Re: [RFC,PATCH] use rcu for fasync_lock

[Jamie Lokier <jamie@shareable.org>]

Davide Libenzi wrote:
> First, f_op->poll() does not allow you to send and event mask, 
> and this requires the driver to indiscriminately wake up both IN and OUT 
> waiters. The second area will be to give the driver to specify some "info" 
> 
> wake_up_info(&wq, XXXX);

I agree totally, both of these are (and always were, isn't it amazing
how long these things take) the way to do it "properly".

> The good thing is that migration can be gradual, beside the initial
> dumb compile fixing to suite the new f_op->poll() interface.

Even that's trivial, if a little time consuming, as it's only a
function signature change.  Actually using the extra argument is
optional for each driver.

-- Jamie