Re: [patch] clone_startup(), 2.5.31-A0

[Jamie Lokier]

Rusty Russell wrote:
> > btw., with the help of these syscalls and futexes, the layer between Linux
> > and pthreads became very thin - almost nonexistant in the majority of
> > cases.
> 
> Woohoo!

Yes, futexes are wonderful, if quite difficult to figure out at first.
(I still haven't understood why the futex example reader-writer locks are
they way are.)

Even things like waitpid() are potentially faster with futexes -- no pid
to lookup!

-- Jamie