Re: [PATCH v8 01/10] arm64/mm: add pte_set_huge() and pte_clear_huge()

[Wen Jiang]

On Fri, 18 Sept 2026 at 07:22, Barry Song <baohua@xxxxxxxxxx> wrote:
>
> On Thu, Sep 17, 2026 at 1:29 PM Wen Jiang <jiangwenxiaomi@xxxxxxxxx> wrote:
> >
> > From: Wen Jiang <jiangwen6@xxxxxxxxxx>
> >
> > vmap installs PTE-level block mappings (PTE_CONT on arm64) by reusing
> > set_huge_pte_at() and huge_ptep_get_and_clear(), which are HugeTLB
> > helpers 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()/pte_clear_huge() to arm64, joining the existing
> > pmd_set_huge()/pud_set_huge() family in mm/mmu.c. They build on
> > __set_ptes() and __get_and_clear_full_ptes() directly, so they do not
> > depend on CONFIG_HUGETLB_PAGE.
> >
> > 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@xxxxxxxxxx>
>
> I tried to apply the patch to mm-new, but it doesn't apply.
>
> % git am raw
> Applying: arm64/mm: add pte_set_huge() and pte_clear_huge()
> error: patch failed: arch/arm64/mm/mmu.c:1872
> error: arch/arm64/mm/mmu.c: patch does not apply
> Patch failed at 0001 arm64/mm: add pte_set_huge() and pte_clear_huge()
> hint: Use 'git am --show-current-patch=diff' to see the failed patch
> When you have resolved this problem, run "git am --continue".
> If you prefer to skip this patch, run "git am --skip" instead.
> To restore the original branch and stop patching, run "git am --abort".
>
> % patch -p1 < raw
> patching file 'arch/arm64/include/asm/pgtable.h'
> patching file 'arch/arm64/mm/mmu.c'
> 1 out of 1 hunks failed--saving rejects to 'arch/arm64/mm/mmu.c.rej'
>
>
> [...]
>
> > +void pte_set_huge(pte_t *ptep, unsigned long addr, phys_addr_t phys,
> > +                 pgprot_t prot, unsigned long size)
> > +{
> > +       unsigned long pfn = __phys_to_pfn(phys);
> > +       pte_t new_pte = pte_mkcont(pfn_pte(pfn, prot));
>
> Maybe we can drop the `new_` prefix.
>

Will do.

> > +       unsigned int nr = size >> PAGE_SHIFT;
> > +
> > +       VM_WARN_ON(!IS_ALIGNED(pfn, CONT_PTES));
> > +       VM_WARN_ON(!IS_ALIGNED(size, CONT_PTE_SIZE));
>
> Do we also need to check that `addr` is aligned?
>

Yes. arch_vmap_pte_range_map_size() already guarantees all three,
since we are already checking two of them here, we should complete the
set. I'll add it in v9.

Thanks,
Wen