Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1475083

Re: [PATCH 2/2] sched/wait: avoid abort_exclusive_wait() in __wait_on_bit_lock()

From Oleg Nesterov <oleg@redhat.com>
Newsgroups linux.kernel
Subject Re: [PATCH 2/2] sched/wait: avoid abort_exclusive_wait() in __wait_on_bit_lock()
Date 2016-09-02 14:10 +0200
Message-ID <scVse-7Ky-15@gated-at.bofh.it> (permalink)
References <saoK6-6YC-21@gated-at.bofh.it> <saoK6-6YC-23@gated-at.bofh.it> <scHpf-70l-21@gated-at.bofh.it> <scHpf-70l-19@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw


On 09/01, Peter Zijlstra wrote:
>
> > 	ret = 0;
> >
> > 	for (;;) {
> > 		prepare_to_wait_exclusive(wq, &q->wait, mode);
> >
> > 		if (test_bit(&q->key.bit_nr, &q->key.flag))
> > 			ret = action(&q->key, mode);
> >
> > 		if (!test_and_set_bit(&q->key.bit_nr, &q->key.flag)) {
> > 			/* we got the lock anyway, ignore the signal */
> > 			ret = 0;
> > 			break;
> > 		}
> >
> > 		if (ret)
> > 			break;
> > 	}
> > 	finish_wait(wq, &q->wait);
> >
> > 	return ret;
> >
> >
> > Would not that work too?
>
> Nope, because we need to do that finish_wait() before
> test_and_set_bit()..

Yes, I meant

	int __sched
	__wait_on_bit_lock(wait_queue_head_t *wq, struct wait_bit_queue *q,
				wait_bit_action_f *action, unsigned mode)
	{
		int ret = 0;

		for (;;) {
			prepare_to_wait_exclusive(wq, &q->wait, mode);
			if (test_bit(q->key.bit_nr, q->key.flags))
				ret = action(&q->key, mode);

			finish_wait(wq, &q->wait);

			if (!test_and_set_bit(q->key.bit_nr, q->key.flags))
				return 0;
			else if (ret)
				return ret;

		}
	}

> Also the problem with doing finish_wait() unconditionally would be
> destroying the FIFO order. With a bit of bad luck you'd get starvation
> cases :/

OK, I didn't think about that, thanks.

Oleg.

Back to linux.kernel | Previous | NextPrevious in thread | Next in thread | Find similar | Unroll thread


Thread

[PATCH 0/2] sched/wait: abort_exclusive_wait() should pass  TASK_NORMAL to wake_up() Oleg Nesterov <oleg@redhat.com> - 2016-08-26 15:00 +0200
  [PATCH 2/2] sched/wait: avoid abort_exclusive_wait() in  __wait_on_bit_lock() Oleg Nesterov <oleg@redhat.com> - 2016-08-26 15:00 +0200
    Re: [PATCH 2/2] sched/wait: avoid abort_exclusive_wait() in  __wait_on_bit_lock() Peter Zijlstra <peterz@infradead.org> - 2016-09-01 23:10 +0200
      Re: [PATCH 2/2] sched/wait: avoid abort_exclusive_wait() in         __wait_on_bit_lock() Oleg Nesterov <oleg@redhat.com> - 2016-09-02 14:10 +0200
    Re: [PATCH 2/2] sched/wait: avoid abort_exclusive_wait() in  __wait_on_bit_lock() Peter Zijlstra <peterz@infradead.org> - 2016-09-01 23:20 +0200
      Re: [PATCH 2/2] sched/wait: avoid abort_exclusive_wait() in  __wait_on_bit_lock() Peter Zijlstra <peterz@infradead.org> - 2016-09-02 00:20 +0200
        Re: [PATCH 2/2] sched/wait: avoid abort_exclusive_wait() in         __wait_on_bit_lock() Oleg Nesterov <oleg@redhat.com> - 2016-09-02 14:10 +0200
          Re: [PATCH 2/2] sched/wait: avoid abort_exclusive_wait() in  __wait_on_bit_lock() Peter Zijlstra <peterz@infradead.org> - 2016-09-02 15:30 +0200
      Re: [PATCH 2/2] sched/wait: avoid abort_exclusive_wait() in         __wait_on_bit_lock() Oleg Nesterov <oleg@redhat.com> - 2016-09-02 14:10 +0200
  [PATCH 1/2] sched/wait: abort_exclusive_wait() should pass  TASK_NORMAL to wake_up() Oleg Nesterov <oleg@redhat.com> - 2016-08-26 15:00 +0200
    Re: [PATCH 1/2] sched/wait: abort_exclusive_wait() should pass  TASK_NORMAL to wake_up() Peter Zijlstra <peterz@infradead.org> - 2016-09-01 13:40 +0200
      Re: [PATCH 1/2] sched/wait: abort_exclusive_wait() should pass         TASK_NORMAL to wake_up() Oleg Nesterov <oleg@redhat.com> - 2016-09-01 23:10 +0200
        Re: [PATCH 1/2] sched/wait: abort_exclusive_wait() should pass  TASK_NORMAL to wake_up() Peter Zijlstra <peterz@infradead.org> - 2016-09-02 00:50 +0200
  Re: [PATCH 0/2] sched/wait: abort_exclusive_wait() should pass  TASK_NORMAL to wake_up() Peter Zijlstra <peterz@infradead.org> - 2016-09-01 13:10 +0200

csiph-web