On Tue, Aug 3, 2010 at 3:51 PM, Dan Carpenter <erro...@gmail.com> wrote:
> On Tue, Aug 03, 2010 at 11:35:17PM +0200, Julia Lawall wrote:
>> diff --git a/arch/powerpc/kernel/pci_of_scan.c
>> b/arch/powerpc/kernel/pci_of_scan.c
>> index 6ddb795..62dd363 100644
>> --- a/arch/powerpc/kernel/pci_of_scan.c
>> +++ b/arch/powerpc/kernel/pci_of_scan.c
>> @@ -336,8 +336,7 @@ static void __devinit __of_scan_bus(struct device_node
>> *node,
>> if (dev->hdr_type == PCI_HEADER_TYPE_BRIDGE ||
>> dev->hdr_type == PCI_HEADER_TYPE_CARDBUS) {
>> struct device_node *child = pci_device_to_OF_node(dev);
>> - if (dev)
>> - of_scan_pci_bridge(child, dev);
>> + of_scan_pci_bridge(child, dev);
>
> The intention was probably to check "child" instead of "dev".
> pci_device_to_OF_node() can return NULL. On the other hand the code
> has been this way for a year and no one has complained...
Still, it should be fixed. It is likely that I'll be generalizing
this code for other architectures in the near future. I'll spin a
patch.
g.
_______________________________________________
Linuxppc-dev mailing list
Linuxppc-dev@lists.ozlabs.org
https://lists.ozlabs.org/listinfo/linuxppc-dev
