From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43)
	id 1MsLoB-0003yP-DG
	for qemu-devel@nongnu.org; Mon, 28 Sep 2009 15:22:51 -0400
Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43)
	id 1MsLoA-0003xC-1c
	for qemu-devel@nongnu.org; Mon, 28 Sep 2009 15:22:50 -0400
Received: from [199.232.76.173] (port=58803 helo=monty-python.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1MsLo9-0003x2-RQ
	for qemu-devel@nongnu.org; Mon, 28 Sep 2009 15:22:49 -0400
Received: from hall.aurel32.net ([88.191.82.174]:51108)
	by monty-python.gnu.org with esmtps (TLS-1.0:RSA_AES_256_CBC_SHA1:32)
	(Exim 4.60) (envelope-from <aurelien@aurel32.net>)
	id 1MsLo9-0006be-BF
	for qemu-devel@nongnu.org; Mon, 28 Sep 2009 15:22:49 -0400
Date: Mon, 28 Sep 2009 21:22:46 +0200
From: Aurelien Jarno <aurelien@aurel32.net>
Subject: Re: [Qemu-devel] [RFC] tcg: store constants without using
	registers when possible
Message-ID: <20090928192246.GA6691@hall.aurel32.net>
References: <20090928091208.GB22663@hall.aurel32.net>
	<f43fc5580909281108p1720e98dm421b858715da6392@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-15
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <f43fc5580909281108p1720e98dm421b858715da6392@mail.gmail.com>
Sender: Aurelien Jarno <aurelien@aurel32.net>
List-Id: qemu-devel.nongnu.org
List-Unsubscribe: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.gnu.org/pipermail/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Blue Swirl <blauwirbel@gmail.com>
Cc: qemu-devel@nongnu.org

On Mon, Sep 28, 2009 at 09:08:32PM +0300, Blue Swirl wrote:
> On Mon, Sep 28, 2009 at 12:12 PM, Aurelien Jarno <aurelien@aurel32.net> wrote:
> > Currently only implemented for x86/x86_64. It may also be implemented
> > for targets that keep one register for TCG internal use.
> 
> Nice, but do you think the op will be used often?

It is used each time a constant is written to a global TCG variable and
written back to memory. A constant propagation patch like the ones Filip
proposed a few months ago increase the number of those cases.

> > --- a/tcg/sparc/tcg-target.c
> > +++ b/tcg/sparc/tcg-target.c
> > @@ -392,6 +392,15 @@ static inline void tcg_out_st(TCGContext *s, TCGType type, int arg,
> >         tcg_out_ldst(s, arg, arg1, arg2, STX);
> >  }
> >
> > +static inline int tcg_out_sti(TCGContext *s, TCGType type, tcg_target_long val,
> > +                              int arg1, tcg_target_long arg2)
> > +{
> > +    /* movl */
> > +    tcg_out_modrm_offset(s, 0xc7, 0, arg1, arg2);
> > +    tcg_out32(s, val);
> > +    return 1;
> > +}
> > +
> 
> Sparc doesn't have store immediate ops, except for val == 0 case:
> 
> static inline int tcg_out_sti(TCGContext *s, TCGType type, tcg_target_long val,
>                               int arg1, tcg_target_long arg2)
> {
>   if (val != 0) {
>     return 0;
>   }
>   /* clr/clrx */
>   tcg_out_st(s, type, TCG_REG_G0, arg1, arg2);
>   return 1;
> }
> 

-- 
Aurelien Jarno	                        GPG: 1024D/F1BCDB73
aurelien@aurel32.net                 http://www.aurel32.net