Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1Kl1eb-0002DV-1N for qemu-devel@nongnu.org; Wed, 01 Oct 2008 09:22:09 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1Kl1eZ-0002Ca-8I for qemu-devel@nongnu.org; Wed, 01 Oct 2008 09:22:08 -0400 Received: from [199.232.76.173] (port=39392 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1Kl1eZ-0002CU-13 for qemu-devel@nongnu.org; Wed, 01 Oct 2008 09:22:07 -0400 Received: from jaguar.mkp.net ([192.139.46.146]:57757) by monty-python.gnu.org with esmtp (Exim 4.60) (envelope-from ) id 1Kl1eY-0004lx-Hq for qemu-devel@nongnu.org; Wed, 01 Oct 2008 09:22:06 -0400 Sender: jes@jaguar.mkp.net Subject: Re: [Qemu-devel] [patch] Introduce per machine based max_cpu variable References: <48D8D694.1000609@sgi.com> <200809231350.34178.paul@codesourcery.com> <48D8E6CA.9040802@sgi.com> <200809231404.49340.paul@codesourcery.com> <48DBAC61.9090906@sgi.com> <48E23C00.2030309@codemonkey.ws> <48E24359.9070607@codemonkey.ws> From: Jes Sorensen Date: 01 Oct 2008 09:22:03 -0400 In-Reply-To: <48E24359.9070607@codemonkey.ws> Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii 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 Cc: jes@sgi.com, Paul Brook >>>>> "Anthony" == Anthony Liguori writes: Anthony> If it were 'max_cpu' then the meaning is pretty clear. On a Anthony> UP system, the max_cpu is 0. Well it depends on how you calculate it, it is meant as max_nr_cpus which IMHO is the most clear. If you prefer, I can rename it to that. Most users will think in terms of 'number of cpus to enable' rather than the highest logical cpu number. Anthony> You could also make it max_aps which makes it even more Anthony> clear. A UP system doesn't have any APs. What do you mean by AP? Anthony> The other option is to have max_cpus == 0 be equivalent to Anthony> max_cpus == 1. But don't print anything. Basically, I don't Anthony> like the idea of modifying every machine definition when they Anthony> all have almost the exact same value for the field. I really don't understand this - the code as it is today is really inconsistent. What my patch does is simply to make it explicit what it does and make it consistent. I have taken great care to do it in a way, so if in the unlikely situation it breaks an existing behavior, it will be very clear what is wrong and how to fix it. I'll even be willing to put my name in there and volunteer to fix it up for people if this actually happens :-) Cheers, Jes