From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from [140.186.70.92] (port=40443 helo=eggs.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1Q7D8I-0006H1-I8 for qemu-devel@nongnu.org; Tue, 05 Apr 2011 16:45:51 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Q7D8H-0000B8-8M for qemu-devel@nongnu.org; Tue, 05 Apr 2011 16:45:50 -0400 Received: from mail-vx0-f173.google.com ([209.85.220.173]:53311) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Q7D8H-0000B4-6E for qemu-devel@nongnu.org; Tue, 05 Apr 2011 16:45:49 -0400 Received: by vxb41 with SMTP id 41so709323vxb.4 for ; Tue, 05 Apr 2011 13:45:48 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <4D9B7BA6.3020800@codemonkey.ws> References: <1302029716.31961.24.camel@nibbler.dlib.indiana.edu> <4D9B7BA6.3020800@codemonkey.ws> Date: Tue, 5 Apr 2011 21:45:48 +0100 Message-ID: Subject: Re: [Qemu-devel] Questions about chardev (and ps/2 mouse) From: Peter Maydell Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Anthony Liguori Cc: qemu-devel@nongnu.org On 5 April 2011 21:29, Anthony Liguori wrote: > On 04/05/2011 01:55 PM, Brian Wheeler wrote: >> >> I'm trying to disable the ps/2 mouse in favor of the serial mouse >> provided by -chardev msmouse since the ps/2 mouse has had lots of >> problems with older OSes (OpenStep, OS/2, Win 3.1, etc). =C2=A0I've got = a >> couple of questions: >> >> 1) If I change a VMState something by appending a field (i.e. >> "unplugged"), do I have to update version_id? > > Yes. =C2=A0You can use a subsection though to avoid bumping the version_i= d. > >> =C2=A0 How do the >> minimum_version_id{,_old} fields fit in? > > This is for load, not save. =C2=A0It's used to basically deprecate loadin= g from > very old versions. There was an email thread about this recently: http://web.archiveorange.com/archive/v/1XS1vynR1fW0cG5xll6r I need to go back and reread it and propose a docs patch for people to shoot down; I think having a clear description of "what do I have to do if I've added a new field to a device model" is important, because that will often be the only interaction a developer has with the migration/vmstate stuff. -- PMM