Mailing-List: contact cygwin-developers-help AT cygwin DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-developers-subscribe AT cygwin DOT com>
List-Archive: <http://sources.redhat.com/ml/cygwin-developers/>
List-Post: <mailto:cygwin-developers AT cygwin DOT com>
List-Help: <mailto:cygwin-developers-help AT cygwin DOT com>, <http://sources.redhat.com/ml/#faqs>
Sender: cygwin-developers-owner AT cygwin DOT com
Delivered-To: mailing list cygwin-developers AT cygwin DOT com
Message-ID: <015e01c243ce$8390e3f0$6132bc3e@BABEL>
From: "Conrad Scott" <Conrad DOT Scott AT dsl DOT pipex DOT com>
To: <cygwin-developers AT cygwin DOT com>
Subject: getsockname / recvfrom bug (?)
Date: Wed, 14 Aug 2002 21:09:36 +0100
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000

Having changed the socket code to use recvfrom, I've now found an
apparent bug whereby recvfrom will fail with WSAEINVAL.  This was
already a problem in cygwin (AFAICT) but my change to make all
reads use recvfrom has made it potentially more troublesome.
(Interestingly, I see the same problem with getsockname but not,
e.g., with recv, or send, or sendto even).

Anyhow the problem occurs if you dup(2) a socket and then operate
on the duplicate copy w/o closing the original socket and without
forking; i.e. it's unlikely that anything will be affected since
dup'ing a socket is normally followed by one or both of those two
actions, and then the bug doesn't show.

I'm just writing a raw winsock2 program to check what's going on
but I thought I'd bring it to Chris and Corinna's attention
particularly, in case they'd rather not release that last patch of
mine until I've found some workaround.  As I said, it's not a new
bug and it's only exposed under slightly wierd circumstances, but
I've made it a little more likely to happen.

As soon as I've any more information on this, I'll get back to the
list.  BTW, if this is a "well known" problem, I'd be interested
in that information too (and yes, I''ve checked the mailing list
archives to no avail as yet).

Cheers,

// Conrad