Re: [PATCH 03/13] filemap: align the index to mapping_min_order in the page cache

From: "Pankaj Raghav (Samsung)" <kernel@pankajraghav.com>
To: Matthew Wilcox <willy@infradead.org>
Cc: linux-xfs@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	 linux-kernel@vger.kernel.org, david@fromorbit.com,
	chandan.babu@oracle.com,  akpm@linux-foundation.org,
	mcgrof@kernel.org, ziy@nvidia.com, hare@suse.de,
	 djwong@kernel.org, gost.dev@samsung.com, linux-mm@kvack.org,
	 Pankaj Raghav <p.raghav@samsung.com>
Subject: Re: [PATCH 03/13] filemap: align the index to mapping_min_order in the page cache
Date: Tue, 27 Feb 2024 11:06:37 +0100	[thread overview]
Message-ID: <37kubwweih4zwvxzvjbhnhxunrafawdqaqggzcw6xayd6vtrfl@dllnk6n53akf> (raw)
In-Reply-To: <Zdyi6lFDAHXi8GPz@casper.infradead.org>

On Mon, Feb 26, 2024 at 02:40:42PM +0000, Matthew Wilcox wrote:
> On Mon, Feb 26, 2024 at 10:49:26AM +0100, Pankaj Raghav (Samsung) wrote:
> > From: Luis Chamberlain <mcgrof@kernel.org>
> > 
> > Supporting mapping_min_order implies that we guarantee each folio in the
> > page cache has at least an order of mapping_min_order. So when adding new
> > folios to the page cache we must ensure the index used is aligned to the
> > mapping_min_order as the page cache requires the index to be aligned to
> > the order of the folio.
> 
> This seems like a remarkably complicated way of achieving:
> 
> diff --git a/mm/filemap.c b/mm/filemap.c
> index 5603ced05fb7..36105dad4440 100644
> --- a/mm/filemap.c
> +++ b/mm/filemap.c
> @@ -2427,9 +2427,11 @@ static int filemap_update_page(struct kiocb *iocb,
>  }
>  
>  static int filemap_create_folio(struct file *file,
> -		struct address_space *mapping, pgoff_t index,
> +		struct address_space *mapping, loff_t pos,
>  		struct folio_batch *fbatch)
>  {
> +	pgoff_t index;
> +	unsigned int min_order;
>  	struct folio *folio;
>  	int error;
>  
> @@ -2451,6 +2453,8 @@ static int filemap_create_folio(struct file *file,
>  	 * well to keep locking rules simple.
>  	 */
>  	filemap_invalidate_lock_shared(mapping);
> +	min_order = mapping_min_folio_order(mapping);
> +	index = (pos >> (min_order + PAGE_SHIFT)) << min_order;

That is some cool mathfu. I will add a comment here as it might not be
that obvious to some people (i.e me).

Thanks.

>  	error = filemap_add_folio(mapping, folio, index,
>  			mapping_gfp_constraint(mapping, GFP_KERNEL));
>  	if (error == -EEXIST)
> @@ -2511,8 +2515,7 @@ static int filemap_get_pages(struct kiocb *iocb, size_t count,
>  	if (!folio_batch_count(fbatch)) {
>  		if (iocb->ki_flags & (IOCB_NOWAIT | IOCB_WAITQ))
>  			return -EAGAIN;
> -		err = filemap_create_folio(filp, mapping,
> -				iocb->ki_pos >> PAGE_SHIFT, fbatch);
> +		err = filemap_create_folio(filp, mapping, iocb->ki_pos, fbatch);
>  		if (err == AOP_TRUNCATED_PAGE)
>  			goto retry;
>  		return err;