Bruce Evans wrote:
The sched_locks are now of course thread_locks. I put the vm unlock before the thread lock since the above order seems to risk a LOR. That may have been a mistake -- we would prefer not to be switched after deciding to do it ourself.
No, there can't be this LOR since thread_lock is a spinmutex while vm_page_queue_mtx is a sleepable mutex, so for our well known-rules about locking you cannot have the opposite situation. And if you don't follow John's pattern I think you get a race too since there is a breaking point in the protected path.
Index: vm_zeroidle.c=================================================================== RCS file: /usr/cvs/src/sys/vm/vm_zeroidle.c,v retrieving revision 1.45 diff -u -r1.45 vm_zeroidle.c --- vm_zeroidle.c 18 May 2007 07:10:50 -0000 1.45 +++ vm_zeroidle.c 7 Jun 2007 14:58:39 -0000 @@ -110,8 +110,10 @@ if (m != NULL && (m->flags & PG_ZERO) == 0) { vm_pageq_remove_nowakeup(m); mtx_unlock(&vm_page_queue_free_mtx); + critical_exit(); pmap_zero_page_idle(m); mtx_lock(&vm_page_queue_free_mtx); + critical_enter(); m->flags |= PG_ZERO; vm_pageq_enqueue(PQ_FREE + m->pc, m); ++vm_page_zero_count;Next I will try this. I put the critical_exit() before the vm unlock. mtx_unlock() should be allowed to switch if it wants. However, we would prefer to keep context switches disabled in the above -- just drop the lock so that other CPUs can proceed.
This is good since mtx_unlock will force a preemption point here. Attilio _______________________________________________ [email protected] mailing list http://lists.freebsd.org/mailman/listinfo/cvs-all To unsubscribe, send any mail to "[EMAIL PROTECTED]"
