From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1KQiU4-0008Fk-Gn for qemu-devel@nongnu.org; Wed, 06 Aug 2008 08:51:20 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1KQiU2-0008Bs-QM for qemu-devel@nongnu.org; Wed, 06 Aug 2008 08:51:19 -0400 Received: from [199.232.76.173] (port=47040 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1KQiU2-0008Bf-Gc for qemu-devel@nongnu.org; Wed, 06 Aug 2008 08:51:18 -0400 Received: from smtp.eu.citrix.com ([62.200.22.115]:32515) by monty-python.gnu.org with esmtp (Exim 4.60) (envelope-from ) id 1KQiU1-00043y-6c for qemu-devel@nongnu.org; Wed, 06 Aug 2008 08:51:18 -0400 Date: Wed, 6 Aug 2008 13:50:55 +0100 From: Samuel Thibault Subject: Re: [Xen-devel] Re: [Qemu-devel] [PATCH 0/7] merge some xen bits into qemu Message-ID: <20080806125055.GG4448@implementation.uk.xensource.com> References: <48973F8E.8080109@codemonkey.ws> <20080805104630.GM4478@implementation.uk.xensource.com> <489835A3.6030804@redhat.com> <20080805112935.GO4478@implementation.uk.xensource.com> <48985336.2020709@redhat.com> <20080805150328.GT4478@implementation.uk.xensource.com> <20080805154140.GV4478@implementation.uk.xensource.com> <48997981.1030703@redhat.com> <20080806102338.GA4448@implementation.uk.xensource.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: Reply-To: qemu-devel@nongnu.org List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Markus Armbruster Cc: xen-devel@lists.xensource.com, Ian Jackson , Gerd Hoffmann , qemu-devel@nongnu.org Markus Armbruster, le Wed 06 Aug 2008 08:43:49 -0400, a écrit : > It *is* quite move, becauses it accomplishes a lot: it goes from a > heavily modified fork of an oldish version all the way to merge with > upstream, as far as PV is concerned. Then why doing it in qemu before having it tested in the xen unstable tree? That's not the way I usually see merging happen. > For what it's worth, I went over significant parts of Gerd's patch > (all the generic stuff + pvfb) with a fine comb, comparing it to what > we have now. I consider it sound. I'm not saying it's not fine. I had a look and the code looked fine indeed. But what I'm afraid of is the delta with Ian would have to bear when merging: is it save/restore safe, does it work with PCI pass-through, VT-D, etc.? > If that's where we want to go, we can of course still argue whether we > should go in leaps or baby steps, and whether Gerd's leap lands in > quite the right spot. Baby steps are much easier to review. That's how things are usually done, and here it looks to me like it is feasible to achieve in Xen (and have it tested) before event thinking about importing a pile of code in qemu where it won't receive as much testing as the xen-unstable tree receives. Samuel