On Tue, Nov 03 2020 at 09:54, Marc Zyngier wrote:
> On 2020-11-02 22:18, Thomas Gleixner wrote:
>> On Mon, Nov 02 2020 at 17:16, Thomas Gleixner wrote:
>>> On Mon, Nov 02 2020 at 11:30, Marc Zyngier wrote:
>>>> --- a/drivers/pci/probe.c
>>>> +++ b/drivers/pci/probe.c
>>>> @@ -871,6 +871,8 @@ static void pci_set_bus_msi_domain(struct pci_bus
>>>> *bus)
>>>> d = pci_host_bridge_msi_domain(b);
>>>>
>>>> dev_set_msi_domain(&bus->dev, d);
>>>> + if (!d)
>>>> + bus->bus_flags |= PCI_BUS_FLAGS_NO_MSI;
>>>
>>> Hrm, that might break legacy setups (no irqdomain support). I'd rather
>>> prefer to explicitly tell the pci core at host registration time.
>>
>> s/might break/ breaks / Just validated :)
>
> For my own edification, can you point me to the failing case?
Any architecture which selects PCI_MSI_ARCH_FALLBACKS and does not have
irqdomain support runs into:
if (!d)
bus->bus_flags |= PCI_BUS_FLAGS_NO_MSI;
which in turn makes pci_msi_supported() return 0 and consequently makes
pci_enable_msi[x]() fail.
Thanks,
tglx
