On 10/01/2019 14:17, Robin Murphy wrote:
> On 10/01/2019 13:35, Jean-Philippe Brucker wrote:
>> When executed for a PCI_ROOT_COMPLEX type, iort_match_node_callback()
>> expects the opaque pointer argument to be a PCI bus device. At the
>> moment rc_dma_get_range() passes the PCI endpoint instead of the bus,
>> and we've been lucky to have pci_domain_nr(ptr) return 0 instead of
>> crashing. Pass the bus device to iort_scan_node().
>
> Oops - that makes sense, and I can now see the combination of
> circumstances that would have led me to overlook it at the time. Sorry
> about that! However..
>
>> Reported-by: Eric Auger <eric.auger@xxxxxxxxxx>
>> Fixes: 5ac65e8c8941 ("ACPI/IORT: Support address size limit for root complexes")
>> Signed-off-by: Jean-Philippe Brucker <jean-philippe.brucker@xxxxxxx>
>> ---
>> drivers/acpi/arm64/iort.c | 3 ++-
>> 1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/acpi/arm64/iort.c b/drivers/acpi/arm64/iort.c
>> index fdd90ffceb85..c08afe44c488 100644
>> --- a/drivers/acpi/arm64/iort.c
>> +++ b/drivers/acpi/arm64/iort.c
>> @@ -952,9 +952,10 @@ static int rc_dma_get_range(struct device *dev, u64 *size)
>> {
>> struct acpi_iort_node *node;
>> struct acpi_iort_root_complex *rc;
>> + struct pci_bus *pbus = to_pci_dev(dev)->bus;
>
> ...is this still right if the device is beyond a switch or other bridge
> and not on the root bus?
My understanding is that the pci_dev still has a valid ->bus
representing its nearest upstream bridge, and the domain_nr attribute of
each bus is copied from its parent by pci_alloc_bus(). Other places in
the kernel seem to do the same thing, including iort_iommu_configure()
in this file
Thanks,
Jean
