kernel does not support hotplug removal of virtio devices

[Scott Moser]

hotplug removal of virtio devices (in kvm guests) will not work without commit at http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=29f9f12ec7 . Fedora bug: https://bugzilla.redhat.com/show_bug.cgi?id=490479 .

To reproduce:
- create a disk image to attach: qemu-img create -f qcow2 /tmp/my-disk.img 1G
- boot jaunty guest, with -server or -generic kernel so as to have access to acpiphp.ko module (bug 364916). I've verified on 2.6.28.11.15
- inside guest: 'modprobe acpiphp.ko' and 'modprobe virtio_blk'
- from kvm monitor: pci_add 0 storage file=/tmp/my-disk.img,if=virtio
  watch to see what slot this is put on
- inside guest: dmesg and cat /proc/partitions should show a new 'vd*' device
- from kvm monitor: pci_del 0 X
  where X is the slot from the add
- inside guest: see message on console or dmesg:
  "modprobe virtio_blk ... Device 'virtio2' does not have a release() function, it is broken and must be fixed"

[Scott Moser (smoser)]

I've verified this is functioning with kvm on lucid beta 1 (2.6.32-16-generic / x86_64) host, with lucid beta 1 guest. Thus, marking this fix released. Only difference from the text in the bug open comment is that 'pci_add auto storage file=/tmp/my-disk.img,if=virtio' and then, to detach: pci_del 0:X (':', not space)