Re: [PATCH V3 3/4] mm/sparse-vmemmap: Generalise vmemmap_populate_hugepages()

[Will Deacon]

On Sat, Jul 02, 2022 at 04:00:20PM +0800, Huacai Chen wrote:
> From: Feiyang Chen <chenfeiyang@xxxxxxxxxxx>
> 
> Generalise vmemmap_populate_hugepages() so ARM64 & X86 & LoongArch can
> share its implementation.
> 
> Signed-off-by: Feiyang Chen <chenfeiyang@xxxxxxxxxxx>
> Signed-off-by: Huacai Chen <chenhuacai@xxxxxxxxxxx>
> ---
>  arch/arm64/mm/mmu.c      | 53 ++++++-----------------
>  arch/loongarch/mm/init.c | 63 ++++++++-------------------
>  arch/x86/mm/init_64.c    | 92 ++++++++++++++--------------------------
>  include/linux/mm.h       |  6 +++
>  mm/sparse-vmemmap.c      | 54 +++++++++++++++++++++++
>  5 files changed, 124 insertions(+), 144 deletions(-)
> 
> diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c
> index 626ec32873c6..b080a65c719d 100644
> --- a/arch/arm64/mm/mmu.c
> +++ b/arch/arm64/mm/mmu.c
> @@ -1158,49 +1158,24 @@ int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node,
>  	return vmemmap_populate_basepages(start, end, node, altmap);
>  }
>  #else	/* !ARM64_KERNEL_USES_PMD_MAPS */
> +void __meminit vmemmap_set_pmd(pmd_t *pmd, void *p, int node,
> +			       unsigned long addr, unsigned long next)
> +{
> +	pmd_set_huge(pmd, __pa(p), __pgprot(PROT_SECT_NORMAL));
> +}
> +
> +int __meminit vmemmap_check_pmd(pmd_t *pmd, int node, unsigned long addr,
> +				unsigned long next)
> +{
> +	vmemmap_verify((pte_t *)pmd, node, addr, next);
> +	return 1;
> +}

nit, but please can you use 'pmdp' instead of 'pmd' for the pointers? We're
pretty consistent elsewhere for arch/arm64 and it makes the READ_ONCE()
usage easier to follow once functions end up loading the entry.

Thanks,

Will