On Mon, Jul 22, 2013 at 6:08 PM, Bjorn Helgaas <bhelgaas@xxxxxxxxxx> wrote:
> On Wed, Jul 3, 2013 at 8:04 AM, Mika Westerberg
> <mika.westerberg@xxxxxxxxxxxxxxx> wrote:
>> In hotplug case (especially with Thunderbolt enabled systems) we might need
>> to call pcibios_resource_survey_bus() several times for a bus. The function
>> ends up calling pci_claim_resource() for each bridge resource that then
>> fails claiming that the resource exists already (which it does). Once this
>> happens the resource is invalidated thus preventing devices behind the
>> bridge to allocate their resources.
>>
>> To fix this we do what has been done in pcibios_allocate_dev_resources()
>> and check 'parent' of the given resource. If it is non-NULL it means that
>> the resource has been allocated already and we can skip it. We do the same
>> for ROM resources as well.
>>
>> Signed-off-by: Mika Westerberg <mika.westerberg@xxxxxxxxxxxxxxx>
>
> This looks reasonable to me.
Acked-by: Bjorn Helgaas <bhelgaas@xxxxxxxxxx>
> However, the use of "r->parent" to determine whether the resource is
> already allocated is nothing specific to x86. So I think we might be
> missing an opportunity to make this more consistent across
> architectures. I looked at other callers of pci_claim_resource() for
> bridge and ROM resources, and it looks like we might be able to make
> similar changes in:
>
> frv pcibios_allocate_bus_resources()
> ia64 pcibios_fixup_resources() (via pcibios_fixup_bridge_resources())
> mn10300 pcibios_allocate_bus_resources()
> mn10300 pcibios_assign_resources() (ROM)
> mn10300 pcibios_fixup_device_resources()
> parisc lba_fixup_bus()
>
> I really hate the idea of fixing something for x86 but not for other
> arches, so maybe somebody could take a deeper look at this and see if
> it would make sense to change the other arches, too.
>
> Bjorn
>
>> ---
>> arch/x86/pci/i386.c | 4 ++++
>> 1 file changed, 4 insertions(+)
>>
>> diff --git a/arch/x86/pci/i386.c b/arch/x86/pci/i386.c
>> index 94919e3..db6b1ab 100644
>> --- a/arch/x86/pci/i386.c
>> +++ b/arch/x86/pci/i386.c
>> @@ -210,6 +210,8 @@ static void pcibios_allocate_bridge_resources(struct pci_dev *dev)
>> r = &dev->resource[idx];
>> if (!r->flags)
>> continue;
>> + if (r->parent) /* Already allocated */
>> + continue;
>> if (!r->start || pci_claim_resource(dev, idx) < 0) {
>> /*
>> * Something is wrong with the region.
>> @@ -318,6 +320,8 @@ static void pcibios_allocate_dev_rom_resource(struct pci_dev *dev)
>> r = &dev->resource[PCI_ROM_RESOURCE];
>> if (!r->flags || !r->start)
>> return;
>> + if (r->parent) /* Already allocated */
>> + return;
>>
>> if (pci_claim_resource(dev, PCI_ROM_RESOURCE) < 0) {
>> r->end -= r->start;
>> --
>> 1.8.3.2
>>
--
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
