Bug#240070: marked as done (g++ generates wrong code)
Your message dated Sun, 28 Mar 2004 14:44:14 +0200
with message-id <16486.51358.996613.834520@gargle.gargle.HOWL>
and subject line Bug#240070: g++ generates wrong code
has caused the attached Bug report to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what I am
talking about this indicates a serious mail system misconfiguration
somewhere. Please contact me immediately.)
Debian bug tracking system administrator
(administrator, Debian Bugs database)
--------------------------------------
Received: (at submit) by bugs.debian.org; 25 Mar 2004 17:39:46 +0000
>From martin.haller@online.de Thu Mar 25 09:39:46 2004
Return-path: <martin.haller@online.de>
Received: from moutng.kundenserver.de [212.227.126.186]
by spohr.debian.org with esmtp (Exim 3.35 1 (Debian))
id 1B6YpW-00051m-00; Thu, 25 Mar 2004 09:39:46 -0800
Received: from [212.227.126.155] (helo=mrelayng.kundenserver.de)
by moutng.kundenserver.de with esmtp (Exim 3.35 #1)
id 1B6YpV-0007eu-00; Thu, 25 Mar 2004 18:39:45 +0100
Received: from [212.9.166.27] (helo=grobi.hamsternet)
by mrelayng.kundenserver.de with asmtp (TLSv1:EDH-RSA-DES-CBC3-SHA:168)
(Exim 3.35 #1)
id 1B6YpU-0003tP-00; Thu, 25 Mar 2004 18:39:45 +0100
Received: from mah by grobi.hamsternet with local (Exim 3.35 #1 (Debian))
id 1B6Yn2-0006gP-00; Thu, 25 Mar 2004 18:37:12 +0100
Date: Thu, 25 Mar 2004 18:37:12 +0100
From: Martin Haller <martin.haller@online.de>
To: submit@bugs.debian.org
Cc: mah@lf.net, martin.haller@online.de
Subject: g++ generates wrong code
Message-ID: <[🔎] 20040325173712.GA32667@grobi.hamsternet>
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="dDRMvlgZJXvWKvBx"
Content-Disposition: inline
User-Agent: Mutt/1.3.28i
Sender: Martin Haller <martin.haller@online.de>
X-Provags-ID: kundenserver.de abuse@kundenserver.de auth:f7ae1794aa6626a14ec6e854745623ce
Delivered-To: submit@bugs.debian.org
X-Spam-Checker-Version: SpamAssassin 2.60-bugs.debian.org_2004_03_25
(1.212-2003-09-23-exp) on spohr.debian.org
X-Spam-Status: No, hits=-7.0 required=4.0 tests=BAYES_00,HAS_PACKAGE
autolearn=no version=2.60-bugs.debian.org_2004_03_25
X-Spam-Level:
--dDRMvlgZJXvWKvBx
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Package: g++-2.95
Version: 2.95.4-11woody1
Package: g++-3.0
Version: 3.0.4-7
Hi,
the bug wich I describe below seems to be in both packages - this is
the reason why I write one report with two packages - sorry.
Problem:
--------
Compiling my test code with and without -O2 compiler option generates
different behaviors. The test code is attached (it is a TCP Server
wich select()s on the different connections - it's really ugly and
unstructured - please tell me if you don't come along with it!).
Reproduction:
-------------
- compile the attached server.cc mit the contributed Makefile; one
time with and one time without -O2 Option
- start the "server" binaries and connect to port 2000 with telnet:
$ telnet localhost 2000
- One time the connection will be established, the other time,
accept() will fail with error: invalid argument
I have reproduced this behavior on different Debian-3.0 boxes. To be
sure I tested it on FreeBSD and (old) Redhat as well: They don't have
this effect.
I have another test code example, wich one time segfaults and the
other time works. Please tell me if you are interested in more code.
Overview:
---------
System: Compiler: with -O2 without -O2
---------------------------------------------------------------------
Debian-3.0 g++-2.954-11woody1 works fails!
