Hi Oscar, hi Michael,
Le 20/05/2024 à 11:14, Oscar Salvador a écrit :
> On Fri, May 17, 2024 at 09:00:00PM +0200, Christophe Leroy wrote:
>> set_huge_pte_at() expects the real page size, not the psize which is
>
> "expects the size of the huge page" sounds bettter?
Parameter 'pzize' already provides the size of the hugepage, but not in
the way set_huge_pte_at() expects it.
psize has one of the values defined by MMU_PAGE_XXX macros defined in
arch/powerpc/include/asm/mmu.h while set_huge_pte_at() expects the size
as a value.
>
>> the index of the page definition in table mmu_psize_defs[]
>>
>> Fixes: 935d4f0c6dc8 ("mm: hugetlb: add huge page size param to set_huge_pte_at()")
>> Signed-off-by: Christophe Leroy <christophe.leroy@xxxxxxxxxx>
>
> Reviewed-by: Oscar Salvador <osalvador@xxxxxxx>
>
> AFAICS, this fixup is not related to the series, right? (yes, you will
> the parameter later)
> I would have it at the very beginning of the series.
You are right, I should have submitted it separately.
Michael can you take it as a fix for 6.10 ?
>
>
>> ---
>> arch/powerpc/mm/nohash/8xx.c | 3 ++-
>> 1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/arch/powerpc/mm/nohash/8xx.c b/arch/powerpc/mm/nohash/8xx.c
>> index 43d4842bb1c7..d93433e26ded 100644
>> --- a/arch/powerpc/mm/nohash/8xx.c
>> +++ b/arch/powerpc/mm/nohash/8xx.c
>> @@ -94,7 +94,8 @@ static int __ref __early_map_kernel_hugepage(unsigned long va, phys_addr_t pa,
>> return -EINVAL;
>>
>> set_huge_pte_at(&init_mm, va, ptep,
>> - pte_mkhuge(pfn_pte(pa >> PAGE_SHIFT, prot)), psize);
>> + pte_mkhuge(pfn_pte(pa >> PAGE_SHIFT, prot)),
>> + 1UL << mmu_psize_to_shift(psize));
>>
>> return 0;
>> }
>> --
>> 2.44.0
>>
>
