[Bug 270785] Performance and power efficiency regression due to pthread_cond_timedwait() changes

From: <bugzilla-noreply_at_freebsd.org>
Date: Thu, 13 Apr 2023 16:35:27 UTC

--- Comment #3 from Mark Johnston <markj@FreeBSD.org> ---
So minecraft is generating many, many more calls to _umtx_op() than before.  If
the kernel commit is reverted, and nothing else changes, does the number of
calls stay the same?  If not, then there may indeed be a kernel bug.

It would be useful to see the timeout being passed.  Some output from a dtrace
script like:

# dtrace -n 'fbt::do_wait:entry /args[3] && pid == <pid of minecraft

would be useful.

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