From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:35507) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VIEa6-0006kX-Ri for qemu-devel@nongnu.org; Sat, 07 Sep 2013 05:13:32 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1VIEa0-0004sv-Lw for qemu-devel@nongnu.org; Sat, 07 Sep 2013 05:13:26 -0400 Received: from moutng.kundenserver.de ([212.227.17.10]:61291) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VIEa0-0004qp-C7 for qemu-devel@nongnu.org; Sat, 07 Sep 2013 05:13:20 -0400 Message-ID: <522AEE2D.1090408@Vivier.EU> Date: Sat, 07 Sep 2013 11:13:17 +0200 From: Laurent Vivier MIME-Version: 1.0 References: <1378500474-26067-1-git-send-email-laurent@vivier.eu> In-Reply-To: Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit Subject: Re: [Qemu-devel] [PATCH 0/4] Define default CPU at configure time List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Peter Maydell Cc: Riku Voipio , QEMU Developers Le 07/09/2013 10:43, Peter Maydell a écrit : > On 6 September 2013 21:47, Laurent Vivier wrote: >> This series of patch allows to set the default CPU used by linux-user qemu. >> >> When qemu is used in a container as a binfmt interpreter we can't use the >> environment variable QEMU_CPU to set the CPU id. > I definitely don't like this. We shouldn't be hardcoding this at > configure time. There doesn't seem to be any particular reason > why QEMU_CPU is the only option you might want to set for > your binfmt-interpreter either. In fact, I don't understand why a given CPU id has been chosen to be the default value. I think there is a default value not because this is the best/most used value but because we need a default. It's hardcoded at development time. So I think this is better to be able to choose this default value without editing any file. A configuration option is the best solution for me. Moreover, this doesn't change the default behavior. > Is there a reason why using a wrapper script which sets > options to the real qemu binary won't work? Personal reason : I don't like wrapper. Kernel must load two executables, wrapper must analyze environment, auxv, parameters and pass what it needs to pass. But I don't want to push these patches at any reason : I just want to share. Thank you for your comments. Regards, Laurent