Jim Houston wrote:
> 
> Hi Linus, George,
> 
> I like the direction that this ERESTART_RESTARTBLOCK patch is
> going.
> 
> It might be nice to clear the restart_block.fun in handle_signal()
> in the ERESTART_RESTARTBLOCK path which returns -EINTR.  This eliminates
> the chance of a stale restart.

How is this?

Note that we do leave tracks in user land when this happens
as the return array is set on the restart exit.  Since we
don't know what it was AND the standard says NOTHING about
it being set (or not set) if there is no error, I don't
think this is a problem.
-- 
George Anzinger   george@mvista.com
High-res-timers: 
http://sourceforge.net/projects/high-res-timers/
Preemption patch:
http://www.kernel.org/pub/linux/kernel/people/rml