Comment # 5 on bug 979853 from Borislav Petkov

Yikes, rIP corruption this time. Even with the fixed microcode...

I don't have another idea right now besides maybe trying latest upstream
4.6 on one of them (would that be even possible?) to rule out some bug
the fix for which we haven't backported yet...

Also, how reproducible is this? Are you seeing the same type of splat or
everytime different ones? Can you copy all dmesg logs somewhere for me
to stare at?

I'm assuming those machines have ECC DIMMs and you have mcelog running
on them?

Can you point me to a full dmesg of one from the beginning?

Thanks.