Comment 215 for bug 1690085

(In reply to Francesco Biscani from comment #129)
> My understanding is that the segfault problem (the only one publicly
> admitted by AMD so far) was a manufacturing issue, which was solved
> somewhere around the introduction of the TR. See also here:
>
Ryzen cpus manufactured before week 20 or so have the segfault problem with high stress compilation that average users are not doing. You get a new cpu Amd after submitting RMA.

>
> The other issue, the hang/reboot when idle, seems to be persisting also on
> newer processors, and, going by memory (again, anecdotal) I do remember at
> one point someone reporting this very problem on a TR.
>
> The current speculation/wishful thinking is that the idle bug may be solved
> by a future AGESA/microcode update, which should be imminent:

This bug is made by the intel developers who does not test their code with other hardware and older intel hardware is affected this problem too. Easy workaround is found in this bug report, put for those who can not follow discussions, here it is again:
To prevent random kernel lock ups, enable RCU_NOCB_CPU and boot the kernel with the rcu_nocbs=0-X command line parameter. X is the cpu thread count -1.