Bug#939170: linux: does not suspend completely, locks up

[Daniel M.]

Source: linux
Version: 5.2.9-2
Severity: important

Dear Maintainer,

*** Reporter, please consider answering these questions, where appropriate ***

* What led up to the situation?
* What exactly did you do (or not do) that was effective (or
ineffective)?
* What was the outcome of this action?
* What outcome did you expect instead?

*** End of the template - remove these template lines ***

Since the update from 4.19 to 5.2.9 my Lenovo Thinkpad E460 no longer suspends properly. If I
close the lid or chose Suspend from the KDE launcher, the system will start to suspend but never
reaches standby. Power light will blink as usual, display backlight turns off but keyboard LEDs
don't turn off, fan is still active.

The system doesn't respond to Ctrl-Alt-Del or change to a Terminal. No input changes anything.
Waiting (30min) doesn't help either. Rebooting to Kernel 4.19 fixes the problem immediatly.

Expected outcome would be a complete suspend/standby state and a complete recovery/powerup on
lid open or press of any key.


-- System Information:
Debian Release: bullseye/sid
APT prefers testing
APT policy: (500, 'testing'), (20, 'unstable'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.19.0-5-amd64 (SMP w/4 CPU cores)
Kernel taint flags: TAINT_OOT_MODULE, TAINT_UNSIGNED_MODULE
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8), LANGUAGE=en_US (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

[Moritz Schlarb]

Source: linux
Version: 5.2.9-2
Followup-For: Bug #939170

Hi everyone,

I'm seeing the same issue with linux-image-5.2.0-2-amd64=5.2.9-2 on Lenovo
Thinkpad X1 Carbon 4th Gen.

Best regards,
Moritz

-- System Information:
Debian Release: bullseye/sid
APT prefers testing

APT policy: (990, 'testing'), (800, 'unstable')
Architecture: amd64 (x86_64)

Kernel: Linux 5.2.0-2-amd64 (SMP w/4 CPU cores)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8), LANGUAGE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /usr/bin/dash

[Daniel M.]

Hi,

this also happens to me when I try systemctl suspend instead of a GUI
action. So it's not a problem of the desktop environment but of kernel
or even systemd.

Maybe something with the thinkpad kernel module?

Cheers,
Daniel

[Marc Bonnor]

Hi,

I think I may have the same problem.

When I shutdown or restart my laptop the process proceeds as normal to
the point where the last message printed on the screen is "Reached
target power-off" and then nothing happens. I then need to push and
hold power button until power switches off.

I setup Systemd persistent logs to allow me to view log messages up
until power off, but again nothing jumps out as being a problem.

I have a Lenovo ThinkPad Laptop;

DMI: LENOVO 20JAA00FAU/20JAA00FAU, BIOS R0MET48W (1.25) 04/22/2019.


I found this Systemd Debugging tip;

poweroff -f

If this command doesn't work, then Systemd is unlikely to be the
problem and it may be kernel related.

When I tried this command the GUI screen froze and there was a loud
annoying continuous beep.

Pressing and holding the power button turned off the laptop.

[Daniel M.]

Hi,

is it possible this one is reported against the wrong package? It
seems like there are a few lenovo laptop / thinkpad owners who
experience this. Is there some (pseudo)package to report this to?
Maybe it gets a bit more traction there...

Cheers,
Daniel

[Felix Rublack]

Hi everyone,

I have the same issue on a ThinkPad T460s (suspend works only half way,
reboot and poweroff stop just short before actual shutdown).

I bisected the problem to the TPM driver. Blacklisting the module in
modprobe.d fixed the problem for me.

Sample configuration: /etc/modprobe.d/blacklist-tpm.conf

# Prevent TPM from loading. It breaks suspend and power cycle.
blacklist tpm
blacklist tpm_crb
blacklist tpm_tis
blacklist tpm_tis_core

Greetings
Felix Rublack

[Daniel M.]

Hi, thanks a lot! I can confirm that this works also on my Thinkpad
E460. Since you can probably provide more details, could you forward
this to kernel.org?

