Re: kern/55229: cond_timedwait_race now fails randomly on real hardware

[Andreas Gustafsson <gson%gson.org@localhost>]

The following reply was made to PR kern/55229; it has been noted by GNATS.

From: Andreas Gustafsson <gson%gson.org@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: ad%NetBSD.org@localhost, thorpej%NetBSD.org@localhost
Subject: Re: kern/55229: cond_timedwait_race now fails randomly on real hardware
Date: Sun, 10 May 2020 11:03:33 +0300

 Looks like this was fixed by ad's commit of src/sys/kern/sys_lwp.c 1.80
 on 2020.05.05.22.12.06:
 
   lwp_unpark(): no need to acquire LWP refs or drop the proc lock.
 
   On the hacky benchmarks I have, held over from the transition to 1:1
   threading, this restores pthread_cond_signal() perf to radixtree/sleepq
   levels, and semes much better than either with pthread_cond_broadcast() and
   10 threads.  It would be interesting to see what might be achieved with a
   lockless lookup, which is within grasp now thanks to pid_table being used
   for lookup.
 
 -- 
 Andreas Gustafsson, gson%gson.org@localhost