[PATCH v8 02/10] powerpc/8xx: add pte_set_huge()

[Christophe Leroy (CS GROUP)]

Le 18/09/2026 à 10:44, Barry Song a écrit :
> On Fri, Sep 18, 2026 at 2:14 PM Christophe Leroy (CS GROUP)
> <chleroy at kernel.org> wrote:
>>
>>
>>
>> Le 17/09/2026 à 07:29, Wen Jiang a écrit :
>>> From: Wen Jiang <jiangwen6 at xiaomi.com>
>>>
>>> vmap installs PTE-level block mappings (SZ_16K/SZ_512K on powerpc/8xx)
>>> by reusing set_huge_pte_at(), a HugeTLB helper gated by
>>> CONFIG_HUGETLB_PAGE. This makes the feature silently unavailable on
>>> CONFIG_HUGETLB_PAGE=n kernels and couples mm/vmalloc.c to HugeTLB
>>> internals it does not otherwise need.
>>>
>>> Add pte_set_huge() to powerpc/8xx, next to the existing
>>> pmd_clear_huge()/pud_clear_huge() in mm/nohash/8xx.c. It builds the huge
>>> PTE and writes it into the backing cells directly, without going through
>>> set_huge_pte_at(), so it does not depend on CONFIG_HUGETLB_PAGE.
>>>
>>> No pte_clear_huge() is needed: 8xx does not implement
>>> arch_vmap_pte_range_unmap_size(), so the vmap unmap path never takes the
>>> block-mapping branch.
>>>
>>> In practice PPC_8xx selects HUGETLBFS unconditionally, so
>>> CONFIG_HUGETLB_PAGE=n does not occur there today; this is a decoupling
>>> cleanup rather than a new configuration.
>>
>> 8xx selects CONFIG_HUGETLB_PAGE because it uses set_huge_pte_at() to map
>> kernel text with large pages. That was added by commit 34536d780683
>> ("powerpc/8xx: Add a function to early map kernel via huge pages")
>>
>> But now pte_set_huge() could be used instead in
>> __early_map_kernel_hugepage() and then CONFIG_HUGETLB_PAGE could be removed.
> 
> Hi Christophe,
> 
> Do you mean the code below? If so, it sounds like a good place to
> clean this up for powerpc as a separate cleanup patch.

Yes this is what I have in mind allthough I have not studied it in details.

> 
> diff --git a/arch/powerpc/mm/nohash/8xx.c b/arch/powerpc/mm/nohash/8xx.c
> index 2518d2a6bdb4..830cdb391161 100644
> --- a/arch/powerpc/mm/nohash/8xx.c
> +++ b/arch/powerpc/mm/nohash/8xx.c
> @@ -87,9 +87,7 @@ static int __ref
> __early_map_kernel_hugepage(unsigned long va, phys_addr_t pa,
>          if (WARN_ON(!ptep))
>                  return -ENOMEM;
> 
> -       set_huge_pte_at(&init_mm, va, ptep,
> -                       arch_make_huge_pte(pfn_pte(pa >> PAGE_SHIFT,
> prot), shift, 0),
> -                       1UL << shift);
> +       pte_set_huge(ptep, va, pa, prot, 1UL << shift);
> 
>          return 0;
>   }
> diff --git a/arch/powerpc/platforms/Kconfig.cputype
> b/arch/powerpc/platforms/Kconfig.cputype
> index f0adbf7ecb7f..efba4d960083 100644
> --- a/arch/powerpc/platforms/Kconfig.cputype
> +++ b/arch/powerpc/platforms/Kconfig.cputype
> @@ -41,7 +41,6 @@ config PPC_8xx
>          select FSL_SOC
>          select PPC_KUEP
>          select HAVE_ARCH_VMAP_STACK
> -       select HUGETLBFS
> 
>   config 44x
>          bool "AMCC 44x, 46x or 47x"
> 
>>
>>>
>>> There is no caller yet: mm/vmalloc.c is converted later in this series,
>>> once the generic fallbacks are in place.
>>>
>>> Signed-off-by: Wen Jiang <jiangwen6 at xiaomi.com>
>>
>> Reviewed-by: Christophe Leroy (CS GROUP) <chleroy at kernel.org>
> 
> Best Regards
> Barry