From owner-freebsd-hackers@FreeBSD.ORG  Thu Apr 24 18:21:28 2003
Return-Path: <owner-freebsd-hackers@FreeBSD.ORG>
Delivered-To: freebsd-hackers@freebsd.org
Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125])
	by hub.freebsd.org (Postfix) with ESMTP id 0BD8337B401
	for <freebsd-hackers@freebsd.org>;
	Thu, 24 Apr 2003 18:21:28 -0700 (PDT)
Received: from foem.leiden.webweaving.org (fia224-72.dsl.hccnet.nl
	[62.251.72.224])
	by mx1.FreeBSD.org (Postfix) with ESMTP id 4E6FC43F85
	for <freebsd-hackers@freebsd.org>;
	Thu, 24 Apr 2003 18:21:20 -0700 (PDT)
	(envelope-from dirkx@webweaving.org)
Received: from foem (foem [10.11.0.2])h3P1LIkF027570
	for <freebsd-hackers@freebsd.org>;
	Fri, 25 Apr 2003 03:21:18 +0200 (CEST)
	(envelope-from dirkx@webweaving.org)
Date: Fri, 25 Apr 2003 03:21:18 +0200 (CEST)
From: Dirk-Willem van Gulik <dirkx@webweaving.org>
X-X-Sender: dirkx@foem.leiden.webweaving.org
To: freebsd-hackers@freebsd.org
Message-ID: <20030425031450.G8323-100000@foem.leiden.webweaving.org>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Subject: recv() returning 0 and EINTR on a still connection.
X-BeenThere: freebsd-hackers@freebsd.org
X-Mailman-Version: 2.1.1
Precedence: list
List-Id: Technical Discussions relating to FreeBSD
	<freebsd-hackers.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-hackers>,
	<mailto:freebsd-hackers-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-hackers>
List-Post: <mailto:freebsd-hackers@freebsd.org>
List-Help: <mailto:freebsd-hackers-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-hackers>,
	<mailto:freebsd-hackers-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Fri, 25 Apr 2003 01:21:28 -0000


After an upgrade from 2.2.8 to 5.0 the following fragement of
code in userland seems to trip under high loads:

server-code
	n = recv(...
	if (n<0) {
		if (errno = EAGAIN)
			..
		if (errno = EINTR)
			..
	}
	if (n == 0)
		bing

with a n == 0 and errno == EINTR. (i.e. NOT n=-1).

Even though the connection does not seem closed from the client end.

Is it possible for n==0 -not- to mean that the connection is closed ?

Or in other words that recv() behaves not like read(2) in this respect.

(In all fairness; the man page for recv(2) does only detail n=-1 for the
EAGAIN. And does not mention any semantics for n==0, unlike read(2) which
says:

RETURN VALUES
     If successful, the number of bytes actually read is returned.  Upon read-
     ing end-of-file, zero is returned.  Otherwise, a -1 is returned and the
     global variable errno is set to indicate the error.

Is recv() fundamentally different ? And is n==0 errno=EINTR a possible
valid return with the meaning; interupted and no bytes read ?

Thanks

Dw