... about behavior at mutex unlock but about behavior at mutex >> re-lock. ... this is about reasonable >> expectations about thread concurrency behavior in a particular ... forgot that default scheduling policy is time-sharing, after thread > #2 has blocked ...

freebsd-threads - Andriy Gapon - May 15 2008 - 05:27

... behavior at mutex unlock but about behavior at mutex > re-lock ... for a scheduler to sacrifice performance to make threads happier. The ... very* wrong. DS _______________________________________________ freebsd-threads@freebsd.org mailing list http ...

freebsd-threads - David Schwartz - May 15 2008 - 16:25

... behavior at mutex unlock but about behavior at mutex > re-lock. ... pthread_cond_wait(...); > //dequeue item > ... > pthread_mutex_unlock(&work mutex); > //perform some short and non- ... that default scheduling policy is time-sharing, after thread #2 has ...

freebsd-threads - David Xu - May 15 2008 - 05:05

... a running thread in this circumstances. Especially if >> execution time between unlock and re- ... while holding the mutex, so probably this is why scheduling algorithm didn't ... rather than always avoiding context switch at all costs and >> depending on ...

freebsd-threads - Andriy Gapon - May 15 2008 - 16:48

... will > be close enough if thread scheduler is sane, so we ... behavior at mutex unlock but about behavior at mutex re-lock. ... attribute. -- Andriy Gapon _______________________________________________ freebsd-threads@freebsd.org mailing list http:// ...

freebsd-threads - Andriy Gapon - May 15 2008 - 04:35

... understand that all this is subject to (thread) scheduler policy, > but I think ... I expect is more reasonable, at least it is more > reasonable ... ("you have good thread scheduler\n"); pthread_mutex_unlock(&mutex); return 0; } -- Andriy ...

freebsd-threads - Andriy Gapon - May 14 2008 - 14:50

... processing of the item >> ... >> } >> >> Because the worker thread (while the queue is not empty) ... the mutex shortly after unlocking it. > > Well in theory, the kernel scheduler will let ... in the queue instead of one at a time. I suppose you could ...

freebsd-threads - Daniel Eischen - May 15 2008 - 13:54

... program. One > thread releases and re-acquires a mutex 10 times ... if > execution time between unlock and re-lock is very small ... how small the region the mutex is protecting. as long as ... context switch at all costs and > depending on scheduler to eventually ...

freebsd-threads - David Xu - May 15 2008 - 08:57

... the mutex shortly after unlocking it. Well in theory, the kernel scheduler will ... in the queue instead of one at a time. > So while improving performance on ... mutex re-acquire-ing > unfairness may be hurting interactivity and thread concurrency and ...

freebsd-threads - Daniel Eischen - May 15 2008 - 09:24

... more "fair", because it comes > at the incredibly high price of ... help - the scheduler might want to keep the starving thread on ... to be fair, the contested mutex case should try to handoff ... order to properly implement priority mutexes, there must be a handoff ...

freebsd-threads - Daniel Eischen - May 16 2008 - 17:55

... in mutex waiter's list. > > Is there any option (thread scheduler, ... that all this is subject to (thread) scheduler policy, > but I ... I expect is more reasonable, at least it is more > reasonable ... 1 some times to re-acquire the mutex without context switches ...

freebsd-threads - David Xu - May 15 2008 - 00:22

... #1 > releases a mutex and then tries to re-acquire it ... that all this is subject to (thread) scheduler policy, > but I ... I expect is more reasonable, at least it is more > reasonable ... freebsd.org/mailman/listinfo/freebsd-threads To unsubscribe, send any mail ...

freebsd-threads - Brent Casavant - May 15 2008 - 01:16

... about timeslices and > scheduling at that level at all. This is ... is because the thread that holds the mutex relinquishes the CPU ... this perfectly. DS _______________________________________________ freebsd-threads@freebsd.org mailing list http ...

freebsd-threads - David Schwartz - May 15 2008 - 18:37

... tries to get mutex... do not care. The way that scheduling works ... the time the scheduler will preempt it at the correct time ... s a period. -Alfred _______________________________________________ freebsd-threads@freebsd.org mailing list http:// ...

freebsd-threads - Alfred Perlstein - May 16 2008 - 16:15

... the lock when it > is preempted. You're analysis is completely correct. However, I ... with a wakeup and immediate schedule of the other thread -- which will have exactly the ... as the queued mutex, the lock-dropping thread will just context switch at a ...

freebsd-threads - Brent Casavant - May 16 2008 - 17:17

... then tries to re-acquire it while thread #2 was > already ... in mutex waiter's list. > > Is there any option (thread scheduler, ... that all this is subject to (thread) scheduler policy, > but I ... I expect is more reasonable, at least it is more > reasonable ...

freebsd-threads - David Schwartz - May 14 2008 - 20:29

... limited situations, and can be implemented at the application level through the use ... waiting for the mutex. Slow path. */ /* Initialize this thread's wait ... (handed to us by unlock) */ return 0; } int queued_mutex_unlock(queued_mutex_t *qm) ...

freebsd-threads - Brent Casavant - May 15 2008 - 18:23

... about timeslices and scheduling at that level at all. This is ... while the other thread is blocked on that mutex for 10 ... timeslices. -- Andriy Gapon _______________________________________________ freebsd-threads@freebsd.org mailing list http:// ...

freebsd-threads - Andriy Gapon - May 15 2008 - 16:56

... in place for the threads or mutex. And actually, >> I think ... for it. -- DE _______________________________________________ freebsd-threads@freebsd.org mailing list http ... .freebsd.org/mailman/listinfo/freebsd-threads To unsubscribe, send any mail ...

freebsd-threads - Daniel Eischen - May 17 2008 - 01:11

... to be fair, the contested mutex case should try > to handoff ... in place for the threads or mutex. And actually, > I think ... believe. -- - Alfred Perlstein _______________________________________________ freebsd-threads@freebsd.org mailing list http ...

freebsd-threads - Alfred Perlstein - May 16 2008 - 21:50