Re: [libvirt] Analysis of the effect of adding PCIe root ports

So here's a rewording of your description (with a couple additional  conditions) to see if I understand everything correctly:   1) during initial domain definition:   A) If there are *no pci controllers at all* (not even a pci-root or  pcie-root) *and there are any unaddressed devices that need a PCI slot*  then auto-add enough controllers for the requested devices, *and* make  sure there are enough empty slots for "N" (do we stick with 4? or make  it 3?) devices to be added later without needing more controllers. (So,  if the domain has no PCI devices, we don't add anything extra, and also  if it only has PCI devices that already have addresses, then we also  don't add anything extra).   B) if there is at least one pci controller specified in the XML, and  there are any unused slots in the pci controllers in the provided XML,  then use them for the unaddressed devices. If there are more devices  that need an address at this time, also add controllers for them, but no  *extra* controllers.   (Note to Rich: libguestfs could avoid the extra controllers either by  adding a pci-root/pcie-root to the XML, or by manually addressing the  devices. The latter would actually be better, since it would avoid the  need for any pcie-root-ports).   2) When adding a device to the persistent config (i.e. offline): if  there is an empty slot on a controller, use it. If not, add a controller  for that device *but no extra controllers*   3) when adding a device to the guest machine (i.e. hotplug / online), if  there is an empty slot on a controller, use it. If not, then fail.   The differences I see from what (I think) you suggested are:   * if there aren't  any unaddressed pci devices (even if there are no  controllers in the config), then we also don't add any extra controllers  (although we will of course add the pci-root or pcie-root, to  acknowledge it is there).   * if another controller is needed for adding a device offline, it's okay  to add it.