On Tue, Sep 25, 2012 at 10:37 AM, Bjorn Helgaas <bhelgaas@xxxxxxxxxx> wrote:
> On Tue, Sep 25, 2012 at 10:29 AM, Yinghai Lu <yinghai@xxxxxxxxxx> wrote:
>>>
>>> I'd prefer a design where the PCI core provides an interface that
>>> means "call this function for every host bridge we know about now
>>> *and* for every one that's added in the future."
>>
>> yes, that is the point to add pci_root_bridge_bus_type. We can register
>> bus notifier on that.
>
> I guess I missed your point. In the patch below (20/29 from your
> series), you're still iterating through all the host bridges, so there
> would have to be something else to handle hot-added host bridges.
>
> Are you saying you plan future patches to change this again to
> something using a bus notifier?
>
> --- a/arch/sparc/kernel/pci.c
> +++ b/arch/sparc/kernel/pci.c
> @@ -997,11 +997,13 @@ static void __devinit pci_bus_slot_names(struct
> device_node *node,
>
> static int __init of_pci_slot_init(void)
> {
> - struct pci_bus *pbus = NULL;
> + struct pci_host_bridge *host_bridge = NULL;
> + struct pci_bus *pbus;
>
> - while ((pbus = pci_find_next_bus(pbus)) != NULL) {
> + for_each_pci_host_bridge(host_bridge) {
> struct device_node *node;
>
> + pbus = hot_bridge->bus;
> if (pbus->self) {
> /* PCI->PCI bridge */
> node = pbus->self->dev.of_node;
that is for initial booting path.
for hot add/remove notifier add need to be done case by case.
-Yinghai
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
