From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1LXGum-0004hJ-O0 for qemu-devel@nongnu.org; Wed, 11 Feb 2009 10:22:16 -0500 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1LXGul-0004h0-7H for qemu-devel@nongnu.org; Wed, 11 Feb 2009 10:22:15 -0500 Received: from [199.232.76.173] (port=53435 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1LXGul-0004gw-14 for qemu-devel@nongnu.org; Wed, 11 Feb 2009 10:22:15 -0500 Received: from qw-out-1920.google.com ([74.125.92.145]:64345) by monty-python.gnu.org with esmtp (Exim 4.60) (envelope-from ) id 1LXGuk-0002Kh-PZ for qemu-devel@nongnu.org; Wed, 11 Feb 2009 10:22:14 -0500 Received: by qw-out-1920.google.com with SMTP id 4so44155qwk.4 for ; Wed, 11 Feb 2009 07:22:12 -0800 (PST) Message-ID: <4992ED09.9030400@codemonkey.ws> Date: Wed, 11 Feb 2009 09:21:45 -0600 From: Anthony Liguori MIME-Version: 1.0 Subject: Re: [Qemu-devel] Support for out-of-code cpu definitions References: <1234358114-23390-1-git-send-email-amit.shah@redhat.com> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Reply-To: qemu-devel@nongnu.org List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Blue Swirl wrote: > On 2/11/09, Amit Shah wrote: > >> This patch series adds support for out-of-code cpu definitions. This helps >> us move the definitions out of target-i386/helper.c. The result is we can >> store the definitions in plain-text; add, modify the definitions without >> having to recompile and most importantly, some management servers can >> define their own CPU types based on the network topology to create >> migration-safe cpu types. >> >> The first patch in the series was sent earlier, just converts the usage of >> bios_dir to get_datafile_dir(), as the path that's referred to stores more >> than just the bios (keymaps, and now, the cpudefs). >> >> The cpudef file currently is very similar to the code that was in helper.c. >> >> Comments welcome on the format. >> > > Instead of this change, IMHO we should wait for either the > configuration file format to be fixed, or use FDT for all HW > definitions. This format is not generic enough, definitely not > suitable for Sparc CPU features. But if the config or FDT does not > appear, we could use this as a starting point. > I agree. If you switch cpudef to FDT, that could provide an intermediate step between what we have today and a full machine description. Regards, Anthony Liguori