From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-patches-return-251349-listarch-gcc-patches=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 17284 invoked by alias); 20 Oct 2009 18:09:48 -0000
Received: (qmail 17269 invoked by uid 22791); 20 Oct 2009 18:09:47 -0000
X-SWARE-Spam-Status: No, hits=-2.5 required=5.0 	tests=AWL,BAYES_00,SPF_HELO_PASS,SPF_PASS
X-Spam-Check-By: sourceware.org
Received: from mx1.redhat.com (HELO mx1.redhat.com) (209.132.183.28)     by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Tue, 20 Oct 2009 18:09:42 +0000
Received: from int-mx02.intmail.prod.int.phx2.redhat.com (int-mx02.intmail.prod.int.phx2.redhat.com [10.5.11.12]) 	by mx1.redhat.com (8.13.8/8.13.8) with ESMTP id n9KI9dS4025111; 	Tue, 20 Oct 2009 14:09:39 -0400
Received: from greed.delorie.com (ovpn01.gateway.prod.ext.phx2.redhat.com [10.5.9.1]) 	by int-mx02.intmail.prod.int.phx2.redhat.com (8.13.8/8.13.8) with ESMTP id n9KI9cCe013145 	(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); 	Tue, 20 Oct 2009 14:09:39 -0400
Received: from greed.delorie.com (greed.delorie.com [127.0.0.1] (may be forged)) 	by greed.delorie.com (8.14.3/8.14.3) with ESMTP id n9KI9bpn008331; 	Tue, 20 Oct 2009 14:09:37 -0400
Received: (from dj@localhost) 	by greed.delorie.com (8.14.3/8.14.3/Submit) id n9KI9aBM008328; 	Tue, 20 Oct 2009 14:09:36 -0400
Date: Tue, 20 Oct 2009 18:10:00 -0000
Message-Id: <200910201809.n9KI9aBM008328@greed.delorie.com>
From: DJ Delorie <dj@redhat.com>
To: Richard Guenther <richard.guenther@gmail.com>
CC: uweigand@de.ibm.com, gcc-patches@gcc.gnu.org
In-reply-to: <84fc9c000910200212r13cf361bpca6f54071e6b0da@mail.gmail.com> 	(message from Richard Guenther on Tue, 20 Oct 2009 11:12:59 +0200)
Subject: Re: m32c support for named addr spaces branch
References: <200910191743.n9JHhBIP025038@d12av02.megacenter.de.ibm.com> 	 <200910192018.n9JKIg0f013560@greed.delorie.com> <84fc9c000910200212r13cf361bpca6f54071e6b0da@mail.gmail.com>
X-IsSubscribed: yes
Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-patches.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-patches/>
List-Post: <mailto:gcc-patches@gcc.gnu.org>
List-Help: <mailto:gcc-patches-help@gcc.gnu.org>
Sender: gcc-patches-owner@gcc.gnu.org
X-SW-Source: 2009-10/txt/msg01314.txt.bz2


> I don't think this observation is correct.  sizetypes are _always_
> sign-extended.

The problem on m32c is like this: size_t is HImode, pointers are
PSImode (24 bits).  Loop optimization sees this:

	char *p = x;
	for (i=0; i<j; i++)
	  *p++ = n;

It replaces the 'i' loop with a 'p'-based loop.  It pre-calculates the
end value of 'p'.  However, it sign extends 'j' before adding it to
'p', or worse, casts 'p' to size_t, does the math, the re-casts it
back to a pointer.

Thus, you end up with the termination value of 'p' being truncated or
mis-computed, and the loop overruns.

My preference is this: pointers and integers are NOT interchangable.
Pointer offsets must be sign or zero extended according to their
underlying types before being added to pointers.  If I add -1 to a
pointer, it had better result in a pointer that's one less than the
previous value, not 65536 more.