From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-74590-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 10033 invoked by alias); 19 May 2003 15:02:09 -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 9989 invoked from network); 19 May 2003 15:02:08 -0000
Received: from unknown (HELO localhost.localdomain) (195.113.19.66)
  by sources.redhat.com with SMTP; 19 May 2003 15:02:08 -0000
Received: from sunsite.ms.mff.cuni.cz (sunsite.mff.cuni.cz [127.0.0.1])
	by localhost.localdomain (8.12.8/8.12.8) with ESMTP id h4JF21qO001512;
	Mon, 19 May 2003 17:02:01 +0200
Received: (from jakub@localhost)
	by sunsite.ms.mff.cuni.cz (8.12.8/8.12.8/Submit) id h4JF21LM001506;
	Mon, 19 May 2003 17:02:01 +0200
Date: Mon, 19 May 2003 15:06:00 -0000
From: Jakub Jelinek <jakub@redhat.com>
To: "H. J. Lu" <hjl@lucon.org>
Cc: Richard Henderson <rth@redhat.com>, Fergus Henderson <fjh@cs.mu.oz.au>,
   John David Anglin <dave@hiauly1.hia.nrc.ca>, gcc@gcc.gnu.org,
   schwab@suse.de
Subject: Re: Does gcc violate the ia64 ABI?
Message-ID: <20030519150200.GB16629@sunsite.ms.mff.cuni.cz>
Reply-To: Jakub Jelinek <jakub@redhat.com>
References: <20030517202234.A28227@lucon.org> <200305182251.h4IMpsCD017880@hiauly1.hia.nrc.ca> <20030519030139.GA23221@ceres.cs.mu.oz.au> <20030519053855.GA25079@redhat.com> <20030519075540.A17193@lucon.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20030519075540.A17193@lucon.org>
User-Agent: Mutt/1.4i
X-SW-Source: 2003-05/txt/msg01761.txt.bz2

On Mon, May 19, 2003 at 07:55:40AM -0700, H. J. Lu wrote:
> > Only tail calls to provably local functions, apparently.
> > I think that's pretty stupid, considering how cheap it is
> > to reload the gp...
> 
> I disagree. Gcc can still do tail calls to external functions as long
> as it reloads gp, which gcc is already doing when gp is used after
> the call, like in this code
> 
> ---
> extern void bar ();
> 
> static void
> bar1 ()
> {
>   bar ();
> }
> 
> extern int x;
> 
> void
> foo ()
> {
>   bar1 ();
>   x += 1;
> }
> ---
> 
> I think we just need to change the code from
> 
> 	reload_gp = gp is used after the tail call
> 
> to
> 
> 	reload_gp = the tail call is external
> 
> It may be a win since we won't reload gp for the local call even if
> gp is used after the tail call. Besides, you have said reloading gp
> is cheap. I don't see fixing gcc will hurt performance.

If bar1 has to reload gp after bar returns, then certainly the call to bar
cannot be a tail call (as a tail call returns from bar stright to foo).

	Jakub