From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:34733)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <aurelien@aurel32.net>) id 1Z0cw5-0006xe-5o
	for qemu-devel@nongnu.org; Thu, 04 Jun 2015 17:44:26 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <aurelien@aurel32.net>) id 1Z0cw4-0001e2-6W
	for qemu-devel@nongnu.org; Thu, 04 Jun 2015 17:44:25 -0400
Received: from hall.aurel32.net ([2001:bc8:30d7:101::1]:45911)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <aurelien@aurel32.net>) id 1Z0cw4-0001dv-0F
	for qemu-devel@nongnu.org; Thu, 04 Jun 2015 17:44:24 -0400
Date: Thu, 4 Jun 2015 23:44:20 +0200
From: Aurelien Jarno <aurelien@aurel32.net>
Message-ID: <20150604214420.GA6313@aurel32.net>
References: <1433447607-31184-1-git-send-email-aurelien@aurel32.net>
	<1433447607-31184-5-git-send-email-aurelien@aurel32.net>
	<5570BE19.7090000@twiddle.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <5570BE19.7090000@twiddle.net>
Subject: Re: [Qemu-devel] [PATCH v1 4/5] tcg/optimize: fold constant test in
 tcg_opt_gen_mov
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Richard Henderson <rth@twiddle.net>
Cc: qemu-devel@nongnu.org

On 2015-06-04 14:07, Richard Henderson wrote:
> On 06/04/2015 12:53 PM, Aurelien Jarno wrote:
> > +
> >  static void tcg_opt_gen_mov(TCGContext *s, TCGOp *op, TCGArg *args,
> >                              TCGArg dst, TCGArg src)
> >  {
> > +    if (temps[src].state == TCG_TEMP_CONST) {
> > +        tcg_opt_gen_movi(s, op, args, dst, temps[src].val);
> > +        return;
> > +    }
> > +
> >      if (temps_are_copies(dst, src)) {
> >          tcg_op_remove(s, op);
> >          return;
> 
> I was going to comment re patch 3/5, but here's better, since it actually
> tidies up all the places I would have pointed out.
> 
> I think you should check for copies first, and only then do the TEMP_CONST
> check.  That way we get to remove an op.

If the state of a temp is TCG_TEMP_CONST, it can't be a copy as its
state should be TCG_TEMP_COPY for that. But it's true that the original
code is checking first for copies and then for constants.

-- 
Aurelien Jarno                          GPG: 4096R/1DDD8C9B
aurelien@aurel32.net                 http://www.aurel32.net