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

bugzilla-noreply at freebsd.org bugzilla-noreply at freebsd.org
Sat Aug 5 02:14:14 UTC 2017


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

--- Comment #207 from Nils Beyer <nbe at renzel.net> ---
(In reply to Nils Beyer from comment #206)

answering myself: yes, maybe I should just listen to what Don said: I do need
that origin patch because without it the programm is not able to MMAP at
0x7f(...) - as Don instructed. And at 0x4f(...) there seems to be no problem.

Using "cpuset -l 0 ./ryzen_provoke_crash" it freezes exactly where my first
system did:
--------------------------------------------------------------------------------
[...]
executing at 0x7fffffffff3a ..........
executing at 0x7fffffffff3b ..........
executing at 0x7fffffffff3c ..........
executing at 0x7fffffffff3d ..........
executing at 0x7fffffffff3e ..........
executing at 0x7fffffffff3f ..........
executing at 0x7fffffffff40 ......
--------------------------------------------------------------------------------

Without using "cpuset -l 0" the program seems to pass the 0x7fffffffff40 mark,
but becomes very, very slow - don't know if it freezes because I aborted it;
took too long for one row...

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


More information about the freebsd-bugs mailing list