Re: [PATCH v5.1] fs: Allow fine-grained control of folio sizes

["Pankaj Raghav (Samsung)" <kernel@pankajraghav.com>]

On Mon, May 27, 2024 at 10:01:23PM +0100, Matthew Wilcox (Oracle) wrote:
> We need filesystems to be able to communicate acceptable folio sizes
> to the pagecache for a variety of uses (e.g. large block sizes).
> Support a range of folio sizes between order-0 and order-31.
> 
> Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
> Co-developed-by: Pankaj Raghav <p.raghav@samsung.com>
> Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
> Reviewed-by: Hannes Reinecke <hare@suse.de>
> ---
> For this version, I fixed the TODO that the maximum folio size was not
> being honoured.  I made some other changes too like adding const, moving
> the location of the constants, checking CONFIG_TRANSPARENT_HUGEPAGE, and
> dropping some of the functions which aren't needed until later patches.
> (They can be added in the commits that need them).  Also rebased against
> current Linus tree, so MAX_PAGECACHE_ORDER no longer needs to be moved).

Thanks for this! So I am currently running my xfstests on the new series
I am planning to send in a day or two based on next-20240523.

I assume this patch is intended to be folded in to the next LBS series?

> 
> diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h
> index 1ed9274a0deb..c6aaceed0de6 100644
> --- a/include/linux/pagemap.h
> +++ b/include/linux/pagemap.h
> @@ -204,13 +204,18 @@ enum mapping_flags {
>  	AS_EXITING	= 4, 	/* final truncate in progress */
>  	/* writeback related tags are not used */
>  	AS_NO_WRITEBACK_TAGS = 5,
> -	AS_LARGE_FOLIO_SUPPORT = 6,
> -	AS_RELEASE_ALWAYS,	/* Call ->release_folio(), even if no private data */
> -	AS_STABLE_WRITES,	/* must wait for writeback before modifying
> +	AS_RELEASE_ALWAYS = 6,	/* Call ->release_folio(), even if no private data */
> +	AS_STABLE_WRITES = 7,	/* must wait for writeback before modifying
>  				   folio contents */
> -	AS_UNMOVABLE,		/* The mapping cannot be moved, ever */
> +	AS_UNMOVABLE = 8,	/* The mapping cannot be moved, ever */
> +	AS_FOLIO_ORDER_MIN = 16,
> +	AS_FOLIO_ORDER_MAX = 21, /* Bits 16-25 are used for FOLIO_ORDER */
>  };
>  
> +#define AS_FOLIO_ORDER_MIN_MASK 0x001f0000
> +#define AS_FOLIO_ORDER_MAX_MASK 0x03e00000

As you changed the mapping flag offset, these masks also needs to be
changed accordingly.

I moved(pun intended) the AS_UNMOVABLE and kept the
AS_FOLIO_ORDER_(MIN|MAX) value the same.

	AS_FOLIO_ORDER_MIN = 8,
	AS_FOLIO_ORDER_MAX = 13, /* Bit 8-17 are used for FOLIO_ORDER */
	AS_UNMOVABLE = 18,	 /* The mapping cannot be moved, ever */
	AS_INACCESSIBLE,	/* Do not attempt direct R/W access to the mapping */

> +#define AS_FOLIO_ORDER_MASK (AS_FOLIO_ORDER_MIN_MASK | AS_FOLIO_ORDER_MAX_MASK)
> +
>  
--
Pankaj