Cheers,
Daniel

[Frank Loeffler]

Hi,

just to add one tiny bit of confirmation: this helped also in my case
(big thanks!):

I am on Buster with a Lenovo Thinkpad T460, with kernels 4.19.0.8 and
5.4.0.0.bpo.4 installed. Suspend works fine with 4.19.0.8. The same
system fails consistently when booted with 5.4, in the same way reported
earlier: the system starts to suspend, but stops somewhere close to the
finish line: display is black (don't remember if backlight was off, but
I think it was), power-led is blinking, but the led-indicators like mute
stay on. I cannot say something about the fan, as it is usually not
running for me anyway. Nothing brings the laptop back at this stage
other than pressing the power-button for like 10 seconds (complete
restart): no shorter press of the power button, no lid action. Both
should, and with the 4.19 kernel, do.

With above blacklisting, suspend now also works for kernel version 5.4.

Big thanks again - and it would be interesting to follow this at the
kernel.org-level (to know when to remove the blacklist again).

Frank Löffler

[Borys Jurgiel]

ThinkPad T460s: Instead of blacklisting the module, I just switched from
"Intel PTT" to "Discrete TPM" in BIOS and everything works perfectly now.

Thanks a lot from me too!
Borys

[Uwe Steinmann]

I'm running a T560 for many years, it always had problems coming
back from suspend. It wasn't related to any kernel version starting
from 5.10.x A half way working work around was to send the laptop
into hybrid sleep. It wouldn't come back from sleep when the lid was
opened or the power button was pressed, but hard resetting it, at least
allowed me to hibernate from disc. After that initial procedure it
worked for weeks or even month with at least opening and closing the
lid once a day. But recently (after more than 3 month running without
a reboot) that trick didn't do it anymore. I couldn't figure out why.
I even updated the bios without any success.

Setting 'Discrete TPM' finally made it work again, even without any work
around.

Uwe

[Andy Smith]

I just installed Debian 11 on a 4th gen Thinkpad X1 Carbon and it
failed to resume from suspend. This was working in Debian 10.

Setting the "TPM Mode" to "Discrete TPM" as mentioned by Uwe
Steinmann also appears to have fixed things for me.

Thanks,
Andy

[Jean-Michel Zwygart]

I updated a Thinkpad X1 Yoga (is detected as X1 Carbon 4th gen) from
Buster to Bullseye a few months ago and after the update the shutdown
reached status power off but didn't power off the machine. Tried
different shutdown command options from terminal, with always the same
result. As relatively new Linux Debian user with I fist thought that I
did something wrong or that my Thinkpad had a problem. I installed other
distro's and the problem disappeared... so I assumed it was a Debian issue.

But as I want to stay on Debian, I register to the Forum and fortunately
found this post.

I modified blacklisted tpm, as mentioned by Frank Löffler, and the
solution worked also for me!

Thanks a lot to Frank for his suggestion, and the others who confirmed
this solution, that made me confident, as new user, to made the change.

Best regards

Jean-Michel

[Jean-Michel Zwygart]

May 13. 2023 UPDATE same issue back again on Bookworm RC2

As I use my Thinkpad X1 Yoga (is detected as X1 Carbon 4th gen) also as test machine, I made a fresh Bookworm install and had the same issue (even worse) as on Bullseye. I say worse because on bullseye the issue was at least visible, as the screen didn't power off completely. But now on Bookworm the display power off... but not the machine. This is why I even didn't realize it after installation... but only when my TP didn't want to start 2-3 days later... the battery was drained off!

After recharging, I  it saw that after shutdown, only sign was that the small led on the ESC/FnLk key was still on! Something that I didn't see earlier when closing the the laptop lid.

I added, as per Felix Rublack recommendation, the same entries in /etc/modprobe.d/blacklist-tpm.conf (Message #40),

>> blacklist tpm
>> blacklist tpm_crb
>> blacklist tpm_tis
>> blacklist tpm_tis_core

but unfortunately this doesn't work on Bookworm!

As I'm a relatively new Linux/Debian user with limited knowledge I would appreciate any suggestion how to solve the problem.

Thanks in advance for any suggestion!

