Re: [PATCH 06/24] swap: rework swapin_no_readahead arguments

[Chris Li <chrisl@kernel.org>]

On Sun, Nov 19, 2023 at 11:48 AM Kairui Song <ryncsn@gmail.com> wrote:
>
> From: Kairui Song <kasong@tencent.com>
>
> Make it use alloc_pages_mpol instead of vma_alloc_folio, and accept
> mm_struct directly as an argument instead of taking a vmf as argument.
> Make its arguments similar to swap_{cluster,vma}_readahead, to
> make the code more aligned.

It is unclear to me what is the value of making the
swap_{cluster,vma}_readahead more aligned here.

> Also prepare for following commits which will skip vmf for certain
> swapin paths.

The following patch 07/24 does not seem to use this patch.
Can it merge with the other patch that uses it?

As it is, this patch does not serve a stand alone value to justify it.

Chris

>
> Signed-off-by: Kairui Song <kasong@tencent.com>
> ---
>  mm/swap_state.c | 15 +++++++--------
>  1 file changed, 7 insertions(+), 8 deletions(-)
>
> diff --git a/mm/swap_state.c b/mm/swap_state.c
> index fd0047ae324e..ff6756f2e8e4 100644
> --- a/mm/swap_state.c
> +++ b/mm/swap_state.c
> @@ -867,17 +867,17 @@ static struct page *swap_vma_readahead(swp_entry_t targ_entry, gfp_t gfp_mask,
>   * in.
>   */
>  static struct page *swapin_no_readahead(swp_entry_t entry, gfp_t gfp_mask,
> -                                       struct vm_fault *vmf)
> +                                       struct mempolicy *mpol, pgoff_t ilx,
> +                                       struct mm_struct *mm)
>  {
> -       struct vm_area_struct *vma = vmf->vma;
> -       struct page *page = NULL;
>         struct folio *folio;
> +       struct page *page;
>         void *shadow = NULL;
>
> -       folio = vma_alloc_folio(GFP_HIGHUSER_MOVABLE, 0,
> -                               vma, vmf->address, false);
> +       page = alloc_pages_mpol(gfp_mask, 0, mpol, ilx, numa_node_id());
> +       folio = (struct folio *)page;
>         if (folio) {
> -               if (mem_cgroup_swapin_charge_folio(folio, vma->vm_mm,
> +               if (mem_cgroup_swapin_charge_folio(folio, mm,
>                                                    GFP_KERNEL, entry)) {
>                         folio_put(folio);
>                         return NULL;
> @@ -896,7 +896,6 @@ static struct page *swapin_no_readahead(swp_entry_t entry, gfp_t gfp_mask,
>
>                 /* To provide entry to swap_readpage() */
>                 folio->swap = entry;
> -               page = &folio->page;
>                 swap_readpage(page, true, NULL);
>                 folio->private = NULL;
>         }
> @@ -928,7 +927,7 @@ struct page *swapin_readahead(swp_entry_t entry, gfp_t gfp_mask,
>
>         mpol = get_vma_policy(vmf->vma, vmf->address, 0, &ilx);
>         if (swap_use_no_readahead(swp_swap_info(entry), entry)) {
> -               page = swapin_no_readahead(entry, gfp_mask, vmf);
> +               page = swapin_no_readahead(entry, gfp_mask, mpol, ilx, vmf->vma->vm_mm);
>                 cached = false;
>         } else if (swap_use_vma_readahead()) {
>                 page = swap_vma_readahead(entry, gfp_mask, mpol, ilx, vmf);
> --
> 2.42.0
>
>