System stuck with "Fixing recursive fault, reboot is required"

[SMG]

I have taken a look into it. It seems few steps are not being run when shutdown freeze. These are those steps.

I noticed two other events which may or may not be related, but will mention them in case they are.

In the "normal shutdown", I noticed the screensaver was used several times and also the computer was put to sleep at least once during the boot cycle.

The shutdown with issue was a boot cycle that was not put to sleep and I did not see where the screensaver came up. The boot cycle which had the radeon errors listed in it also had this segment (which the "normal shutdown" did not).

Code: Select all

Sep 10 01:56:58 raj-Lenovo-G50-80 systemd[1]: Stopped Light Display Manager.
Sep 10 01:56:58 raj-Lenovo-G50-80 systemd[1]: lightdm.service: Failed with result 'timeout'.
Sep 10 01:56:58 raj-Lenovo-G50-80 systemd[1]: lightdm.service: Killing process 967 (Xorg) with signal SIGKILL.
Sep 10 01:56:58 raj-Lenovo-G50-80 systemd[1]: lightdm.service: State 'stop-final-sigterm' timed out. Killing.
Sep 10 01:56:48 raj-Lenovo-G50-80 systemd[1]: lightdm.service: Main process exited, code=killed, status=9/KILL
Sep 10 01:56:48 raj-Lenovo-G50-80 systemd[1]: lightdm.service: Killing process 928 (gdbus) with signal SIGKILL.
Sep 10 01:56:48 raj-Lenovo-G50-80 systemd[1]: lightdm.service: Killing process 967 (Xorg) with signal SIGKILL.
Sep 10 01:56:48 raj-Lenovo-G50-80 systemd[1]: lightdm.service: Killing process 887 (lightdm) with signal SIGKILL.
Sep 10 01:56:48 raj-Lenovo-G50-80 systemd[1]: lightdm.service: State 'stop-sigterm' timed out. Killing.

It seems the fact radeon was waiting for a reboot caused problems with LightDM which resulted in the 'timeout'. LightDM is part of the graphics-related code.

I do not know if a sleep cycle might somehow affect the radeon driver, but it's a thought to keep in mind as you determine what might be different on the cases where shutdown works and shutdown does not work.

[rajeevrdev]

How do you think I can check if this is the issue? Should I try disabling screensaver?

[SMG]

How do you think I can check if this is the issue? Should I try disabling screensaver?

The screensaver is separate functionality from the computer going to sleep.

Do you have your system automatically go to sleep after a certain amount of time? I just put my computer to sleep manually, so tracking whether or not I put it to sleep during a boot cycle is something I could manually track and see if that corresponds to when the issue does or does not happen.

I have no specific reason to suspect putting the computer to sleep should help the radeon driver during a shutdown, but just happened to notice that difference in the logs. The fact it does not happen every cycle means something different must be happening. Just trying to come up with possible ideas of items to check.

[rajeevrdev]

I have it enabled on battery. When its powered , I manually put it to sleep. I am yet to notice any correlation between it. Do you think automatic sleep might be one suspect?

[SMG]

Do you think automatic sleep might be one suspect?

Not at this point. The difference between the logs was one did not sleep at all and that was the one which would not shut down.

Maybe we'll get lucky and the recent kernel update will mean the problem is gone. We can think positive.

[rajeevrdev]

I am on the newest kernel and the issue still persists. I have had the redeon graphics card errors on all kernels i tried. Mine is a 2016 laptop. I see this issue with Lenovo and Acer laptops in many places but with no solutions at all. The only good thing is , the shutdown freeze dont happen everytime.

[SMG]

The only good thing is , the shutdown freeze dont happen everytime.

The issues which are hard to easily reproduce, such as this one which does not happen every time, often take longer in finding a solution.

I do not know if using a kernel parameter radeon.dpm=0 card might help this situation. Usually, however, the problem that fixes is a consistent issue rather than a random one and it relates to specific errors in journalctl which your system did not have.

[rajeevrdev]

I upgraded to latest kernel but it showed many errors. Dont know if this is part of the existing problem. There are no problems in normal functioning of Mint. Faced shutdown freeze issue again. Here is the error messages that I saw while upgrading system. Is this relevant?

[SMG]

Here is the error messages that I saw while upgrading system. Is this relevant?

Those are not error messages. Not all messages listed on screen are errors.

Those are warnings about possible missing firmware. Most of that firmware is something quite new computers might need to download because it is not yet included in the linux-firmware package due to its newness.

You are welcome to add the missing firmware, but I do not see anything at first glance of the list that might help. The errors you are receiving at shutdown are related to the radeon driver and not the amdgpu or i915 drivers.