Good day. I looked at the data in your kern.log file
Namely the following lines
mmc0: Timeout waiting for hardware interrupt. (entry dated December 1)
You have a lot of these notes, which surprised me. Then I looked in Google for a message about the problem, I found that the problem is common.
By the way, your last record of the problem is still the same
Dec 7 10:22:10 raspberrypi kernel: [15538.096792] mmc0: Timeout waiting for hardware interrupt.
I concluded that the problem with the card (given that it is a bootable card) also affects the restart of the computer. But this is a “logical conclusion” based on messages on the Internet.
It would be interesting to deal with this error.
For example, see how your computer will behave on the same problem card, but with a different operating system. Because my Buster works worse with some programs than other OSs. Or see what happens if you change the settings in config.txt
So far I know that problems with memory cards are very critical for the fourth version of Raspberry, they started at the very beginning, when the “32 Gb memory card error” occurred.
https://www.raspberrypi.org/forums/view ... 8&t=248968
Then it was fixed, there were error messages "green death light" related (as reported on the forum here) with memory cards in some cases. https://www.raspberrypi.org/forums/view ... 8&t=250183
So I think studying the issue of memory cards might be useful to all of us.
I also know that Buster has problems leading to hardware failures.
This is normal for a young operating system, but it is annoying.
I have an idea that the problems of memory cards are due to the fact that in the fourth version, Raspberries changed the mechanism for working with memory cards. Including, as far as I remember from the messages of RPI development engineers, the option in config.txt clock speed for memory cards disappeared.
Unfortunately, I personally can’t verify this at all, because I only have old primitive Sandisk cards that work well. But perhaps in more advanced modern cards from Samsung, this problem manifests itself. Or is it because Samsung’s memory cards are often faked.
Also on forums they sometimes write that Samsung cards have a “special” controller.
I think the following can be done if there is interest in experiments.
https://www.raspberrypi.org/forums/view ... 9#p1576919
Perhaps if you tried to put a different version of the operating system on the problem card (I did this when I had hardware problems with the sound card in Buster), we could also conclude that the Buster itself is not the source of the error .
PS These are just my thoughts, and I cannot guarantee that I am going the right way.