Re: [PATCH v3 1/3] mm, swap: make SWAPFILE_CLUSTER runtime

[Kairui Song]

On Fri, Jun 12, 2026 at 11:39:15PM +0800, Ritesh Harjani (IBM) wrote:
> On PowerPC Book3S64, MMU is selected at runtime, so macros like
> PMD_SHIFT are effectively runtime variables in the Book3S64 code. THP
> swap code uses these macros to size some of its array data structures
> based on PMD_ORDER e.g. SWAPFILE_CLUSTER macro is used for this very
> purpose.
> Hence this patch initializes SWAPFILE_CLUSTER at runtime and also
> modifies swap_table and swap_memcg_table which were earlier using this
> macro for defining the number of table entries.
> 
> Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@xxxxxxxxx>
> ---
>  mm/swap.h       |  5 +++--
>  mm/swap_table.h |  6 ++----
>  mm/swapfile.c   | 27 ++++++++++++++++++++++-----
>  3 files changed, 27 insertions(+), 11 deletions(-)

Hi Ritesh,

Thanks for the patch.

> 
> diff --git a/mm/swap.h b/mm/swap.h
> index 77d2d14eda42..956879a69ddd 100644
> --- a/mm/swap.h
> +++ b/mm/swap.h
> @@ -26,11 +26,12 @@ extern int page_cluster;
>  #define SWAP_TABLE_HAS_ZEROFLAG		((BITS_PER_LONG - SWAP_CACHE_PFN_MARK_BITS - \
>  					  SWAP_CACHE_PFN_BITS) > SWAP_COUNT_MIN_BITS)
> 
> +extern unsigned int swap_slots_in_cluster __read_mostly;

Maybe __ro_after_init is better for this kind of use case?

> +#define SWAPFILE_CLUSTER	swap_slots_in_cluster
> +
>  #ifdef CONFIG_THP_SWAP
> -#define SWAPFILE_CLUSTER	HPAGE_PMD_NR

So on Book3S64, HPAGE_PMD_NR is also a variable right? Then we don't really
need to change the SWAPFILE_CLUSTER defination here? We just need to adjust
the users of this macro so the build will pass?

Or maybe use another macro instead of HPAGE_PMD_NR here, whichever is
more arch friendly.

That way if that is a build time constant, all users are folded to
just one mask/shift which is super efficient especially for inline helpers,
and there are a lot of them. Only special archs live with the dynamic
load overhead.