From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-57215-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 31550 invoked by alias); 27 Jul 2002 04:20:07 -0000
Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Archive: <http://gcc.gnu.org/ml/gcc/>
List-Post: <mailto:gcc@gcc.gnu.org>
List-Help: <http://gcc.gnu.org/ml/>
Sender: gcc-owner@gcc.gnu.org
Received: (qmail 31543 invoked from network); 27 Jul 2002 04:20:05 -0000
Received: from unknown (HELO potter.sfbay.redhat.com) (205.180.83.107)
  by sources.redhat.com with SMTP; 27 Jul 2002 04:20:05 -0000
Received: from flamingo.sfbay.redhat.com (remus.sfbay.redhat.com [172.16.27.252])
	by potter.sfbay.redhat.com (8.11.6/8.11.6) with ESMTP id g6R4HvQ07649;
	Fri, 26 Jul 2002 21:17:58 -0700
Received: (from aldyh@localhost)
	by flamingo.sfbay.redhat.com (8.11.6/8.11.6) id g6R4Ud006121;
	Fri, 26 Jul 2002 21:30:39 -0700
X-Authentication-Warning: flamingo.sfbay.redhat.com: aldyh set sender to aldyh@redhat.com using -f
To: Joern Rennecke <joern.rennecke@superh.com>
Cc: Steve Ellcey <sje@cup.hp.com>, gcc@gcc.gnu.org, dave@hiauly1.hia.nrc.ca
Subject: Re: Vector modes under hppa64-hpu
References: <3D406473.5E2DF7AA@superh.com>
From: Aldy Hernandez <aldyh@redhat.com>
Date: Sat, 27 Jul 2002 07:35:00 -0000
In-Reply-To: <3D406473.5E2DF7AA@superh.com>
Message-ID: <m31y9px05s.fsf@flamingo.sfbay.redhat.com>
User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.1.90
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-SW-Source: 2002-07/txt/msg01304.txt.bz2

>>>>> "Joern" == Joern Rennecke <joern.rennecke@superh.com> writes:

 >   	  else
 >   	    t = simplify_gen_subreg (submode, target, mode, i * subsize);
 > ! 	  if (CONSTANT_P (op0) || GET_CODE (op0) == REG)
 >   	    a = simplify_gen_subreg (submode, op0, mode, i * subsize);
 >   	  else
 >   	    a = extract_bit_field (op0, subbitsize, i * subbitsize, unsignedp,
 >   				   NULL_RTX, submode, submode, size);
 > ! 	  if (CONSTANT_P (op1) || GET_CODE (op1) == REG)
 >   	    b = simplify_gen_subreg (submode, op1, mode, i * subsize);
 >   	  else
 >   	    b = extract_bit_field (op1, subbitsize, i * subbitsize, unsignedp,


 > That doesn't work when submode is smaller than a word.  At register allocation
 > time, every subreg of a single hard register is assumed to be a low part,
 > even if its not.

Joern.  Could we keep the above code and add a check for the submode
being >= word?  It seems like the extract_bit_field is a bit of
overkill when the submode is just the word size.

What do you think?

Aldy