From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([140.186.70.92]:52278) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1RWYE2-0001OB-FE for qemu-devel@nongnu.org; Fri, 02 Dec 2011 13:52:47 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1RWYE1-0007ZH-8x for qemu-devel@nongnu.org; Fri, 02 Dec 2011 13:52:46 -0500 Received: from am1ehsobe001.messaging.microsoft.com ([213.199.154.204]:23244 helo=AM1EHSOBE001.bigfish.com) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1RWYE1-0007ZC-0M for qemu-devel@nongnu.org; Fri, 02 Dec 2011 13:52:45 -0500 Message-ID: <4ED91E72.9040303@freescale.com> Date: Fri, 2 Dec 2011 12:52:34 -0600 From: Scott Wood MIME-Version: 1.0 References: <20111103195452.21259.93021.stgit@bling.home> <4ED43AD9.5090509@au1.ibm.com> <4ED43CFE.8040009@au1.ibm.com> <1322538856.19120.126.camel@bling.home> <1322594768.19120.194.camel@bling.home> <1322610273.19120.209.camel@bling.home> <1322672293.19120.232.camel@bling.home> <1322774717.26545.118.camel@bling.home> <4ED91876.2030408@freescale.com> In-Reply-To: Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [RFC PATCH] vfio: VFIO Driver core framework List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Bhushan Bharat-R65777 Cc: "chrisw@sous-sol.org" , Wood Scott-B07421 , Alexey Kardashevskiy , "kvm@vger.kernel.org" , Stuart Yoder , "joerg.roedel@amd.com" , "agraf@suse.de" , "pmac@au1.ibm.com" , "qemu-devel@nongnu.org" , "iommu@lists.linux-foundation.org" , "konrad.wilk@oracle.com" , "aafabbri@cisco.com" , Yoder Stuart-B08248 , Alex Williamson , "avi@redhat.com" , "linux-pci@vger.kernel.org" , "dwg@au1.ibm.com" , "benve@cisco.com" On 12/02/2011 12:45 PM, Bhushan Bharat-R65777 wrote: > Scott, I am not sure if there is any real use case where device needed to assigned beyond 2 level (host + immediate guest) in nested virtualization. Userspace drivers in the guest is a more likely scenario than nested virtualization, at least for us. Our hardware doesn't support nested virtualization, so it would have to be some slow emulation-based approach (worse than e500v2, since we don't have multiple PID registers). > But if there any exists then will not it be better to virtualizes the iommu (PAMU for Freescale)? We can't virtualize the PAMU in any sort of transparent manner. It's not flexible enough to handle arbitrary mappings. The guest will need to cooperate with the host to figure out what mappings it can do. -Scott