[plug] AMD Machine Check errors

[Brad Campbell]

Does anyone know anything about AMD MCE's?

My main "server" at home is a Ryzen machine. Built in 2017 it's based on an ASUS X370-Prime Pro motherboard, and has had a few processors.
It had an 3950X for a long time, but that started to really be a pain in the bum with random reboots. I tried pretty much everything, 
and it was looking like a Cstate related issue. I had silicon based issues with the 1800X that was only resolved with an RMA and what I read
indicated that was likely again, but the warranty was long passed so I upgraded it to a 5950X.

The 5950X doesn't reboot, but in the last couple of months I've had 2 MCE's. I thought perhaps the 5950X was a bit more civilised
and might throw an MCE rather than just cause a panic and reboot.

After a month or so of searching, I've managed to find pretty much nothing that would allow me to decode these, so last resort
I thought I'd ask here.

[2117285.345782] mce: [Hardware Error]: Machine check events logged
[2117285.347945] mce: [Hardware Error]: CPU 1: Machine Check: 0 Bank 26: 90ffa9cb41e9a91e
[2117285.350174] mce: [Hardware Error]: TSC 0 
[2117285.352381] mce: [Hardware Error]: PROCESSOR 2:a20f12 TIME 1758400575 SOCKET 0 APIC 2 microcode a201210

[4685571.986956] mce: [Hardware Error]: Machine check events logged
[4685571.989655] mce: [Hardware Error]: CPU 1: Machine Check: 0 Bank 27: 8000000105c75163
[4685571.992434] mce: [Hardware Error]: TSC 0 
[4685571.995175] mce: [Hardware Error]: PROCESSOR 2:a20f12 TIME 1760968859 SOCKET 0 APIC 2 microcode a201210


I've replaced the CPU, the RAM and the PSU. I'm assuming it's the 2017 vintage Motherboard, but then I have 2 identical boards, bought
within a month of each other. One has been on 24x7 and the other probably has 200 hours on it. The problem remains regardless of the board.
I've switched them twice to test.

Anyone have a lead on how I might find out what these actually mean?

Regards,
Brad