From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([140.186.70.92]:58309)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <weil@mail.berlios.de>) id 1QpnFf-00041F-2E
	for qemu-devel@nongnu.org; Sat, 06 Aug 2011 16:13:43 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <weil@mail.berlios.de>) id 1QpnFd-0007iS-T9
	for qemu-devel@nongnu.org; Sat, 06 Aug 2011 16:13:43 -0400
Received: from v220110690675601.yourvserver.net ([78.47.199.172]:51104)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <weil@mail.berlios.de>) id 1QpnFd-0007iM-Ky
	for qemu-devel@nongnu.org; Sat, 06 Aug 2011 16:13:41 -0400
Message-ID: <4E3DA06C.7050506@mail.berlios.de>
Date: Sat, 06 Aug 2011 22:13:32 +0200
From: Stefan Weil <weil@mail.berlios.de>
MIME-Version: 1.0
References: <CAAu8pHvSJvL-ypbieMj=h3=xKpmYz3iuzi9BpwsZcrNiS2Svug@mail.gmail.com>
In-Reply-To: <CAAu8pHvSJvL-ypbieMj=h3=xKpmYz3iuzi9BpwsZcrNiS2Svug@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [PATCH] TCG: fix copy propagation
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: Blue Swirl <blauwirbel@gmail.com>
Cc: TeLeMan <geleman@gmail.com>, qemu-devel <qemu-devel@nongnu.org>, zhur@ispras.ru, Kirill Batuzov <batuzovk@ispras.ru>

Am 06.08.2011 16:06, schrieb Blue Swirl:
> Copy propagation introduced in 22613af4a6d9602001e6d0e7b6d98aa40aa018dc
> considered only global registers. However, register temps and stack
> allocated locals must be handled differently because register temps
> don't survive across brcond.
>
> Fix by propagating only within same class of temps.
>
> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
> ---
> tcg/optimize.c | 13 +++++++------
> tcg/tcg.h | 5 +++++
> 2 files changed, 12 insertions(+), 6 deletions(-)
>
> diff --git a/tcg/optimize.c b/tcg/optimize.c
> index a3bfa5e..748ecf9 100644
> --- a/tcg/optimize.c
> +++ b/tcg/optimize.c
> @@ -185,12 +185,13 @@ static int op_to_movi(int op)
> }
> }
>
> -static void tcg_opt_gen_mov(TCGArg *gen_args, TCGArg dst, TCGArg src,
> - int nb_temps, int nb_globals)
> +static void tcg_opt_gen_mov(TCGContext *s, TCGArg *gen_args, TCGArg dst,
> + TCGArg src, int nb_temps, int nb_globals)
> {
> reset_temp(dst, nb_temps, nb_globals);
> assert(temps[src].state != TCG_TEMP_COPY);
> - if (src >= nb_globals) {
> + if (src >= nb_globals &&
> + tcg_arg_is_local(s, src) == tcg_arg_is_local(s, dst)) {
> assert(temps[src].state != TCG_TEMP_CONST);
[snip]

Hi Blue,

your patch fixes qemu-system-x86_64 which now seems to work on 32 bit 
hosts, too.

qemu-system-mips64(el) still fail with the same abort. They work when I 
remove the
if block in tcg_opt_gen_mov.

The Debian kernel for qemu-system-mips64 which I used for the test is 
available on
http://qemu.weilnetz.de/mips64/.

I could not reproduce the crash with qemu-system-ppc64 - neither with 
nor without
your patch.

Kind regards,
Stefan