On Mon, Feb 27, 2012 at 9:57 AM, Bjorn Helgaas <bhelgaas@xxxxxxxxxx> wrote:
> On Sun, Feb 26, 2012 at 4:33 PM, Yinghai Lu <yinghai@xxxxxxxxxx> wrote:
>> use that device for pci_root_bus bridge pointer.
>>
>> With that make code more simple.
>>
>> Also we can use pci_release_bus_bridge_dev() to release allocated
>> pci_host_bridge during removing path.
>>
>> At last, we can use root bus bridge pointer to get host bridge pointer instead
>> of going over host bridge list, so could kill that host bridge list.
>>
>> Signed-off-by: Yinghai Lu <yinghai@xxxxxxxxxx>
>>
>> ---
>> drivers/pci/host-bridge.c | 14 ---------
>> drivers/pci/pci.h | 2 -
>> drivers/pci/probe.c | 65 ++++++++++++++++++++++++----------------------
>> include/linux/pci.h | 4 ++
>> 4 files changed, 38 insertions(+), 47 deletions(-)
>>
>> Index: linux-2.6/drivers/pci/host-bridge.c
>> ===================================================================
>> --- linux-2.6.orig/drivers/pci/host-bridge.c
>> +++ linux-2.6/drivers/pci/host-bridge.c
>> - dev = kzalloc(sizeof(*dev), GFP_KERNEL);
>> - if (!dev)
>> - goto err_dev;
>> + return NULL;
>>
...
here old code allocate dummy struct device.
>> + bridge = pci_alloc_host_bridge(b);
>> + if (!bridge)
>> + goto err_out;
>> +
>> + bridge->dev.parent = parent;
>> + bridge->dev.release = pci_release_bus_bridge_dev;
>> + dev_set_name(&bridge->dev, "pci%04x:%02x", pci_domain_nr(b), bus);
>> + error = device_register(&bridge->dev);
>> if (error)
>> - goto dev_reg_err;
>> - b->bridge = get_device(dev);
>> + goto bridge_dev_reg_err;
>> + b->bridge = get_device(&bridge->dev);
...
>> --- linux-2.6.orig/include/linux/pci.h
>> +++ linux-2.6/include/linux/pci.h
>> @@ -375,11 +375,13 @@ struct pci_host_bridge_window {
>> };
>>
>> struct pci_host_bridge {
>> - struct list_head list;
>> + struct device dev;
>> struct pci_bus *bus; /* root bus */
>> struct list_head windows; /* pci_host_bridge_windows */
>> };
>
> This doesn't feel right to me. You're allocating a new struct device
> here, but the arch likely already has one. In fact, I think we even
> passed it in to pci_create_root_bus().
no.
original pci_create_root_bus() will create one dummy struct device and
use it as parent for bus->dev.
>
> On x86, we already have an ACPI device for the host bridge, and now
> we'll have a second one. (Actually a *third* one, because PNPACPI
> also has one, but that's a long-standing problem.)
not related.
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
