Re: [PATCH RFC] mm: madvise: use walk_page_range_vma() for madvise_free_single_vma()

[Anshuman Khandual]

On 6/3/25 07:01, Barry Song wrote:
> From: Barry Song <v-songbaohua@xxxxxxxx>
> 
> We've already found the VMA before calling madvise_free_single_vma(),
> so calling walk_page_range() and doing find_vma() again seems
> unnecessary. It also prevents potential optimizations for MADV_FREE
> to use a per-VMA lock.
> 
> Cc: "Liam R. Howlett" <Liam.Howlett@xxxxxxxxxx>
> Cc: Lorenzo Stoakes <lorenzo.stoakes@xxxxxxxxxx>
> Cc: David Hildenbrand <david@xxxxxxxxxx>
> Cc: Vlastimil Babka <vbabka@xxxxxxx>
> Cc: Jann Horn <jannh@xxxxxxxxxx>
> Cc: Suren Baghdasaryan <surenb@xxxxxxxxxx>
> Cc: Lokesh Gidra <lokeshgidra@xxxxxxxxxx>
> Cc: Tangquan Zheng <zhengtangquan@xxxxxxxx>
> Signed-off-by: Barry Song <v-songbaohua@xxxxxxxx>
> ---
>  mm/madvise.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/mm/madvise.c b/mm/madvise.c
> index d408ffa404b3..c6a28a2d3ff8 100644
> --- a/mm/madvise.c
> +++ b/mm/madvise.c
> @@ -826,7 +826,7 @@ static int madvise_free_single_vma(struct madvise_behavior *madv_behavior,
>  
>  	mmu_notifier_invalidate_range_start(&range);
>  	tlb_start_vma(tlb, vma);
> -	walk_page_range(vma->vm_mm, range.start, range.end,
> +	walk_page_range_vma(vma, range.start, range.end,
>  			&madvise_free_walk_ops, tlb);
>  	tlb_end_vma(tlb, vma);
>  	mmu_notifier_invalidate_range_end(&range);

This indeed looks like an improvement dropping a redundant VMA look up
and also opening up potential optimization later using per-VMA lock.

Reviewed-by: Anshuman Khandual <anshuman.khandual@xxxxxxx>