Re: [libvirt] [PATCH v6 15/17] qemu: initially reserve one open pcie-root-port for hotplug

Instead this patch just reserves one slot for a "future potential" PCIe device after doing the assignment for actual devices, but only if the only PCI controller defined prior to starting address assignment was pcie-root, and only if we auto-added at least one PCI controller during address assignment.  This assures two things:

This is set to reserve a single free port for now, but could be increased in the future if public sentiment goes in that direction (it's easy to increase later, but essential impossible to decrease)

diff --git a/src/qemu/qemu_domain_address.c b/src/qemu/qemu_domain_address.c index fbbcfb2..15d7c1a 100644 --- a/src/qemu/qemu_domain_address.c +++ b/src/qemu/qemu_domain_address.c @@ -1929,6 +1929,36 @@ qemuDomainAssignPCIAddresses(virDomainDefPtr def,           if (qemuDomainAssignDevicePCISlots(def, qemuCaps, addrs) < 0)               goto cleanup;    +        /* Only for *new* domains with pcie-root (and no other +         * manually specified PCI controllers in the definition): If, +         * after assigning addresses/reserving slots for all devices, +         * we see that any extra buses have been auto-added, we +         * understand that the application has left management of PCI +         * addresses and controllers up to libvirt. In order to allow +         * such applications to easily support hotplug, we will do a +         * "one time" reservation of one extra PCIE|HOTPLUGGABLE +         * slots, which should cause su to auto-add 1 extra

+         * pcie-root-ports The single slot in this root-port will be