--- Marc Herbert <[EMAIL PROTECTED]>
escreveu:

> Bruno Corsi dos Santos <[EMAIL PROTECTED]>
> writes:
> 
> > Other doubt is about the use of select syscall to
> wait
> > for data at the sockets. I know that select
> syscall
> > fails when the number of file descriptors in the
> > process application exceeds 1024. I didn't checked
> it
> > yet in carob, but I've already this problem with
> other
> > libraries.
> 
> Is this a fd_set / FD_SETSIZE bitmask issue ? Isn't
> an old and
> disappearing limitation of some legacy libcs ?
Yes, this limitation may be due to the FD_SETSIZE. I
don't know if it is disappearing, but I had some
problems few time ago, and an epoll syscall solved
that.

> 
> 
> > So, I think we would make available a
> > compilation using epoll syscall instead of select
> > syscall to avoid this problem.
> 
> This first carob implementation is using select()
> for portability
> reasons. Since the semantics are not too different,
> we could imagine
> some #ifdef USING_POLL -> poll() implementation if
> it does not clutter
> too much the code.
Sure. I think it could use poll syscall. I've never
used that, but I think there is no limitations like
select. But I think that poll is a bit different of
select and it would clutter the code, but do you know
other way?

> 
> However epoll looks totally different and very
> linux-specific/
> non-portable, right? What would be the benefit of
> using it?
I think in this case there is no benefit using epoll
instead of poll because it doesn't have to wait for
data from too many sockets. We should use it if we had
to wait data at thousands of sockets, by example. You
are right because the epoll is not portable at all. 

> 
> Thanks for your feedback.
> 
> 
> Marc.

Thanks,
Bruno



                
_______________________________________________________ 
Novidade no Yahoo! Mail: receba alertas de novas mensagens no seu celular. 
Registre seu aparelho agora! 
http://br.mobile.yahoo.com/mailalertas/ 
 


_______________________________________________
Carob mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/carob

Reply via email to