Re: [PATCH 1/2] xen: make include/xen/unaligned.h usable on all architectures

[Jan Beulich <jbeulich@xxxxxxxx>]

On 05.12.2023 15:01, Julien Grall wrote:
> On 05/12/2023 13:59, Jan Beulich wrote:
>> On 05.12.2023 14:46, Julien Grall wrote:
>>> On 05/12/2023 13:41, Juergen Gross wrote:
>>>> On 05.12.23 14:31, Julien Grall wrote:
>>>>> Anyway, given you don't seem to have a use-case yet, I would simply to
>>>>> consider to surround the declaration with an a config which can be
>>>>> selected if unaligned access is supported.
>>>>
>>>> Like in xen/common/lzo.c et al?
>>>
>>> Just to clarify, I am suggesting to add in unaligned.h:
>>>
>>> #ifdef CONFIG_HAS_UNALIGNED_ACCESS
>>>
>>> your definitions
>>>
>>> #endif
>>
>> But that would be wrong: HAS_UNALIGNED_ACCESS would be there to indicate
>> one does _not_ need any special accessors.
> 
> I am guessing you are disagreeing on the name rather than the concept? 
> If so, what about CONFIG_UNALIGNED_ACCESS_ALLOWED?

Not really, no. Of course the name needs to properly express the purpose.
But I don't see why a Kconfig control would be appropriate here. You simply
want to provide accessors to unaligned data. Nobody needs to use them, but
when you have to, they ought to be there. A Kconfig control (of the name
you suggested first) would be helpful to not penalize architectures which
can do unaligned accesses without any helpers (in case the code generated
through the helpers turned out sub-optimal).

Jan