From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:46209) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Vqm8v-0003Gc-8V for qemu-devel@nongnu.org; Wed, 11 Dec 2013 10:56:10 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Vqm8u-0001RW-3F for qemu-devel@nongnu.org; Wed, 11 Dec 2013 10:56:09 -0500 Received: from mx1.redhat.com ([209.132.183.28]:21303) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Vqm8t-0001RJ-RE for qemu-devel@nongnu.org; Wed, 11 Dec 2013 10:56:08 -0500 Received: from int-mx02.intmail.prod.int.phx2.redhat.com (int-mx02.intmail.prod.int.phx2.redhat.com [10.5.11.12]) by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id rBBFu6H0019041 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for ; Wed, 11 Dec 2013 10:56:06 -0500 Message-ID: <52A88B12.8060303@redhat.com> Date: Wed, 11 Dec 2013 16:56:02 +0100 From: Paolo Bonzini MIME-Version: 1.0 References: <1385401393-14291-1-git-send-email-pbonzini@redhat.com> <20131128102652.GA24261@redhat.com> <52A714AC.3050703@redhat.com> <20131210150542.GA27998@amt.cnet> <20131210172144.GA29151@amt.cnet> <20131210210241.GC8476@redhat.com> <20131211134118.GB14071@amt.cnet> <20131211142000.GA23355@redhat.com> <52A87A89.4000901@redhat.com> <20131211164529.0b4d92a6@thinkpad> In-Reply-To: <20131211164529.0b4d92a6@thinkpad> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH rebased for-1.8] i386: pc: align gpa<->hpa on 1GB boundary (v6) List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Igor Mammedov Cc: Marcelo Tosatti , qemu-devel@nongnu.org, "Michael S. Tsirkin" Il 11/12/2013 16:45, Igor Mammedov ha scritto: >> > I'm not sure if it is fixable. You need a 2M mountpoint to bind the 3G-4G >> > range correctly, a 1G mountpoint for everything else, and QEMU only allows >> > to specify one path. > we could do it with hugepage memdev backend. > something like: > > -object hugepage-ram,id=mem1gb,size=3G,host-node=0,mem-path=/1gb-hugepage-fs > -device dimm,id=hp1g,memdev=mem1gb,node=0 > -object hugepage-ram,id=mem2mb,size=500Mb,host-node=1,mem-path=/2mb-hugepage-fs > -device dimm,id=hp2mb,memdev=mem2mb,node=1 > > that basically would allow to distribute initial memory in any way user would > like. If you allow for DIMMs, you can just use a small initial amount of memory (2GB), and cold-plug DIMMs at 4GB. Then you get exactly the same result as Gerd's patch. :) But the beauty of Marcelo's idea was that the user didn't need to do anything, and the guest did not see anything. It's a great approach for backwards-compatibility, no doubt about that. Paolo