Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe AT sources DOT redhat DOT com>
List-Archive: <http://sources.redhat.com/ml/cygwin/>
List-Post: <mailto:cygwin AT sources DOT redhat DOT com>
List-Help: <mailto:cygwin-help AT sources DOT redhat DOT com>, <http://sources.redhat.com/ml/#faqs>
Sender: cygwin-owner AT sources DOT redhat DOT com
Delivered-To: mailing list cygwin AT sources DOT redhat DOT com
From: "Markus Hoenicka" <Markus DOT Hoenicka AT uth DOT tmc DOT edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <14897.28668.934000.940116@gargle.gargle.HOWL>
Date: Fri, 8 Dec 2000 23:34:20 +0000 (CT0)
To: Corinna Vinschen <cygwin AT cygwin DOT com>
Subject: Re: socket read()/write() problem in 1.1.6
In-Reply-To: <00120813285702.12851@cygbert>
References: <1529f1887b DOT 1887b1529f AT uth DOT tmc DOT edu>
	<00120813285702 DOT 12851 AT cygbert>
X-Mailer: VM 6.75 under Emacs 20.6.1

Let me please add another important (I think) bit of information. In
my previous mail I reported that a call to fork() is essential to
reproduce the weird read()/write() problem. It turns out that only the
child is affected. If I modify the server to let the parent answer the
client request after forking, everything is fine again. This should
IMHO rule out any timing related issues.

A small change to server program will do the trick:
------------------------------------------------------
--- server5b.c	Fri Dec  8 16:49:27 2000
+++ server5c.c	Fri Dec  8 21:55:46 2000
@@ -52,22 +52,22 @@
 	  printf("adding client on fd %d\n", client_sockfd);
 	}
 	else {
-	  if (fork() == 0) { /* child */
+	  if (fork() != 0) { /* parent */
 	    iread(fd, ch, 10);
 	    sleep(1);
-	    printf("child serving client on fd %d; ch=%s<<\n", fd, ch);
+	    printf("parent serving client on fd %d; ch=%s<<\n", fd, ch);
 	    ch[0]++;
 	    iwrite(fd, ch, 5);
 	    iwrite(fd, ch+5, 5);
 	    close(fd);
 	    FD_CLR(fd, &readfds);
-	    printf("child done on fd %d; ch=%s<<\n", fd, ch);
-	    exit(0);
+	    printf("parent done on fd %d; ch=%s<<\n", fd, ch);
 	  }
-	  else { /* parent */
+	  else { /* child */
 	    close(fd);
 	    FD_CLR(fd, &readfds);
-	    printf("parent done on fd %d\n", fd);
+	    printf("child done on fd %d\n", fd);
+	    exit(0);
 	  }
 	}
       }
------------------------------------------------------

So my (un)educated guess is that it is less a read()/write() problem
per se but rather a problem how an inherited socket is set up in the
child process after forking.

regards,
Markus

-- 
Markus Hoenicka, PhD
UT Houston Medical School
Dept. of Integrative Biology and Pharmacology
6431 Fannin MSB4.114
Houston, TX 77030
(713) 500-6313, -7477
(713) 500-7444 (fax)
Markus DOT Hoenicka AT uth DOT tmc DOT edu
http://ourworld.compuserve.com/homepages/hoenicka_markus/


--
Want to unsubscribe from this list?
Send a message to cygwin-unsubscribe AT sourceware DOT cygnus DOT com