Re: [RFC PATCH v1] mm/madvise: prefetch private file COW swap entries

["Lorenzo Stoakes (ARM)" <ljs@kernel.org>]

On Fri, Sep 04, 2026 at 06:09:19PM +0800, zhaozhengzhuo wrote:
...


> Fixes: 1998cc048901 ("mm: make madvise(MADV_WILLNEED) support swap file prefetch")

Fixes totally inappropriate.

> Reported-by: Mike Kaplinskiy <mike@recall.ai>

Missing Closes tag.

> Link: https://lore.kernel.org/linux-mm/CABeknB_S2XJSHFgnHdgnN0rjzHhH4oQJs_APq9fvxHztQ_pgiA@mail.gmail.com/
> Signed-off-by: zhaozhengzhuo <zhaozhengzhuo@uniontech.com>

(Very likely) missing Assisted-by tag.

Please follow kernel procedure.

https://docs.kernel.org/process/coding-assistants.html
https://docs.kernel.org/process/generated-content.html

I already sent at patch at
https://lore.kernel.org/linux-mm/apm68kgHC0NK2zVI@gremlin/ before this one, so
I'm going to take care of this thanks.

> ---
>  mm/madvise.c | 32 ++++++++++++++++++++++++++------
>  1 file changed, 26 insertions(+), 6 deletions(-)
>
> diff --git a/mm/madvise.c b/mm/madvise.c
> index 73c2901b9adb..96cbce6c7f8b 100644
> --- a/mm/madvise.c
> +++ b/mm/madvise.c
> @@ -193,10 +193,16 @@ static int madvise_update_vma(vm_flags_t new_flags,
>  }
>
>  #ifdef CONFIG_SWAP
> +struct swapin_walk_ctx {
> +	struct vm_area_struct *vma;
> +	bool swapped;
> +};

This is silly.

> +
>  static int swapin_walk_pmd_entry(pmd_t *pmd, unsigned long start,
>  		unsigned long end, struct mm_walk *walk)
>  {
> -	struct vm_area_struct *vma = walk->private;
> +	struct swapin_walk_ctx *swc = walk->private;
> +	struct vm_area_struct *vma = swc->vma;
>  	struct swap_io_ctx ctx = {};
>  	pte_t *ptep = NULL;
>  	spinlock_t *ptl;
> @@ -223,8 +229,10 @@ static int swapin_walk_pmd_entry(pmd_t *pmd, unsigned long start,
>
>  		folio = read_swap_cache_async(&ctx, entry, GFP_HIGHUSER_MOVABLE,
>  					vma, addr);
> -		if (folio)
> +		if (folio) {
> +			swc->swapped = true;
>  			folio_put(folio);
> +		}

This is pointless, it's not a good trade off.

>  	}
>
>  	if (ptep)
> @@ -297,10 +305,22 @@ static long madvise_willneed(struct madvise_behavior *madv_behavior)
>  	loff_t offset;
>
>  #ifdef CONFIG_SWAP
> -	if (!file) {
> -		walk_page_range_vma(vma, start, end, &swapin_walk_ops, vma);
> -		lru_add_drain(); /* Push any new pages onto the LRU now */
> -		return 0;
> +	bool private_file = file && !(vma->vm_flags & VM_SHARED);

This is not correct. My version uses the correct CoW predicate.

> +
> +	/*
> +	 * A private file mapping can contain anonymous COW pages. Once such
> +	 * pages are swapped out, their PTEs contain swap entries even though
> +	 * the VMA still has vm_file set. Prefetch those pages as well; file
> +	 * readahead can only fetch the original file contents.
> +	 */
> +	if (!file || (private_file && vma->anon_vma)) {
> +		struct swapin_walk_ctx ctx = { .vma = vma };
> +
> +		walk_page_range_vma(vma, start, end, &swapin_walk_ops, &ctx);
> +		if (ctx.swapped)
> +			lru_add_drain(); /* Push any new pages onto the LRU now */
> +		if (!file)
> +			return 0;

General structure is OK, but again I already sent a patch.

So please no v2, I will handle this thanks!

>  	}
>
>  	if (shmem_mapping(file->f_mapping)) {
> --
> 2.43.0

--
Cheers, Lorenzo