On 02.11.2023 10:24, Oleksii wrote:
> On Thu, 2023-11-02 at 09:39 +0100, Jan Beulich wrote:
>> On 01.11.2023 11:15, Oleksii Kurochko wrote:
>>> --- a/xen/include/xen/pci.h
>>> +++ b/xen/include/xen/pci.h
>>> @@ -68,7 +68,18 @@ typedef union {
>>> };
>>> } pci_sbdf_t;
>>>
>>> +#ifdef CONFIG_HAS_PCI
>>> #include <asm/pci.h>
>>> +#else
>>
>> This minimal scope of the #ifdef will do for now, but will likely
>> want
>> extending down the road. Even what's visible in context is already an
>> entity which should be entirely unused in the code base when
>> !HAS_PCI.
>>
>>> +struct arch_pci_dev { };
>>> +
>>> +static always_inline bool is_pci_passthrough_enabled(void)
>>
>> Perhaps s/always_inline/inline/ as this is moved here. We really
>> shouldn't
>> use always_inline unless actually have a clear purpose.
> Could it be fixed during the commit ( in case there won't be any other
> critical comments about this patch )?
Oh, yes, sure. I should have said so.
Jan
