On 18.07.2022 12:53, Marek Marczykowski-Górecki wrote:
> On Thu, Jul 14, 2022 at 12:17:53PM +0200, Jan Beulich wrote:
>> On 06.07.2022 17:32, Marek Marczykowski-Górecki wrote:
>>> --- a/xen/drivers/passthrough/iommu.c
>>> +++ b/xen/drivers/passthrough/iommu.c
>>> @@ -651,6 +651,46 @@ bool_t iommu_has_feature(struct domain *d, enum
>>> iommu_feature feature)
>>> return is_iommu_enabled(d) && test_bit(feature,
>>> dom_iommu(d)->features);
>>> }
>>>
>>> +#define MAX_EXTRA_RESERVED_RANGES 20
>>> +struct extra_reserved_range {
>>> + xen_pfn_t start;
>>
>> Why not unsigned long?
>
> Isn't this the correct type for the page number?
In the public interface - yes. But internally we (almost) universally
use unsigned long. xen_pfn_t is inefficient to use on Arm32 (and then
presumably also other future 32-bit counterparts of 64-bit
architectures, inheriting the choices made for Arm and differing from
what we have on x86).
Jan