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

Groups > linux.kernel > #1397159

Re: [PATCH 2/4] locking/rwsem: Drop superfluous waiter refcount

Show all headers | View raw

On Mon, May 09, 2016 at 08:56:07AM -0700, Davidlohr Bueso wrote:
> On Mon, 09 May 2016, Peter Zijlstra wrote:
> 
> >On Sun, May 08, 2016 at 09:56:08PM -0700, Davidlohr Bueso wrote:
> >>Read waiters are currently reference counted from the time it enters
> >>the slowpath until the lock is released and the waiter is awoken. This
> >>is fragile and superfluous considering everything occurs within down_read()
> >>without returning to the caller, and the very nature of the primitive does
> >>not suggest that the task can disappear from underneath us. In addition,
> >>spurious wakeups can make the whole refcount useless as get_task_struct()
> >>is only called when setting up the waiter.
> >
> >So I think you're wrong here; imagine this:
> >
> >
> >	rwsem_down_read_failed()			rwsem_wake()
> >	  get_task_struct();
> >	  raw_spin_lock_irq(&wait_lock);
> >	  list_add_tail(&waiter.list, &wait_list);
> >	  raw_spin_unlock_irq(&wait_lock);
> >							  raw_spin_lock_irqsave(&wait_lock)
> >							  __rwsem_do_wake()
> >	  while (true) {
> >	    set_task_state(tsk, TASK_UNINTERRUPTIBLE);
> >							    waiter->task = NULL
> >	    if (!waiter.task) // true
> >	      break;
> >
> >	  __set_task_state(tsk, TASK_RUNNING);
> >
> >	do_exit();
> >							    wake_up_process(tsk); /* BOOM */
> 
> I may be missing something, but rwsem_down_read_failed() will not return until
> after the wakeup is done by the rwsem_wake() thread. 

The above never gets to schedule(), and even if it did, a spurious
wakeup could've happened, no?

> So racing with do_exit() isn't
> going to occur because the task is still blocked at that point. This is even more
> so with delaying the wakeup. Similarly, we don't do this for writers either, which
> could also suffer from similar scenarios.

The write side is different; it serializes on wait_lock. See how it
takes wait_lock again, after blocking, and removes itself from the
wait_list.

Readers do not do this, they rely on the waker to remove them, and
therefore suffer this problem.

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

Thread

[PATCH -tip 0/4] locking/rwsem (xadd): Reader waiter optimizations Davidlohr Bueso <dave@stgolabs.net> - 2016-05-09 07:00 +0200
  [PATCH 4/4] locking/rwsem: Rework zeroing reader waiter->task Davidlohr Bueso <dave@stgolabs.net> - 2016-05-09 07:10 +0200
    Re: [PATCH 4/4] locking/rwsem: Rework zeroing reader waiter->task Peter Zijlstra <peterz@infradead.org> - 2016-05-09 10:00 +0200
  [PATCH 3/4] locking/rwsem: Enable lockless waiter wakeup(s) Davidlohr Bueso <dave@stgolabs.net> - 2016-05-09 07:10 +0200
    Re: [PATCH 3/4] locking/rwsem: Enable lockless waiter wakeup(s) Peter Zijlstra <peterz@infradead.org> - 2016-05-09 09:50 +0200
  [PATCH 2/4] locking/rwsem: Drop superfluous waiter refcount Davidlohr Bueso <dave@stgolabs.net> - 2016-05-09 07:10 +0200
    Re: [PATCH 2/4] locking/rwsem: Drop superfluous waiter refcount Peter Zijlstra <peterz@infradead.org> - 2016-05-09 09:40 +0200
      Re: [PATCH 2/4] locking/rwsem: Drop superfluous waiter refcount Davidlohr Bueso <dave@stgolabs.net> - 2016-05-09 18:00 +0200
        Re: [PATCH 2/4] locking/rwsem: Drop superfluous waiter refcount Peter Zijlstra <peterz@infradead.org> - 2016-05-09 18:20 +0200
          Re: [PATCH 2/4] locking/rwsem: Drop superfluous waiter refcount Davidlohr Bueso <dave@stgolabs.net> - 2016-05-09 21:00 +0200
            Re: [PATCH 2/4] locking/rwsem: Drop superfluous waiter refcount Peter Zijlstra <peterz@infradead.org> - 2016-05-09 21:10 +0200
  [PATCH 1/4] locking/rwsem: Avoid stale ->count for rwsem_down_write_failed() Davidlohr Bueso <dave@stgolabs.net> - 2016-05-09 07:10 +0200
    Re: [PATCH 1/4] locking/rwsem: Avoid stale ->count for  rwsem_down_write_failed() Peter Hurley <peter@hurleysoftware.com> - 2016-05-09 07:40 +0200
      Re: [PATCH 1/4] locking/rwsem: Avoid stale ->count for  rwsem_down_write_failed() Peter Zijlstra <peterz@infradead.org> - 2016-05-09 09:20 +0200
      Re: [PATCH 1/4] locking/rwsem: Avoid stale ->count for  rwsem_down_write_failed() Davidlohr Bueso <dave@stgolabs.net> - 2016-05-09 16:40 +0200
  Re: [PATCH -tip 0/4] locking/rwsem (xadd): Reader waiter optimizations Waiman Long <waiman.long@hpe.com> - 2016-05-10 03:40 +0200

csiph-web