Re: [PATCH 04/24] mm/swap: avoid setting page lock bit and doing extra unlock check
From: Chris Li
Date: Sun Nov 19 2023 - 23:18:17 EST
On Sun, Nov 19, 2023 at 11:48 AM Kairui Song <ryncsn@xxxxxxxxx> wrote:
>
> From: Kairui Song <kasong@xxxxxxxxxxx>
>
> When swapping in a page, mem_cgroup_swapin_charge_folio is called for new
> allocated folio, nothing else is referencing the folio so no need to set
> the lock bit. This avoided doing unlock check on error path.
>
> Signed-off-by: Kairui Song <kasong@xxxxxxxxxxx>
> ---
> mm/swap_state.c | 20 +++++++++-----------
> 1 file changed, 9 insertions(+), 11 deletions(-)
>
> diff --git a/mm/swap_state.c b/mm/swap_state.c
> index ac4fa404eaa7..45dd8b7c195d 100644
> --- a/mm/swap_state.c
> +++ b/mm/swap_state.c
> @@ -458,6 +458,8 @@ struct page *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask,
You move the mem_cgroup_swapin_charge_folio() inside the for loop:
for (;;) {
int err;
/*
* First check the swap cache. Since this is normally
* called after swap_cache_get_folio() failed, re-calling
* that would confuse statistics.
*/
folio = filemap_get_folio(swap_address_space(entry),
swp_offset(entry));
> mpol, ilx, numa_node_id());
> if (!folio)
> goto fail_put_swap;
> + if (mem_cgroup_swapin_charge_folio(folio, NULL, gfp_mask, entry))
> + goto fail_put_folio;
Wouldn't it cause repeat charging of the folio when it is racing
against others in the for loop?
>
> /*
> * Swap entry may have been freed since our caller observed it.
> @@ -483,13 +485,9 @@ struct page *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask,
> /*
> * The swap entry is ours to swap in. Prepare the new page.
> */
> -
> __folio_set_locked(folio);
> __folio_set_swapbacked(folio);
>
> - if (mem_cgroup_swapin_charge_folio(folio, NULL, gfp_mask, entry))
> - goto fail_unlock;
> -
The original code makes the charge outside of the for loop. Only the
winner can charge once.
Chris