On Thu, Oct 10, 2013 at 05:20:44PM -0700, Josh Triplett wrote:
> On Wed, Oct 09, 2013 at 05:28:33PM -0700, Paul E. McKenney wrote:
> > On Wed, Oct 09, 2013 at 05:12:40PM -0700, Eric Dumazet wrote:
> > > On Wed, 2013-10-09 at 16:40 -0700, Josh Triplett wrote:
> > >
> > > > that. Constructs like list_del_rcu are much clearer, and not
> > > > open-coded. Open-coding synchronization code is almost always a Bad
> > > > Idea.
> > >
> > > OK, so you think there is synchronization code.
> > >
> > > I will shut up then, no need to waste time.
> >
> > As you said earlier, we should at least get rid of the memory barrier
> > as long as we are changing the code.
> >
> > Josh, what would you suggest as the best way to avoid the memory barrier,
> > keep sparse happy, and not be too ugly?
>
> The more I think about it, the more I realize that assigning an __rcu
> pointer to an __rcu pointer *without* a memory barrier is a sufficiently
> uncommon case that you probably *should* just write an open-coded
> assignment. Just please put a very clear comment right before it.
Fair enough, will do! Given earlier email, I believe that Eric is
fine with this, and if he isn't I am sure he will let us know. ;-)
> I'd originally thought it might make sense to have a macro similar to
> rcu_assign_pointer, but I just don't think this is a common enough case,
> and we don't want people thinking they can use this in general for __rcu
> to __rcu assignments (most of which still need a memory barrier).
Yep, it is a rather small fraction of rcu_assign_pointer() instances.
Thanx, Paul
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
