lkml.org 
[lkml]   [2009]   [Aug]   [24]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: Is adding requeue_delayed_work() a good idea
Date

> > In my particular case it doesn't really matter. In the queued case it
> > could leave it to run whenever it gets to the head of the workqueue. In
> > the already running case then I think the timer should be reset. The
> > main point is that if I do requeue_delayed_work() I want to make sure
> > the work runs all the way through from the beginning at some point in
> > the future. The pattern I have in mind is something like:
> >
> > spin_lock_irqsave(&mydata_lock);
> > new_timeout = add_item_to_timeout_list();
> > requeue_delayed_work(wq, &process_timeout_list_work, new_timeout);
> > spin_unlock_irqsave(&mydata_lock);
> >
> > so if the process_timeout_list_work runs early or twice it doesn't
> > matter; I just want to make sure that the work runs from the beginning
> > and sees the new item I added to the list at some point after the
> > requeue.
>
> Hmm. But, asuming that process_timeout_list_work->func() takes mydata_lock
> too, you can just use queue_delayed_work() ?
>
> process_timeout_list_work can't miss new items, queue_delayed_work()
> can only fail if dwork is pending and its ->func has not started yet.

Maybe I misunderstand the code or misunderstand you, but looking at it:

int queue_delayed_work_on(int cpu, struct workqueue_struct *wq,
struct delayed_work *dwork, unsigned long delay)
{
int ret = 0;
struct timer_list *timer = &dwork->timer;
struct work_struct *work = &dwork->work;
if (!test_and_set_bit(WORK_STRUCT_PENDING, work_data_bits(work))) {
BUG_ON(timer_pending(timer));
BUG_ON(!list_empty(&work->entry));
//...
}
return ret;
}
if I have dwork already queued and try to requeue it for earlier it
seems that this will just silently return with the later expiration
still in effect.

Maybe I wasn't totally clear on what I want to happen; an example would
be if, say, I have delayed work queued with timer set to expire in 100
seconds and then a new piece of work is added that needs to expire in 1
second. So I want to requeue delayed work to run in 1 second, not wait
the 100 seconds for the original timeout work.

> I think we need requeue_xxx helper(s), but the exact semantics is not
> clear to me.

Not clear to me either unfortunately.

- R.


\
 
 \ /
  Last update: 2009-08-24 23:13    [from the cache]
©2003-2011 Jasper Spaans