On Friday 16 January 2015 10:23:11 Arnd Bergmann wrote:

On Friday 16 January 2015 09:44:09 Yijing Wang wrote:

quoted

@@ -2064,7 +2073,7 @@ struct pci_bus *pci_scan_root_bus(struct device *parent, u32 db,
 {
        struct pci_host_bridge *host;
 
-       host = pci_create_host_bridge(parent, db, resources, sysdata);
+       host = pci_create_host_bridge(parent, db, resources, sysdata, NULL);
        if (!host)
                return NULL;
 

Same comment as for patch 10: If we leave this out of the pci_create_host_bridge
argument but set it later, ?e interface becomes simpler for callers that
don't have custom pci_host_bridge_ops.

This would break your phb_prepare() callback, but I assume a host driver
can just make a direct function call before entering pci_create_host_bridge
for this. Let me know if I'm missing something here.

I've read the later patches now that explain why it's needed, so nevermind
my comment above.

	Arnd