kuznet@ms2.inr.ac.ru wrote:
> Signal queueing makes this issue much more important, because events
> are not merged and signal overhead f.e. when receiving on ethernet
> becomes non-trivial.
Calling them signals is a bit bizarre too, as under load the *last*
thing you do is pick them up via relatively slow signal delivery.
IMO a callback device that you simply read() the siginfo from would be
much the cleaner API. For a start that would give >32 queues, ability
to group queues (the device itself may send siginfos to another queue),
no problems agreeing with your library which ones you're using, and a
"flat file" API which is generally quite nice.. But it's a bit late
now.
-- Jamie
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue Mar 07 2000 - 21:00:21 EST