Re: [PATCH 13/25] aio: Convert read_events() to hrtimers

[Kent Overstreet <koverstreet@google.com>]

On Wed, Nov 28, 2012 at 04:24:36PM -0800, Zach Brown wrote:
> > -	int			i = 0;
> > +	DEFINE_WAIT(wait);
> > +	struct hrtimer_sleeper t;
> > +	size_t i = 0;
> 
> Changing i to size_t is kind of surprising.  Is that on purpose?

I doubt it matters due to limits on ringbuffer size elsewhere but size_t
is more correct, since we're counting objects in an array.

> > -		set_task_state(tsk, TASK_RUNNING);
> > -		remove_wait_queue(&ctx->wait, &wait);
> > -
> >  		if (unlikely(ret <= 0))
> >  			break;
> >  
> > @@ -879,11 +844,10 @@ static int read_events(struct kioctx *ctx,
> >  		event ++;
> >  		i ++;
> >  	}
> > -
> > -	if (timeout)
> > -		clear_timeout(&to);
> >  out:
> > -	destroy_timer_on_stack(&to.timer);
> > +	hrtimer_cancel(&t.timer);
> > +	destroy_hrtimer_on_stack(&t.timer);
> > +	finish_wait(&ctx->wait, &wait);
> 
> I'd move the finish_wait() up to where TASK_RUNNING was set previously
> so that we can't call copy_to_user() while still set to
> TASK_INTERRUPTIBLE.

Slightly less efficient, but yeah that's more correct and this code all
changes later anyways.

--
To unsubscribe, send a message with 'unsubscribe linux-aio' in
the body to majordomo@kvack.org.  For more info on Linux AIO,
see: http://www.kvack.org/aio/
Don't email: <a href=mailto:"aart@kvack.org">aart@kvack.org</a>