From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([209.51.188.92]:32996) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1guapw-0000uF-M1 for qemu-devel@nongnu.org; Fri, 15 Feb 2019 05:35:17 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1guapu-0003V2-G9 for qemu-devel@nongnu.org; Fri, 15 Feb 2019 05:35:16 -0500 Received: from mx1.redhat.com ([209.132.183.28]:53590) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1guapt-0003Kx-EU for qemu-devel@nongnu.org; Fri, 15 Feb 2019 05:35:14 -0500 Date: Fri, 15 Feb 2019 10:34:55 +0000 From: Daniel =?utf-8?B?UC4gQmVycmFuZ8Op?= Message-ID: <20190215103455.GB14454@redhat.com> Reply-To: Daniel =?utf-8?B?UC4gQmVycmFuZ8Op?= References: <20190118133826.22447-1-berrange@redhat.com> <20190128174857.GT4136@habkost.net> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <20190128174857.GT4136@habkost.net> Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH] hw: report invalid disable-legacy|modern usage for virtio-1-only devs List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Eduardo Habkost Cc: qemu-devel@nongnu.org, Gonglei , "Michael S. Tsirkin" , Gerd Hoffmann On Mon, Jan 28, 2019 at 03:48:57PM -0200, Eduardo Habkost wrote: > On Fri, Jan 18, 2019 at 01:38:26PM +0000, Daniel P. Berrang=C3=A9 wrote= : > > A number of virtio devices (gpu, crypto, mouse, keyboard, tablet) onl= y > > support the virtio-1 (aka modern) mode. Currently if the user launche= s > > QEMU, setting those devices to enable legacy mode, QEMU will silently > > create them in modern mode, ignoring the user's (mistaken) request. > >=20 > > This patch introduces proper data validation so that an attempt to > > configure a virtio-1-only devices in legacy mode gets reported as an > > error to the user. > >=20 > > Checking this required introduction of a new field to explicitly trac= k > > what operating model is to be used for a device, separately from the > > disable_modern and disable_legacy fields that record the user's > > requested configuration. > >=20 > > Signed-off-by: Daniel P. Berrang=C3=A9 >=20 > This breaks the virtio-1-only devices on pc-*-2.6: >=20 > $ ./x86_64-softmmu/qemu-system-x86_64 -device virtio-vga -machine pc-= i440fx-2.6 > qemu-system-x86_64: -device virtio-vga: Unable to set disable-legacy=3D= off on a virtio-1.0 only device >=20 > Probably the simplest way to fix that is to append the following to > hw_compat_2_6: >=20 > { virtio-vga, disable-modern, off }, > { virtio-gpu-pci, disable-modern, off }, > { virtio-input-pci, disable-modern, off }, > { virtio-crypto-pci, disable-modern, off } That's no sufficient, as it needs to also set disable-legacy=3Don Once we add all these extra properties, it becomes pointless to try to use the generic settings against virtio-pci in hw_compat_2_6. So I've just turn it into a whitelist, only listing the devices which actually support both modern+legacy modes. This makes it obvious which devices are being affected & avoids silent surprises like this bug you found in my patch. Regards, Daniel --=20 |: https://berrange.com -o- https://www.flickr.com/photos/dberran= ge :| |: https://libvirt.org -o- https://fstop138.berrange.c= om :| |: https://entangle-photo.org -o- https://www.instagram.com/dberran= ge :|