:26:51

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#7 2012-07-08 21:31:19

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#8 2012-07-08 22:01:20

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#9 2012-07-09 05:55:47

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#10 2012-07-09 06:19:22

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#11 2012-07-10 22:41:47

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#12 2012-07-11 06:30:11

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#13 2012-07-11 07:41:36

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#14 2012-07-11 08:29:36

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#15 2012-07-11 09:31:56

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#16 2012-07-12 02:31:43

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#17 2012-07-12 06:46:09

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#18 2012-07-12 13:00:48

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#19 2012-07-15 01:56:52

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#20 2012-07-15 09:53:08

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#21 2012-07-16 10:25:01

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#22 2012-07-16 21:40:24

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#23 2012-07-16 22:25:55

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#24 2012-07-17 06:40:52

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

#25 2012-07-17 14:11:18

Re: [SOLVED]Linux3.4 broke resume/suspend some of the times in Lenovo T420

Board footer