Constant reports of "Hardware Error"

[srcshelton]

PCEngines firmware version

v4.19.0.1

APU variant

apu2e

OS and OS Version

Linux 5.15.32, 6.0.7, 6.1.4

Affected component(s), peripheral(s) or functionality

The errors are reported as "corrected"… but there are a lot of them!

Brief summary

Very frequently (several times a day, at least) I'm seeing the following being logged via syslog():

[ T1030] [Hardware Error]: Corrected error, no action required.
[ T1030] [Hardware Error]: CPU:0 (16:30:1) MC0_STATUS[-|CE|-|AddrV|-|-|-]: 0x9400000000010015
[ T1030] [Hardware Error]: Error Addr: 0x00008ae14b79c000
[ T1030] [Hardware Error]: MC0 Error: L1 TLB multimatch.
[ T1030] [Hardware Error]: cache level: L1, tx: DATA

How reproducible

This has happened across a range of Linux kernels from early 5.x through to at least Linux 6.1 - and I'll test with more recent kernels shortly.

How to reproduce

Steps to reproduce the behavior:

1. Boot Linux
2. Watch the system logs
3. Even if fully correctable, there's something amiss here...

Expected behavior

No L1 TLB multimatch errors.

Actual behavior

Consistent errors over time.

[arjacy]

ECC RAM detecting a memory error?

[srcshelton]

Potentially, but I assume the TLB multimatch is something like this:

https://www.mail-archive.com/[email protected]/msg1575932.html

… which looks to have been fixed years ago in the general case, but still happening here?

(Plus, I'm booting with nopti and still getting this)