On 06/15/2009 04:23 PM, Anthony Liguori wrote:
Avi Kivity wrote:
On 06/15/2009 03:52 PM, Anthony Liguori wrote:
Avi Kivity wrote:
On 06/15/2009 03:41 PM, Michael S. Tsirkin wrote:
We should just tell the user which slots are open.
This might be tricky if the config is passed in with the command
line
flags.
qemu -show-available-pci-slots
Why does the user care?
Let QEMU allocate the PCI slot, then query it to see what slot it
assigned and remember that.
It's a roundabout way of doing things.
Having libvirt do PCI slot allocation scares me. It assumes we can
return a whitelist of available slots, and then let libvirt just
randomly assign things. There's knowledge though in slot assignment
that's board-specific. For instance, depending on how many LNK lines
you have, you may want to put things in slots in such a way to
optimize interrupt balancing or something like that.
How would qemu know which slots to optimize for?
In practice, I don't see that as a real problem. We should (a) add an
ioapic and four more pci links (b) recommend that slots be assigned in
ascending order, and everything works.
I don't see your concern about libvirt allocating slots. If a human can
plug a card into a slot, so can libvirt. Doing an interactive
back-and-forth (equivalent to plugging a card while blindfolded, then
looking to see which slot we hit) is certainly more difficult.
Some platforms have quirks about expecting a particular slot to have a
particular device. It's still an optimal device but it has to be in
that slot. You can't really express that via an available slot list.
I'll be surprised if we ever measure different dma speeds on different
slots in the qemu virtual pci bus. If we do, we'll find a way to
express them:
$ qemu -print-pci
slot 0:01: available 33MHz
slot 0:02: available 33MHz
slot 0:03: available 66MHz
I feel a little silly typing this.
--
error compiling committee.c: too many arguments to function
--
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
