From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1Lzcuv-000540-OZ for qemu-devel@nongnu.org; Thu, 30 Apr 2009 16:31:37 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1Lzcuq-00053o-CX for qemu-devel@nongnu.org; Thu, 30 Apr 2009 16:31:36 -0400 Received: from [199.232.76.173] (port=55231 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1Lzcuq-00053l-5w for qemu-devel@nongnu.org; Thu, 30 Apr 2009 16:31:32 -0400 Received: from mx20.gnu.org ([199.232.41.8]:19244) by monty-python.gnu.org with esmtps (TLS-1.0:RSA_AES_256_CBC_SHA1:32) (Exim 4.60) (envelope-from ) id 1Lzcup-0004NY-QM for qemu-devel@nongnu.org; Thu, 30 Apr 2009 16:31:31 -0400 Received: from verein.lst.de ([213.95.11.210]) by mx20.gnu.org with esmtp (Exim 4.60) (envelope-from ) id 1Lzcup-0006IM-3h for qemu-devel@nongnu.org; Thu, 30 Apr 2009 16:31:31 -0400 Date: Thu, 30 Apr 2009 22:31:22 +0200 From: Christoph Hellwig Subject: Re: [Qemu-devel] [PATCH] Change the virtio-block storage class to PCI_CLASS_STORAGE_SCSI Message-ID: <20090430203122.GA31440@lst.de> References: <49F9F86C.3000009@redhat.com> <20090430200901.GA30530@lst.de> <49FA0788.5000809@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <49FA0788.5000809@redhat.com> List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Dor Laor Cc: Christoph Hellwig , qemu-devel On Thu, Apr 30, 2009 at 11:18:16PM +0300, Dor Laor wrote: > Christoph Hellwig wrote: > >On Thu, Apr 30, 2009 at 10:13:48PM +0300, Dor Laor wrote: > > > >>Windows virtio driver cannot pass DTM (certification) tests > >>while the storage class is PCI_CLASS_STORAGE_UNKNOWN. > >> > > > >Well, it isn't scsi so this is wrong. Tell these certification people > >to fix their test. > > > It's also not unknown... someone I know implemented scsi pass through so > it is > more close to scsi than to unknown.. > These certification people are M$ so it is simpler to do this harmless > change instead > of waiting some light years. It only supports limited scsi cdb passthrough _if_ it's using a raw host device supporting the SG_IO ioctl on Linux. That's a lot of ifs and whens and definitively not anything like a SCSI HBA. Now fortunately the pci class IDs don't matter too much in real life, but I'd rather not change it from the correct to the wrong one. Especially as there might be all kinds of weird breakages in existing setups when the pci id changes underneath on a migration / upgrade of the host. So if you really need this to please a broken testcase add an option to qemu to use a different pci id just for this broken certification test, similarly to various other OS workarounds.