Debian-3.0 g++-3.0.4-7 fails! works
FreeBSD-4.9 g++-2.95.4 works works
Redhat-7.0 g++-2.96 20000731 works works
--
Martin
--dDRMvlgZJXvWKvBx
Content-Type: text/plain; charset=us-ascii
Content-Disposition: attachment; filename=Makefile
CC=g++-3.0
all:
$(CC) -c server.cc -g -O2
# $(CC) -c server.cc -g
$(CC) -o server server.o -lstdc++
clean:
rm -f *.o server
--dDRMvlgZJXvWKvBx
Content-Type: text/x-c++src; charset=us-ascii
Content-Disposition: attachment; filename="server.cc"
#include <stdio.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <netdb.h>
#include <unistd.h>
#include <string.h>
#include <stdlib.h>
#include <unistd.h>
#include <sys/types.h>
#include <sys/time.h>
#include <sys/poll.h>
#include <list.h>
typedef list<int> intlist;
void initFDSet(int *maxfds, fd_set *set, int serverSocket,
intlist *clients)
{
/* Initialize the file descriptor set. */
FD_ZERO (set);
FD_SET (serverSocket, set);
*maxfds = serverSocket;
for (list<int>::iterator i = clients->begin(); i != clients->end(); i++)
{
FD_SET(*i,set);
if (*i > *maxfds)
*maxfds = *i;
}
}
int main()
{
struct sockaddr_in sin;
struct sockaddr_in pin;
socklen_t addrlen;
int i;
struct timeval timeout;
int maxfds;
int _serverSocket;
int client;
list<int> clients;
fd_set set;
if ((_serverSocket = socket(PF_INET, SOCK_STREAM, 0)) == -1) {
perror("socket");
exit(1);
}
i = 1;
setsockopt(_serverSocket,SOL_SOCKET,SO_REUSEADDR, &i, sizeof(i));
/* complete the socket structure */
memset(&sin, 0, sizeof(sin));
sin.sin_family = AF_INET;
sin.sin_addr.s_addr = INADDR_ANY;
sin.sin_port = htons(2000);
/* bind the socket to the port number */
if (bind(_serverSocket, (struct sockaddr *) &sin, sizeof(sin)) == -1) {
perror("bind");
exit(1);
}
/* show that we are willing to listen */
if (listen(_serverSocket, 128) == -1) {
perror("listen");
exit(1);
}
/* Initialize the timeout data structure. */
timeout.tv_sec = 5;
timeout.tv_usec = 0;
initFDSet(&maxfds,&set,_serverSocket,&clients);
while (1)
{
select (maxfds + 1, &set, NULL, NULL, &timeout);
if (FD_ISSET(_serverSocket,&set))
{
// new connection
printf("new connection\n");
if ((client = accept(_serverSocket,
(struct sockaddr *) &pin, &addrlen)) == -1)
{
perror("accept");
exit(1);
}
clients.push_back(client);
}
for (list<int>::iterator i = clients.begin(); i != clients.end(); i++)
{
client = *i;
if (FD_ISSET(client,&set))
{
char buf[512];
int size = recv(client,buf,sizeof(buf),0);
if (! size)
{
clients.remove(client);
close(client);
printf("%X client closed connection\n",client);
break;
}
else
{
buf[size] = 0;
printf("%X recv: %d - %s\n",client,size,buf);
}
}
}
initFDSet(&maxfds,&set,_serverSocket,&clients);
}
}
--dDRMvlgZJXvWKvBx--
---------------------------------------
Received: (at 240070-done) by bugs.debian.org; 28 Mar 2004 12:49:02 +0000
>From doko@cs.tu-berlin.de Sun Mar 28 04:49:02 2004
Return-path: <doko@cs.tu-berlin.de>
Received: from mail.cs.tu-berlin.de [130.149.17.13] (root)
by spohr.debian.org with esmtp (Exim 3.35 1 (Debian))
id 1B7Zio-0001Ol-00; Sun, 28 Mar 2004 04:49:02 -0800
Received: from bolero.cs.tu-berlin.de (doko@bolero.cs.tu-berlin.de [130.149.19.1])
by mail.cs.tu-berlin.de (8.9.3p2/8.9.3) with ESMTP id OAA10754;
Sun, 28 Mar 2004 14:44:15 +0200 (MET DST)
Received: (from doko@localhost)
by bolero.cs.tu-berlin.de (8.12.10+Sun/8.12.8/Submit) id i2SCiFuw023824;
Sun, 28 Mar 2004 14:44:15 +0200 (MEST)
From: Matthias Klose <doko@cs.tu-berlin.de>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <16486.51358.996613.834520@gargle.gargle.HOWL>
Date: Sun, 28 Mar 2004 14:44:14 +0200
To: Martin Haller <martin.haller@online.de>, 240070-done@bugs.debian.org
Subject: Re: Bug#240070: g++ generates wrong code
In-Reply-To: <[🔎] 20040325173712.GA32667@grobi.hamsternet>
References: <[🔎] 20040325173712.GA32667@grobi.hamsternet>
X-Mailer: VM 7.03 under 21.4 (patch 6) "Common Lisp" XEmacs Lucid
Delivered-To: 240070-done@bugs.debian.org
X-Spam-Checker-Version: SpamAssassin 2.60-bugs.debian.org_2004_03_25
(1.212-2003-09-23-exp) on spohr.debian.org
X-Spam-Status: No, hits=-5.0 required=4.0 tests=BAYES_00,HAS_BUG_NUMBER
autolearn=no version=2.60-bugs.debian.org_2004_03_25
X-Spam-Level:
Fixed in the default version of the current testing/unstable
distributen (gcc-3.3.3).
Martin Haller writes:
> Package: g++-2.95
> Version: 2.95.4-11woody1
>
> Package: g++-3.0
> Version: 3.0.4-7
>
>
> Hi,
>
> the bug wich I describe below seems to be in both packages - this is
> the reason why I write one report with two packages - sorry.
>
> Problem:
> --------
>
> Compiling my test code with and without -O2 compiler option generates
> different behaviors. The test code is attached (it is a TCP Server
> wich select()s on the different connections - it's really ugly and
> unstructured - please tell me if you don't come along with it!).
>
> Reproduction:
> -------------
>
> - compile the attached server.cc mit the contributed Makefile; one
> time with and one time without -O2 Option
> - start the "server" binaries and connect to port 2000 with telnet:
> $ telnet localhost 2000
> - One time the connection will be established, the other time,
> accept() will fail with error: invalid argument
>
> I have reproduced this behavior on different Debian-3.0 boxes. To be
> sure I tested it on FreeBSD and (old) Redhat as well: They don't have
> this effect.
>
> I have another test code example, wich one time segfaults and the
> other time works. Please tell me if you are interested in more code.
>
>
> Overview:
> ---------
>
> System: Compiler: with -O2 without -O2
> ---------------------------------------------------------------------
> Debian-3.0 g++-2.954-11woody1 works fails!
> Debian-3.0 g++-3.0.4-7 fails! works
> FreeBSD-4.9 g++-2.95.4 works works
> Redhat-7.0 g++-2.96 20000731 works works
Reply to: