On Mon, Jan 13, 2014 at 11:13:07AM +0100, Gerd Hoffmann wrote:
> On Mi, 2014-01-08 at 17:11 +0100, Paolo Bonzini wrote:
> > Il 08/01/2014 16:07, Yan Vugenfirer ha scritto:
> > >>>
> > >>> +%QEMU-PCI_SERIAL_1_PORT%=ComPort_inst1,
> > >>> PCI\VEN_1B36&DEV_0002&SUBSYS_11001AF4&REV_01
> > >>> +%QEMU-PCI_SERIAL_2_PORT%=ComPort_inst2,
> > >>> PCI\VEN_1B36&DEV_0003&SUBSYS_11001AF4&REV_01
> > >>> +%QEMU-PCI_SERIAL_4_PORT%=ComPort_inst4,
> > >>> PCI\VEN_1B36&DEV_0004&SUBSYS_11001AF4&REV_01
> > >>
> > >> I think checking the subsystem is not necessary (and I think downstreams
> > >> could legitimately change it). Can you check CC and REV but not SUBSYS?
> > >
> > > PNP ID can be reduced to vendor and device ID only, for example:
> > > PCI\VEN_1B36&DEV_0002 . But in this case we cannot check revision.
> >
> > Gerd, Michael, what do you think is better? Not check revision, or
> > enforcing subsystem?
>
> No need to check the revision. There is only one, I doubt this will
> ever change, and should we do a rev2 virtual hardware it is supposed to
> be backward-compatible to rev1 (otherwise we should hand out a new pci
> id to the device).
>
> [ drivers which depend on few features of the hypothetical rev2 hardware
> and don't work with rev1 would need a revision check ]
>
> cheers,
> Gerd
>
Well linux checks revision:
if (pci_dev->revision != VIRTIO_PCI_ABI_VERSION) {
printk(KERN_ERR "virtio_pci: expected ABI version %d,
got %d\n",
VIRTIO_PCI_ABI_VERSION, pci_dev->revision);
return -ENODEV;
}
so it seems better to be consistent.
in any case, we must check subsystem ids, the spec is very
explicit on this point.
--
MST
