[Bug 219399] System panics after several hours of 14-threads-compilation orgies using poudriere on AMD Ryzen...

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=219399

--- Comment #96 from Don Lewis <truckman at FreeBSD.org> ---
(In reply to Nils Beyer from comment #91)
Since the IRET instruction seems to play a factor in this problem, and it is
supposedly microcoded, I'm hoping that a microcode fix is possible for it.

I would not expect this to fix problems found by ryzen_segv_test since that
code does not use signals, so those processes should not be executing the
signal trampoline code that my patch relocates.

It's curious that disabling SMT fixes that problem for you.  Can you try
running ryzen_segv_test under "cpuset -l 2-15" to see if avoiding running the
test on the core that handles interrupts makes a difference?

-- 
You are receiving this mail because:
You are the assignee for the bug.