Bug#350636: marked as done (gcc-4.0: compiles wrong code with -O2)
Your message dated Tue, 31 Jan 2006 00:53:09 +0100
with message-id <8764o1jmsa.fsf@debian.org>
and subject line Bug#350636: code that reproduces the bug and more info
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; 30 Jan 2006 22:15:45 +0000
>From Thorsten_Jordan@gmx.de Mon Jan 30 14:15:45 2006
Return-path: <Thorsten_Jordan@gmx.de>
Received: from mail.gmx.net ([213.165.64.21])
by spohr.debian.org with smtp (Exim 4.50)
id 1F3hJJ-000899-6g
for submit@bugs.debian.org; Mon, 30 Jan 2006 14:15:45 -0800
Received: (qmail invoked by alias); 30 Jan 2006 22:15:12 -0000
Received: from p54A7BE26.dip0.t-ipconnect.de (EHLO [192.168.178.19]) [84.167.190.38]
by mail.gmx.net (mp032) with SMTP; 30 Jan 2006 23:15:12 +0100
X-Authenticated: #779194
Message-ID: <[🔎] 43DE9024.9070307@gmx.de>
Date: Mon, 30 Jan 2006 23:16:04 +0100
From: Thorsten Jordan <Thorsten_Jordan@gmx.de>
User-Agent: Mozilla Thunderbird 1.0.7 (X11/20051017)
X-Accept-Language: en
MIME-Version: 1.0
To: Debian Bug Tracking System <submit@bugs.debian.org>
Subject: gcc-4.0: compiles wrong code with -O2
X-Enigmail-Version: 0.93.0.0
Content-Type: text/plain; charset=ISO-8859-15
Content-Transfer-Encoding: 7bit
X-Y-GMX-Trusted: 0
Delivered-To: submit@bugs.debian.org
X-Spam-Checker-Version: SpamAssassin 2.60-bugs.debian.org_2005_01_02
(1.212-2003-09-23-exp) on spohr.debian.org
X-Spam-Level:
X-Spam-Status: No, hits=-8.0 required=4.0 tests=BAYES_00,HAS_PACKAGE
autolearn=no version=2.60-bugs.debian.org_2005_01_02
Package: gcc-4.0
Version: 4.0.2-8
Severity: normal
i am reading from a stream in c++. The read values differ depending on
compiler optimization level.
Note that i have a little endian machine (x86)!
take this code
inline uint32_t read_u32(istream& in) {
uint32_t i;
in.read((char*)&i, 4);
return i;
}
and elsewhere it is called
...
uint32_t a = read_u32(in);
uint32_t b = read_u32(in);
uint32_t c = read_u32(in);
when i use the values later they're wrong. If i printf() them directly
after the read calls, they are ok later, same when i call a pseudo
function with each value, that justs assigns it to a global variable.
This happens only with -O2, not with -O0 or -O1.
I compared the assembler code of both versions, but could not find
relevant differences, except the call and some stack operations. Maybe
the inlining makes the difference.
Note that the bug does not occour in a simple example source, but the
code is user in a bigger context. I'm afraid this description won't help
much, though...
Note that gcc-3.4 compiles the code without problems and the results are ok.
-- System Information:
Debian Release: testing/unstable
APT prefers unstable
APT policy: (500, 'unstable')
Architecture: i386 (i686)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux 2.6.15
Locale: LANG=de_DE, LC_CTYPE=de_DE (charmap=ISO-8859-1)
Versions of packages gcc-4.0 depends on:
ii binutils 2.16.1cvs20060117-1 The GNU assembler, linker
and bina
ii cpp-4.0 4.0.2-8 The GNU C preprocessor
ii gcc-4.0-base 4.0.2-8 The GNU Compiler Collection
(base
ii libc6 2.3.5-12 GNU C Library: Shared
libraries an
ii libgcc1 1:4.0.2-8 GCC support library
Versions of packages gcc-4.0 recommends:
ii libc6-dev 2.3.5-12 GNU C Library: Development
Librari
ii libmudflap0-dev 4.0.2-8 GCC mudflap support
libraries (dev
-- no debconf information
---------------------------------------
Received: (at 350636-done) by bugs.debian.org; 30 Jan 2006 23:53:42 +0000
>From falk@debian.org Mon Jan 30 15:53:41 2006
Return-path: <falk@debian.org>
Received: from smtp05.web.de ([217.72.192.209])
by spohr.debian.org with esmtp (Exim 4.50)
id 1F3iq5-0006hI-P4
for 350636-done@bugs.debian.org; Mon, 30 Jan 2006 15:53:41 -0800
Received: from [84.59.221.119] (helo=juist)
by smtp05.web.de with asmtp (TLSv1:AES256-SHA:256)
(WEB.DE 4.105 #340)
id 1F3ipa-0007hg-00
for 350636-done@bugs.debian.org; Tue, 31 Jan 2006 00:53:10 +0100
Received: from falk by juist with local (Exim 4.60)
(envelope-from <falk@debian.org>)
id 1F3ipZ-0002u8-Kd
for 350636-done@bugs.debian.org; Tue, 31 Jan 2006 00:53:09 +0100
To: 350636-done@bugs.debian.org
Subject: Re: Bug#350636: code that reproduces the bug and more info
References: <[🔎] 43DE97AB.5060706@gmx.de>
From: Falk Hueffner <falk@debian.org>
X-Face: "iUeUu$b*W_"w?tV83Y3*r:`rh&dRv}$YnZ3,LVeCZSYVuf[Gpo*5%_=/\_!gc_,SS}[~xZ
wY77I-M)xHIx:2f56g%/`SOw"Dx%4Xq0&f\Tj~>|QR|vGlU}TBYhiG(K:2<T^
Date: Tue, 31 Jan 2006 00:53:09 +0100
In-Reply-To: <[🔎] 43DE97AB.5060706@gmx.de> (Thorsten Jordan's message of "Mon,
30 Jan 2006 23:48:11 +0100")
Message-ID: <8764o1jmsa.fsf@debian.org>
User-Agent: Gnus/5.1006 (Gnus v5.10.6) XEmacs/21.5 (cilantro, linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: falk@debian.org
X-Sender: falk.hueffner@web.de
X-Spam-Checker-Version: SpamAssassin 2.60-bugs.debian.org_2005_01_02
(1.212-2003-09-23-exp) on spohr.debian.org
X-Spam-Level:
X-Spam-Status: No, hits=-6.0 required=4.0 tests=BAYES_00,HAS_BUG_NUMBER
autolearn=no version=2.60-bugs.debian.org_2005_01_02
Thorsten Jordan <Thorsten_Jordan@gmx.de> writes:
> inline float read_float(std::istream& in)
> {
> Uint32 u = read_u32(in);
> return *(float*)(&u);
> }
This accesses u, which is of type unsigned, by an lvalue of type
float. That is not allowed in C++. Newer g++ will warn about it:
falk@juist:/tmp% g++ -Wall -O2 test.cc
test.cc: In function 'float read_float(std::istream&)':
test.cc:16: warning: dereferencing type-punned pointer will break strict-aliasing rules
--
Falk
Reply to: