I apparently responded to Jonathan via private email.
The patch below fixes the eventhandler lock ordering
problem.
steve
On Tue, Apr 02, 2002 at 06:25:40PM -0800, Jonathan Mini wrote:
> Steven G. Kargl [[EMAIL PROTECTED]] wrote :
>
>> lock order reversal
>> 1st 0xc02d9b40 eventhandler @ /usr/src/sys/kern/subr_eventhandler.c:163
>> 2nd 0xc4050008 shutdown_post_sync @ /usr/src/sys/kern/subr_eventhandler.c:116
>
> [ ... ]
>
>> I believe the second is known, but the first appears
>> to be related to the recent eventhandler locking.
>
> Sorry about that.
>
> Index: kern/subr_eventhandler.c
> ===================================================================
> RCS file: /usr/cvs/src/sys/kern/subr_eventhandler.c,v
> retrieving revision 1.16
> diff -u -r1.16 subr_eventhandler.c
> --- sys/kern/subr_eventhandler.c 2 Apr 2002 04:18:54 -0000 1.16
> +++ sys/kern/subr_eventhandler.c 3 Apr 2002 02:21:29 -0000
> @@ -101,11 +101,11 @@
> sx_init(&list->el_lock, name);
> list->el_flags = EHE_INITTED;
> }
> + mtx_unlock(&eventhandler_mutex);
>
> /* allocate an entry for this handler, populate it */
> if ((eg = malloc(sizeof(struct eventhandler_entry_generic),
> M_EVENTHANDLER, M_NOWAIT)) == NULL) {
> - mtx_unlock(&eventhandler_mutex);
> return(NULL);
> }
> eg->func = func;
> @@ -125,7 +125,6 @@
> if (ep == NULL)
> TAILQ_INSERT_TAIL(&list->el_entries, &eg->ee, ee_link);
> EHE_UNLOCK(list);
> - mtx_unlock(&eventhandler_mutex);
> return(&eg->ee);
> }
>
>
> --
> Jonathan Mini
> [EMAIL PROTECTED]
>
> <baka^ni> desolation... despair... plastic forks...
>
>
> To Unsubscribe: send mail to [EMAIL PROTECTED]
> with "unsubscribe freebsd-current" in the body of the message
--
Steve
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
