Avi Kivity <avi@xxxxxxxxxx> writes: > On 06/15/2009 12:08 PM, Mark McLoughlin wrote: >>> This last option makes sense to me: in a real world the user has >>> control over where he places the device on the bus, so why >>> not with qemu? >>> >> >> Yep, most people seem to agree that it makes sense to allow this, but >> some believe it should only be via a machine description file, not the >> command line. >> > > I don't understand this opposition. It's clear a machine config file > is a long way in our future. It's also clear lack of stable PCI > addresses hurts us now. Correct. >> However, the first problem is that it isn't a solution to the guest ABI >> problem more generally. >> > > pci_addr was never meant to bring world peace, just stable PCI > addresses. The other issues should be addressed separately. > >> And the second problem is that for e.g. libvirt to use it, it would have >> to be possible to query qemu for what PCI slots were assigned to the >> devices - libvirt would need to be able to parse 'info pci' and match >> the devices listed with the devices specified on the command line. >> > > If all devices (including vga, ide) are set up with pci_addr, then > this is unneeded. You do need to export available slot numbers from > qemu. Not really. QEMU gives just the host bridge a fixed slot[*]. All the other slots are available. The real problem is devices that get implicitly added, like the SCSI controller. Those devices get their slots auto-assigned, which can interfere with slot numbers chosen by the user. We need a way to avoid that, as you suggested elsewhere in this thread. [*] There's an exception or two for oddball targets. -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
