From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-patches-return-205757-listarch-gcc-patches=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 26989 invoked by alias); 12 Oct 2007 16:40:37 -0000
Received: (qmail 26938 invoked by uid 22791); 12 Oct 2007 16:40:36 -0000
X-Spam-Check-By: sourceware.org
Received: from sccrmhc12.comcast.net (HELO sccrmhc12.comcast.net) (204.127.200.82)     by sourceware.org (qpsmtpd/0.31) with ESMTP; Fri, 12 Oct 2007 16:40:29 +0000
Received: from lucon.org ([24.6.230.138]) by comcast.net (sccrmhc12) with ESMTP           id <2007101216402701200eb7joe>; Fri, 12 Oct 2007 16:40:27 +0000
Received: by lucon.org (Postfix, from userid 500) 	id 0CA1EF82E4; Fri, 12 Oct 2007 09:40:27 -0700 (PDT)
Date: Fri, 12 Oct 2007 16:40:00 -0000
From: "H.J. Lu" <hjl@lucon.org>
To: Uros Bizjak <ubizjak@gmail.com>, GCC Patches <gcc-patches@gcc.gnu.org>, 	Jan Hubicka <jh@suse.cz>
Subject: Re: [PATCH,i386] fix PR 11001
Message-ID: <20071012164026.GA1318@lucon.org>
References: <5787cf470710032334y2aefe839x8981f4c84b9a2662@mail.gmail.com> <20071004205533.GG4491@codesourcery.com> <5787cf470710042259x3428b010y7d8bb8ac189157a0@mail.gmail.com> <20071012161246.GE11809@codesourcery.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20071012161246.GE11809@codesourcery.com>
User-Agent: Mutt/1.5.14 (2007-02-12)
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: 2007-10/txt/msg00742.txt.bz2

On Fri, Oct 12, 2007 at 09:12:47AM -0700, Nathan Froyd wrote:
> On Fri, Oct 05, 2007 at 07:59:04AM +0200, Uros Bizjak wrote:
> > Sorry for not being clear here. TARGET_INLINE_ALL_STRINGOPS* should be
> > respected, and libcalls should not be generated in this case. For
> > example, we use -minline-all-stringops to build crtfastmath.o. For
> > some targets, we don't  want to link against libc to use memset
> > routine.
> > 
> > The patch is OK  with above change, but please wait a day or two for
> > possilbe comments from Jan, as he is the original author of stringops
> > code.
> 
> Thanks, the below is what I committed.
> 
> -Nathan
> 
> gcc/
> 	PR 11001
> 	* config/i386/i386.md (strmov): Check for esi and edi usage.
> 	* config/i386/i386.c (decide_alg): Check whether we can use a
> 	rep prefix and adjust algorithm choice accordingly.
> 	(ix86_expand_strlen): Check for eax, ecx, and edi usage.
> 
> Index: gcc/config/i386/i386.c
> ===================================================================
> --- gcc/config/i386/i386.c	(revision 129263)
> +++ gcc/config/i386/i386.c	(working copy)
> @@ -15056,21 +15056,32 @@ decide_alg (HOST_WIDE_INT count, HOST_WI
>  	    int *dynamic_check)
>  {
>    const struct stringop_algs * algs;
> +  /* Algorithms using the rep prefix want at least edi and ecx;
> +     additionally, memset wants eax and memcpy wants esi.  Don't
> +     consider such algorithms if the user has appropriated those
> +     registers for their own purposes.	*/
> +  bool rep_prefix_usable = !(global_regs[2] || global_regs[5]
> +                             || (memset ? global_regs[0] : global_regs[4]));
> +
> +
> +      /* Can't use this if the user has appropriated eax, ecx, or edi.  */
> +      if (global_regs[0] || global_regs[2] || global_regs[5])
> +        return false;
> +

Can you use/add AX_REG, CX_REG, DI_REG, SI_REG instead using 0, 2,
4, 5?

Thanks.

H.J.