From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-93082-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 12030 invoked by alias); 19 Mar 2004 18:57:41 -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 12016 invoked from network); 19 Mar 2004 18:57:39 -0000
Received: from unknown (HELO mail-out3.apple.com) (17.254.13.22)
  by sources.redhat.com with SMTP; 19 Mar 2004 18:57:39 -0000
Received: from mailgate1.apple.com (a17-128-100-225.apple.com [17.128.100.225])
	by mail-out3.apple.com (8.12.11/8.12.11) with ESMTP id i2JIvdF1011461
	for <gcc@gcc.gnu.org>; Fri, 19 Mar 2004 10:57:39 -0800 (PST)
Received: from relay2.apple.com (relay2.apple.com) by mailgate1.apple.com
 (Content Technologies SMTPRS 4.3.6) with ESMTP id <T686f2d9c41118064e13b0@mailgate1.apple.com>;
 Fri, 19 Mar 2004 10:57:38 -0800
Received: from [17.201.20.186] (gambrinus.apple.com [17.201.20.186])
	by relay2.apple.com (8.12.11/8.12.11) with ESMTP id i2JIvLoM000977;
	Fri, 19 Mar 2004 18:57:22 GMT
In-Reply-To: <OF6E029669.0E5E3F05-ONC2256E5C.000482C4-C2256E5C.0004D314@il.ibm.com>
References: <OF6E029669.0E5E3F05-ONC2256E5C.000482C4-C2256E5C.0004D314@il.ibm.com>
Mime-Version: 1.0 (Apple Message framework v612)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <FA3D1E5D-79D6-11D8-A952-000A95D7CD40@apple.com>
Content-Transfer-Encoding: 7bit
Cc: gcc@gcc.gnu.org, David Edelsohn <dje@makai.watson.ibm.com>,
        Dale Johannesen <dalej@apple.com>
From: Dale Johannesen <dalej@apple.com>
Subject: Re: rs6000.md/altivec.md problem in setting of vector registers
Date: Fri, 19 Mar 2004 20:59:00 -0000
To: Dorit Naishlos <DORIT@il.ibm.com>
X-SW-Source: 2004-03/txt/msg01163.txt.bz2

On Mar 18, 2004, at 4:52 PM, Dorit Naishlos wrote:
> I managed to hack something that causes Reload to make the "right" 
> decision
> and generate the same code as it generates for i386 - i.e., the spill 
> code
> is created out of the loop.
>
> The hack modifies the macro CANNOT_CHANGE_MODE_CLASS in rs6000.h to not
> allow a mode change from a vector mode to a smaller mode in 
> ALTIVEC_REGS or
> GENERAL_REGS. As a result, these register classes cannot be considered 
> for
> allocation in this case; instead, Reload directly creates stores to 
> memory,
> outside the loop, like for i386.

I haven't tried it, but this might well break passing of vector 
parameters in int regs.