From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-42169-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 20888 invoked by alias); 4 Dec 2001 06:34:35 -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 20867 invoked from network); 4 Dec 2001 06:34:32 -0000
Received: from unknown (HELO taltos.codesourcery.com) (66.92.14.122)
  by sources.redhat.com with SMTP; 4 Dec 2001 06:34:32 -0000
Received: from zack by taltos.codesourcery.com with local (Exim 3.33 #1 (Debian))
	id 16B99z-0004v4-00; Mon, 03 Dec 2001 22:34:31 -0800
Date: Mon, 03 Dec 2001 22:34:00 -0000
From: Zack Weinberg <zack@codesourcery.com>
To: Richard Henderson <rth@redhat.com>, gcc@gcc.gnu.org
Subject: Re: messy combine problem - PPC
Message-ID: <20011204063431.GB18650@codesourcery.com>
References: <20011203204934.GG292@codesourcery.com> <20011203150606.D7284@redhat.com> <20011203235312.GK292@codesourcery.com> <20011203182706.B7420@redhat.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20011203182706.B7420@redhat.com>
User-Agent: Mutt/1.3.24i
X-SW-Source: 2001-12/txt/msg00105.txt.bz2

On Mon, Dec 03, 2001 at 06:27:06PM -0800, Richard Henderson wrote:
> On Mon, Dec 03, 2001 at 03:53:12PM -0800, Zack Weinberg wrote:
> > I'll try this, but will it inhibit desirable optimizations?
> 
> No.

I thought about it a bit more, and it can't ever inhibit an optimization,
but I am still worried about not recording registers that die in call
instructions.  Consider

  result = (*fptr) (args...);

where fptr is not used after that, and combine will see RTL such as 

  (set (reg:SI result)
       (call (reg:P fptr)))

or is there something else that takes care of that?

> > > Also, that should really be using regs_invalidated_by_call rather
> > > than call_used_regs.  Not that it'll make too much difference, 
> > > since we don't do a whole lot of combining with fixed registers.
> > 
> > Huh?
> 
>        for (i = 0; i < FIRST_PSEUDO_REGISTER; i++)
> -        if (call_used_regs[i])
> +	 if (TEST_HARD_REG_BIT (regs_invalidated_by_call, i))

Oh, you meant the loop in record_dead_and_set_regs.

zw