Re: [PATCHv4 2/2] memory barrier: adding smp_mb__after_lock

[Jarek Poplawski <jarkao2@gmail.com>]

On Fri, Jul 03, 2009 at 09:41:26AM +0200, Jiri Olsa wrote:
> On Thu, Jul 02, 2009 at 07:39:04AM -0700, Davide Libenzi wrote:
> > On Thu, 2 Jul 2009, Eric Dumazet wrote:
> > 
> > > Jiri Olsa a écrit :
> > > > Adding smp_mb__after_lock define to be used as a smp_mb call after
> > > > a lock.  
> > > > 
> > > > Making it nop for x86, since {read|write|spin}_lock() on x86 are 
> > > > full memory barriers.
> > > > 
> > > > wbr,
> > > > jirka
> > > > 
> > > > 
> > > > Signed-off-by: Jiri Olsa <jolsa@redhat.com>
> > > 
> > > 
> > > Maybe we should remind that sk_has_helper() is always called
> > > right after a call to read_lock() as in :
> > > 
> > > 	read_lock(&sk->sk_callback_lock);
> > > 	if (sk_has_sleeper(sk))
> > > 		wake_up_interruptible_all(sk->sk_sleep);
> > 
> > Agreed, that'd be to have it in the source code comment.
> > 
> > 
> > - Davide
> > 
> 
> ok, I'll add it to the 1/2 part in v5
> 

Btw., there is a tiny typo:

- receive callbacks. Adding fuctions sock_poll_wait and sock_has_sleeper
+ receive callbacks. Adding fuctions sock_poll_wait and sk_has_sleeper

Jarek P.