On 18/12/2025 19:19, Andrew Cooper wrote:
> On 18/12/2025 5:28 pm, Oleksii Kurochko wrote:
>> diff --git a/xen/arch/arm/domain.c b/xen/arch/arm/domain.c
>> index 47973f99d9..e566023340 100644
>> --- a/xen/arch/arm/domain.c
>> +++ b/xen/arch/arm/domain.c
>> @@ -473,36 +473,21 @@ void dump_pageframe_info(struct domain *d)
>>
>> }
>>
>> -/*
>> - * The new VGIC has a bigger per-IRQ structure, so we need more than one
>> - * page on ARM64. Cowardly increase the limit in this case.
>> - */
>> -#if defined(CONFIG_NEW_VGIC) && defined(CONFIG_ARM_64)
>> -#define MAX_PAGES_PER_VCPU 2
>> -#else
>> -#define MAX_PAGES_PER_VCPU 1
>> -#endif
>> -
>> struct vcpu *alloc_vcpu_struct(const struct domain *d)
>> {
>> struct vcpu *v;
>>
>> - BUILD_BUG_ON(sizeof(*v) > MAX_PAGES_PER_VCPU * PAGE_SIZE);
>> - v = alloc_xenheap_pages(get_order_from_bytes(sizeof(*v)), 0);
>> - if ( v != NULL )
>> - {
>> - unsigned int i;
>> -
>> - for ( i = 0; i < DIV_ROUND_UP(sizeof(*v), PAGE_SIZE); i++ )
>> - clear_page((void *)v + i * PAGE_SIZE);
>> - }
>> + BUILD_BUG_ON(sizeof(*v) > PAGE_SIZE);
>> + v = alloc_xenheap_pages(0, 0);
>
> I know this is only interim until the next patch, but
> alloc_xenheap_page() to match the free function used.
>
> Personally, I'd merge patches 2 and 3 together, because everything you
> touch in this patch is deleted by the next one.
>
> But, whatever the ARM maintainers prefer.
I'm in favor of Andrew's suggestion. There's no point in introducing something
in one patch and dropping it in the very next one.
~Michal
