On 06.05.2022 11:56, Julien Grall wrote:
>
>
> On 06/05/2022 08:27, Michal Orzel wrote:
>> On 05.05.2022 13:13, Jan Beulich wrote:
>>> On 05.05.2022 12:36, Michal Orzel wrote:
>>>> --- a/xen/arch/arm/include/asm/grant_table.h
>>>> +++ b/xen/arch/arm/include/asm/grant_table.h
>>>> @@ -29,12 +29,21 @@ static inline void gnttab_mark_dirty(struct domain *d,
>>>> mfn_t mfn)
>>>> #endif
>>>> }
>>>> +static inline bool gnttab_host_mapping_get_page_type(bool ro, struct
>>>> domain *ld,
>>>> + struct domain *rd)
>>>> +{
>>>> + return false;
>>>> +}
>>>> +
>>>> +static inline bool gnttab_release_host_mappings(struct domain *d)
>>>> +{
>>>> + return true;
>>>> +}
>>>
>>> Looking at x86 I think all three instances of struct domain * want to
>>> be const struct domain *. Then
>>> Reviewed-by: Jan Beulich <[email protected]>
>>>
>>> Jan
>>>
>> Thanks. I think we should mark all parameters as const meaning also const
>> bool ro.
>
> Hmmmm... ro is not a pointer and so the value can only be modified within the
> inline helper. So isn't it a bit pointless to set it to const?
>
> If that's the only comment on the next version, this could be dealt on commit.
>
> Cheers,
>
>From the code point of view it is pointless.
However it is also about self-documenting the code.
If this is something that cannot occur in Xen, I'd be greatful for dealing with
this on commit.
Cheers,
Michal
