On Fri, 2025-03-14 at 05:05 +0800, Kemeng Shi wrote:
> 1. Factor out general swap_entries_put_map() helper to drop entries belong
> to one cluster. If entries are last map, free entries in batch, otherwise
> put entries with cluster lock acquired and released only once.
> 2. Iterate and call swap_entries_put_map() for each cluster in
> swap_entries_put_nr() to leverage batch-remove for last map belong to one
> cluster and reduce lock acquire/release in fallback case.
> 3. As swap_entries_put_nr() won't handle SWAP_HSA_CACHE drop, rename it to
> swap_entries_put_map_nr().
>
> Signed-off-by: Kemeng Shi <shikemeng@xxxxxxxxxxxxxxx>
> ---
> mm/swapfile.c | 58 +++++++++++++++++++++++++--------------------------
> 1 file changed, 29 insertions(+), 29 deletions(-)
>
> diff --git a/mm/swapfile.c b/mm/swapfile.c
> index 2d0f5d630211..ebac9ff74ba7 100644
> --- a/mm/swapfile.c
> +++ b/mm/swapfile.c
> @@ -1473,25 +1473,10 @@ struct swap_info_struct *get_swap_device(swp_entry_t entry)
> return NULL;
> }
>
> -static unsigned char swap_entry_put(struct swap_info_struct *si,
> - swp_entry_t entry)
Need to update comment in free_swap_and_cache_nr() now that swap_entry_put() goes away.
* the swap once per folio in the common case. If we do
* swap_entry_put() and __try_to_reclaim_swap() in the same loop, the
* latter will get a reference and lock the folio for every individual ...
> +static bool swap_entries_put_map(struct swap_info_struct *si,
> + swp_entry_t entry, int nr)
> {
> - struct swap_cluster_info *ci;
> unsigned long offset = swp_offset(entry);
> - unsigned char usage;
> -
> - ci = lock_cluster(si, offset);
> - usage = swap_entry_put_locked(si, ci, entry, 1);
> - unlock_cluster(ci);
> -
> - return usage;
> -}
> -