Jean-Michel Zwygart

[Nolan]

On Sat, 13 May 2023 17:31:35 +0200 Jean-Michel Zwygart
<jmzw...@bluewin.ch> wrote:
> May 13. 2023 UPDATE same issue back again on Bookworm RC2
>
> As I use my Thinkpad X1 Yoga (is detected as X1 Carbon 4th gen) also as
> test machine, I made a fresh Bookworm install and had the same issue
> (even worse) as on Bullseye. I say worse because on bullseye the issue
> was at least visible, as the screen didn't power off completely. But now
> on Bookworm the display power off... but not the machine. This is why I
> even didn't realize it after installation... but only when my TP didn't
> want to start 2-3 days later... the battery was drained off!
>
> After recharging, I  it saw that after shutdown, only sign was that the
> small led on the ESC/FnLk key was still on! Something that I didn't see
> earlier when closing the the laptop lid.
>
> I added, as per Felix Rublack recommendation, the same entries in
> /etc/modprobe.d/blacklist-tpm.conf (Message #40),
>
> >> blacklist tpm
> >> blacklist tpm_crb
> >> blacklist tpm_tis
> >> blacklist tpm_tis_core
>
> but unfortunately this doesn't work on Bookworm!

Looks like between bullseye and bookworm TPM support switched from
modular to built in, so the old workaround will not work without a
kernel rebuild.

bullseye:~# grep TPM /boot/config-5.10.0-25-amd64
CONFIG_TCG_TPM=m
CONFIG_HW_RANDOM_TPM=y
CONFIG_TCG_VTPM_PROXY=m

bookworm:~# grep TPM /boot/config-6.1.0-11-amd64
CONFIG_TCG_TPM=y
CONFIG_TCG_VTPM_PROXY=m

[Diederik de Haas]

On Wednesday, 23 August 2023 18:31:50 CEST Nolan wrote:
> Looks like between bullseye and bookworm TPM support switched from
> modular to built in, so the old workaround will not work without a
> kernel rebuild.
>
> bullseye:~# grep TPM /boot/config-5.10.0-25-amd64
> CONFIG_TCG_TPM=m
> CONFIG_HW_RANDOM_TPM=y
> CONFIG_TCG_VTPM_PROXY=m
>
> bookworm:~# grep TPM /boot/config-6.1.0-11-amd64
> CONFIG_TCG_TPM=y
> CONFIG_TCG_VTPM_PROXY=m

diederik@bagend:~/dev/debian/salsa/kernel-team/linux$ grep -r CONFIG_TCG_TPM debian/config/
debian/config/arm64/config:CONFIG_TCG_TPM=m
debian/config/kernelarch-x86/config:CONFIG_TCG_TPM=m
debian/config/config.cloud:CONFIG_TCG_TPM=m

The Debian kernel configuration didn't change, so I guess it's one of its
dependencies which turned the `=m` into `=y`?

[Diederik de Haas]

On Wednesday, 23 August 2023 19:32:19 CEST Diederik de Haas wrote:
> diederik@bagend:~/dev/debian/salsa/kernel-team/linux$ grep -r CONFIG_TCG_TPM
> debian/config/ debian/config/arm64/config:CONFIG_TCG_TPM=m
> debian/config/kernelarch-x86/config:CONFIG_TCG_TPM=m
> debian/config/config.cloud:CONFIG_TCG_TPM=m
>
> The Debian kernel configuration didn't change, so I guess it's one of its
> dependencies which turned the `=m` into `=y`?

This seems the same or similar to https://bugs.debian.org/1041007

[Nolan]

On Wed, 23 Aug 2023 09:31:50 -0700 Nolan <deb...@sigbus.net> wrote:>
Looks like between bullseye and bookworm TPM support switched from
> modular to built in, so the old workaround will not work without a
> kernel rebuild.
>
> bullseye:~# grep TPM /boot/config-5.10.0-25-amd64
> CONFIG_TCG_TPM=m
> CONFIG_HW_RANDOM_TPM=y
> CONFIG_TCG_VTPM_PROXY=m
>
> bookworm:~# grep TPM /boot/config-6.1.0-11-amd64
> CONFIG_TCG_TPM=y
> CONFIG_TCG_VTPM_PROXY=m

I tried another workaround that appears to work.

Disable the TPM in the BIOS. It is in the BIOS setup under the
"Security" tab: "Security Chip". I set that to disabled and "systemctl
suspend" now looks like it works, all the keyboard lights go out.
Pressing power resumes normally.