On Mon, Oct 14, 2013 at 02:23:55AM -0700, Paul E. McKenney wrote:
> On Mon, Oct 14, 2013 at 11:05:08AM +0200, Peter Zijlstra wrote:
> > On Sat, Oct 12, 2013 at 07:06:56PM +0200, Oleg Nesterov wrote:
> > > it even disables irqs, so this should always imply rcu_read_lock() with
> > > any implementation,
> >
> > Not so; I could make an RCU implementation that drives the state machine
> > from rcu_read_unlock(). Such an implementation doesn't need the
> > interrupt driven poll-state driver we currently have and could thus
> > subvert that assumption :-)
> >
> > Then again, there's a good reason PaulMck didn't pick this
> > implementation.
>
> True enough, but there really are some out-of-tree RCU implementations
> that do take this approach and where disabling interrupts would not
> block preemptible RCU. So please do not rely on this implementation
> detail. You never know...
Actually, the current implementation of SRCU is not blocked by CPUs
disabling interrupts!
Thanx, Paul
> > > In fact I do not even understand why getaffinity() doesn't simply
> > > return ->cpus_allowed, but this is off-topic.
> >
> > Yeah, me neither :-(, it always surprises me. But changing it is likely
> > to break stuff so there we are.
>
> I know that feeling...
>
> 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/
