From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-prs-return-30325-listarch-gcc-prs=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 10936 invoked by alias); 2 Mar 2003 22:46:01 -0000
Mailing-List: contact gcc-prs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Archive: <http://gcc.gnu.org/ml/gcc-prs/>
List-Post: <mailto:gcc-prs@gcc.gnu.org>
List-Help: <mailto:gcc-prs-help@gcc.gnu.org>
Sender: gcc-prs-owner@gcc.gnu.org
Received: (qmail 10914 invoked by uid 71); 2 Mar 2003 22:46:01 -0000
Resent-Date: 2 Mar 2003 22:46:01 -0000
Resent-Message-ID: <20030302224601.10913.qmail@sources.redhat.com>
Resent-From: gcc-gnats@gcc.gnu.org (GNATS Filer)
Resent-Cc: gcc-prs@gcc.gnu.org, gcc-bugs@gcc.gnu.org
Resent-Reply-To: gcc-gnats@gcc.gnu.org, kfoltman@onet.pl
Received: (qmail 8856 invoked by uid 48); 2 Mar 2003 22:36:44 -0000
Message-Id: <20030302223644.8855.qmail@sources.redhat.com>
Date: Sun, 02 Mar 2003 22:46:00 -0000
From: kfoltman@onet.pl
Reply-To: kfoltman@onet.pl
To: gcc-gnats@gcc.gnu.org
X-Send-Pr-Version: gnatsweb-2.9.3 (1.1.1.1.2.31)
Subject: target/9908: wrong code generated for virtual method call (gas bug?)
X-SW-Source: 2003-03/txt/msg00055.txt.bz2
List-Id: <gcc-prs.sourceware.org>


>Number:         9908
>Category:       target
>Synopsis:       wrong code generated for virtual method call (gas bug?)
>Confidential:   no
>Severity:       critical
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          wrong-code
>Submitter-Id:   net
>Arrival-Date:   Sun Mar 02 22:46:01 UTC 2003
>Closed-Date:
>Last-Modified:
>Originator:     Krzysztof Foltman
>Release:        gcc-3.2.2
>Organization:
>Environment:
Debian-x86, mingw - both give same effect
>Description:
Wrong call type is generated in -masm=intel mode. I'm not sure, but it looks like the .S file contains correct call (call dword ptr label+8), but it's assembled as near jump (E8 08 00 00 00) instead of indirect jump (FF 15 08 00 00 00). Checked by compiling both AT&T and Intel versions with -S, assembling with as and unassembling with ndisasmw.
>How-To-Repeat:
========== cut =======================

struct B {
  virtual int foo() { }
};

void check(void *pV)
{
  ((B *)pV)->foo();
}

int main(int argc, char *argv)
{
  B b;
  check((B *)&b);
}

========== cut =======================

# gcc-3.2 -O3 -masm=att -o inh inh.cpp -l stdc++ && ./inh
(this works)

# gcc-3.2 -O3 -masm=intel -o inh inh.cpp -l stdc++ && ./inh
(this segfaults)

100% repeatable both on mingw and debian gcc-3.2
>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted: