From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gdb-patches-return-37614-listarch-gdb-patches=sources.redhat.com@sources.redhat.com>
Received: (qmail 21897 invoked by alias); 28 Dec 2004 13:15:21 -0000
Mailing-List: contact gdb-patches-help@sources.redhat.com; run by ezmlm
Precedence: bulk
List-Subscribe: <mailto:gdb-patches-subscribe@sources.redhat.com>
List-Archive: <http://sources.redhat.com/ml/gdb-patches/>
List-Post: <mailto:gdb-patches@sources.redhat.com>
List-Help: <mailto:gdb-patches-help@sources.redhat.com>, <http://sources.redhat.com/ml/#faqs>
Sender: gdb-patches-owner@sources.redhat.com
Received: (qmail 21797 invoked from network); 28 Dec 2004 13:15:13 -0000
Received: from unknown (HELO NUTMEG.CAM.ARTIMI.COM) (217.40.111.177)
  by sourceware.org with SMTP; 28 Dec 2004 13:15:13 -0000
Received: from mace ([192.168.1.25]) by NUTMEG.CAM.ARTIMI.COM with Microsoft SMTPSVC(6.0.3790.211);
	 Tue, 28 Dec 2004 13:13:15 +0000
From: "Dave Korn" <dave.korn@artimi.com>
To: "'H. J. Lu'" <hjl@lucon.org>,
	<binutils@sources.redhat.com>
Cc: "'GDB'" <gdb-patches@sources.redhat.com>
Subject: RE: PATCH: Fix read_leb128 in readelf for 64bit host
Date: Tue, 28 Dec 2004 14:57:00 -0000
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
In-Reply-To: <20041227190130.GA21178@lucon.org>
Message-ID: <NUTMEGNE6f17ro43pwu00000a9c@NUTMEG.CAM.ARTIMI.COM>
X-OriginalArrivalTime: 28 Dec 2004 13:13:15.0031 (UTC) FILETIME=[FD54E270:01C4ECDE]
X-SW-Source: 2004-12/txt/msg00449.txt.bz2

> -----Original Message-----
> From: binutils-owner On Behalf Of H. J. Lu
> Sent: 27 December 2004 19:02

> On Sat, Dec 25, 2004 at 04:42:29PM -0800, H. J. Lu wrote:
> > read_leb128 in readelf assumes long == int == 32bit. It doesn't work
> > with 64bit host. Does this patch look right?
> > 
> > 
> 
> I am going to check in this patch. Gdb 6.3 has the same problem. I
> am enclosing a patch here.

> --- binutils/readelf.c.leb	2004-12-10 14:20:22.000000000 -0800
> +++ binutils/readelf.c	2004-12-27 10:49:33.689234088 -0800
> @@ -6933,7 +6933,7 @@ read_leb128 (unsigned char *data, int *l
>  {
>    unsigned long int result = 0;
>    unsigned int num_read = 0;
> -  int shift = 0;
> +  unsigned int shift = 0;
>    unsigned char byte;
>  
>    do
> @@ -6941,7 +6941,7 @@ read_leb128 (unsigned char *data, int *l
>        byte = *data++;
>        num_read++;
>  
> -      result |= (byte & 0x7f) << shift;
> +      result |= ((unsigned long int) (byte & 0x7f)) << shift;
>  
>        shift += 7;
>  
> @@ -6951,8 +6951,8 @@ read_leb128 (unsigned char *data, int *l
>    if (length_return != NULL)
>      *length_return = num_read;
>  
> -  if (sign && (shift < 32) && (byte & 0x40))
> -    result |= -1 << shift;
> +  if (sign && (shift < 8 * sizeof (result)) && (byte & 0x40))
> +    result |= -1L << shift;
>  
>    return result;
>  }
> 

  IIRC the C spec says shift amounts >=32 are undefined behaviour, even in the
presence of larger integer types.  Is shift ever going to be >= 32 for 64 bit
hosts in these functions?  (I suspect it will).  If so, the shift must be
decomposed into two smaller shifts, no?

    cheers, 
      DaveK
-- 
Can't think of a witty .sigline today....