Date: Thu, 14 Dec 2000 17:55:58 +0000 From: David Malone <dwmalone@maths.tcd.ie> To: mikko@dynas.se Cc: David Malone <dwmalone@maths.tcd.ie>, freebsd-hackers@freebsd.org, dwmalone@maths.tcd.ie Subject: Re: Rejecting a connection: is accept(2) correct? Message-ID: <200012141755.aa62114@salmon.maths.tcd.ie> In-Reply-To: Your message of "Thu, 14 Dec 2000 09:44:15 PST." <Pine.BSF.4.21.0012140936520.336-100000@explorer.rsa.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> Umm.. yes, that is the normal way of doing things, but that is not
> what (I think) it says in the man-page, which is why I was wondering.
> "One can obtain user connection request data without confirming the
> connection..." (and then goes on about recvmsg(),sendmsg() and
> {set,get}sockept() with little or no details).
> It does not state which socket this is supposed to operate on: the
> listening socket, or the new one. If it is the listening socket,
> it would be an innovation, and open up assorted new possibilities.
I'd guess it's the new socket 'cos the previous paragraph of the
man page talks about protocols which require explicit confirmation,
which can be caused by reading or writing to the socket. I'd guess
it's just indicating that recvmsg with a zero iovlen doesn't count
as a read or write.
> If it is the new socket, it is old news :-)
If it does work, it would mean that we could use tcp wrappers on
`stream wait' services, among other things. Maybe we need an API
for this.
David.
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi? <200012141755.aa62114>