> On 22 May 2023, at 13:41, Jan Beulich <jbeul...@suse.com> wrote:
>
> On 22.05.2023 12:20, Luca Fancellu wrote:
>>
>>
>>> On 18 May 2023, at 19:30, Julien Grall <jul...@xen.org> wrote:
>>>
>>> Hi Luca,
>>>
>>> One more remark.
>>>
>>> On 24/04/2023 07:02, Luca Fancellu wrote:
>>>> #else /* !CONFIG_ARM64_SVE */
>>>> @@ -46,6 +50,15 @@ static inline unsigned int get_sys_vl_len(void)
>>>> return 0;
>>>> }
>>>> +static inline int sve_context_init(struct vcpu *v)
>>>> +{
>>>> + return 0;
>>>
>>> The call is protected by is_domain_sve(). So I think we want to return an
>>> error just in case someone is calling it outside of its intended use.
>>
>> Regarding this one, since it should not be called when SVE is not enabled,
>> are you ok if I’ll do this:
>>
>> static inline int sve_context_init(struct vcpu *v)
>> {
>> ASSERT_UNREACHABLE();
>> return 0;
>> }
>
> Do you need such a stub in the first place? I.e. can't you arrange for
> DCE to take care of unreachable function calls, thus letting you get
> away with just an always-visible declaration (and no definition when
> !ARM64_SVE)?
>
Right, I always forgot about this approach, I’ll try that
> Jan
