Subject: Re: possible bug in NetBSD asynchronous I/O
To: Emmanuel Dreyfus <p99dreyf@criens.u-psud.fr>
From: Ignatios Souvatzis <is@netbsd.org>
List: tech-kern
Date: 04/28/2001 22:34:40
On Sat, Apr 28, 2001 at 09:40:00PM +0200, Emmanuel Dreyfus wrote:
> > FASYNC == O_ASYNC which is documented to send SIGIO. That is
> > what asynchronous/io is.
>
> According to Design & Implementation of 4.4BSD OS, Async I/O is about
> sending "a signal (SIGIO) when read or write becomes possible".
>
> In our fcntl(2) man page, we describe O_ASYNC like this: "Enable the
> SIGIO signal to be sent to the process group when I/O is possible, e.g.,
> upon availability of data to be read."
>
> I do not exepct to receive SIGIO because the reader at the other end of
> a pipe has readen a byte. Do you expect such a behavior? Since NetBSD
> seems to be the only OS doing this, I beleive there is a real problem
> here.
Hm, when the reader reads one byte, I can write one byte, and if I'm async
writing, I deserve to get notified by SIGIO, IMO.
-is