This patch has to be squashed with the previous one otherwise itI do not think so. The previous patch passes the parent to the pci_create_root_bus(), indeed, so pci_create_root_bus() set companion device before calling pcibios_root_bridge_prepare() (for both cases x86 and ia64). In turn, pcibios_root_bridge_prepare() will do nothing, since bridge->dev.parent != NULL. After all, pcibios_root_bridge_prepare() is dead code.
has bisectability issues (ie the previous patch passes the parent
to the pci_create_root_bus() call so, without this patch applied,
the companion initialization would fail).
Thanks!
Other than that:
Reviewed-by: Lorenzo Pieralisi<lorenzo.pieralisi@xxxxxxx>