On Fri, 2012-03-16 at 11:13 -0400, Andy Adamson wrote:
> On Thu, Mar 15, 2012 at 8:10 PM, Myklebust, Trond
> <Trond.Myklebust@xxxxxxxxxx> wrote:
> > On Thu, 2012-03-15 at 14:40 -0400, andros@xxxxxxxxxx wrote:
> >> From: Andy Adamson <andros@xxxxxxxxxx>
> >>
> >> Signed-off-by: Andy Adamson <andros@xxxxxxxxxx>
> >> ---
> >> include/linux/sunrpc/sched.h | 1 +
> >> net/sunrpc/sched.c | 27 +++++++++++++++++++++++++++
> >> 2 files changed, 28 insertions(+), 0 deletions(-)
> >>
> >> diff --git a/include/linux/sunrpc/sched.h b/include/linux/sunrpc/sched.h
> >> index dc0c3cc..fce0873 100644
> >> --- a/include/linux/sunrpc/sched.h
> >> +++ b/include/linux/sunrpc/sched.h
> >> @@ -235,6 +235,7 @@ void rpc_sleep_on_priority(struct rpc_wait_queue *,
> >> void rpc_wake_up_queued_task(struct rpc_wait_queue *,
> >> struct rpc_task *);
> >> void rpc_wake_up(struct rpc_wait_queue *);
> >> +void rpc_drain_queue(struct rpc_wait_queue *);
> >> struct rpc_task *rpc_wake_up_next(struct rpc_wait_queue *);
> >> struct rpc_task *rpc_wake_up_first(struct rpc_wait_queue *,
> >> bool (*)(struct rpc_task *, void *),
> >> diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
> >> index 1c570a8..11928ff 100644
> >> --- a/net/sunrpc/sched.c
> >> +++ b/net/sunrpc/sched.c
> >> @@ -551,6 +551,33 @@ void rpc_wake_up(struct rpc_wait_queue *queue)
> >> EXPORT_SYMBOL_GPL(rpc_wake_up);
> >>
> >> /**
> >> + * rpc_drain_queue - empty the queue and wake up all rpc_tasks
> >> + * @queue: rpc_wait_queue on which the tasks are sleeping
> >> + *
> >> + * Grabs queue->lock
> >> + */
> >> +void rpc_drain_queue(struct rpc_wait_queue *queue)
> >> +{
> >> + struct rpc_task *task;
> >> + struct list_head *head;
> >> +
> >> + spin_lock_bh(&queue->lock);
> >> + head = &queue->tasks[queue->maxpriority];
> >> + for (;;) {
> >> + while (!list_empty(head)) {
> >> + task = list_entry(head->next, struct rpc_task,
> >> + u.tk_wait.list);
> >> + rpc_wake_up_task_queue_locked(queue, task);
> >> + }
> >> + if (head == &queue->tasks[0])
> >> + break;
> >> + head--;
> >> + }
> >> + spin_unlock_bh(&queue->lock);
> >> +}
> >> +EXPORT_SYMBOL_GPL(rpc_drain_queue);
> >> +
> >
> > Confused... How is this function any different from rpc_wake_up()?
>
> Because it actually drains the queues where rpc_wake_up does not. See
> the attached output where I added the same printks to both
> rpc_drain_queue and rpc_wake_up.
So you are seeing a bug in rpc_wake_up()? I'm surprised; a bug of that
magnitude should have caused a lot of hangs. Can you please look into
what is happening.
--
Trond Myklebust
Linux NFS client maintainer
NetApp
Trond.Myklebust@xxxxxxxxxx
www.netapp.com
��.n��������+%������w��{.n�����{��w��� jg��� �����ݢj����G���� ���j:+v���w�m������w�������h�����٥
