From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:37925)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <rth7680@gmail.com>) id 1VoykJ-0004Mc-Tk
	for qemu-devel@nongnu.org; Fri, 06 Dec 2013 11:59:28 -0500
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <rth7680@gmail.com>) id 1VoykB-00025n-GQ
	for qemu-devel@nongnu.org; Fri, 06 Dec 2013 11:59:19 -0500
Received: from mail-yh0-x22a.google.com ([2607:f8b0:4002:c01::22a]:52067)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <rth7680@gmail.com>) id 1VoykB-00025f-DI
	for qemu-devel@nongnu.org; Fri, 06 Dec 2013 11:59:11 -0500
Received: by mail-yh0-f42.google.com with SMTP id z6so662112yhz.15
	for <qemu-devel@nongnu.org>; Fri, 06 Dec 2013 08:59:10 -0800 (PST)
Sender: Richard Henderson <rth7680@gmail.com>
Message-ID: <52A20256.3040706@twiddle.net>
Date: Sat, 07 Dec 2013 05:59:02 +1300
From: Richard Henderson <rth@twiddle.net>
MIME-Version: 1.0
References: <1386335953-28876-1-git-send-email-peter.maydell@linaro.org>
	<1386335953-28876-2-git-send-email-peter.maydell@linaro.org>
In-Reply-To: <1386335953-28876-2-git-send-email-peter.maydell@linaro.org>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [PATCH v2 01/13] target-arm: A64: add support for
 conditional select
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: Peter Maydell <peter.maydell@linaro.org>, qemu-devel@nongnu.org
Cc: patches@linaro.org, Michael Matz <matz@suse.de>, Claudio Fontana <claudio.fontana@linaro.org>, Dirk Mueller <dmueller@suse.de>, Will Newton <will.newton@linaro.org>, Laurent Desnogues <laurent.desnogues@gmail.com>, =?ISO-8859-1?Q?Alex_Benn=E9e?= <alex.bennee@linaro.org>, kvmarm@lists.cs.columbia.edu, Christoffer Dall <christoffer.dall@linaro.org>

On 12/07/2013 02:19 AM, Peter Maydell wrote:
> +    tcg_rd = cpu_reg(s, rd);
> +
> +    if (cond >= 0x0e) { /* condition "always" */
> +        tcg_src = read_cpu_reg(s, rn, sf);
> +        tcg_gen_mov_i64(tcg_rd, tcg_src);
> +    } else {
> +        /* OPTME: we could use movcond here, at the cost of duplicating
> +         * a lot of the arm_gen_test_cc() logic.
> +         */
> +        int label_match = gen_new_label();
> +        int label_continue = gen_new_label();
> +
> +        arm_gen_test_cc(cond, label_match);
> +        /* nomatch: */
> +        tcg_src = cpu_reg(s, rm);

Sorry for missing this in the first round: For the silly corner case of Rd ==
XZR, tcg_rd is dead after the branch.

We could either move the tcg_rd assignment down into each basic block with the
assignment to tcg_src, or simply add

    if (rd == 31) {
        /* silly no-op write; until we use movcond we must special-case
           this to avoid a dead temporary across basic blocks.  */
        return;
    }

Either solution is ok by me.


r~