Re: [PATCH v4 07/36] fs: Add a filesystem flag for large pages

From: Dave Chinner
Date: Thu May 21 2020 - 17:55:36 EST


On Fri, May 15, 2020 at 06:16:27AM -0700, Matthew Wilcox wrote:
> From: "Matthew Wilcox (Oracle)" <willy@xxxxxxxxxxxxx>
>
> The page cache needs to know whether the filesystem supports pages >
> PAGE_SIZE.
>
> Signed-off-by: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx>
> ---
> include/linux/fs.h | 1 +
> include/linux/pagemap.h | 5 +++++
> 2 files changed, 6 insertions(+)
>
> diff --git a/include/linux/fs.h b/include/linux/fs.h
> index 55c743925c40..777783c8760b 100644
> --- a/include/linux/fs.h
> +++ b/include/linux/fs.h
> @@ -2241,6 +2241,7 @@ struct file_system_type {
> #define FS_HAS_SUBTYPE 4
> #define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */
> #define FS_DISALLOW_NOTIFY_PERM 16 /* Disable fanotify permission events */
> +#define FS_LARGE_PAGES 8192 /* Remove once all fs converted */
> #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */
> int (*init_fs_context)(struct fs_context *);
> const struct fs_parameter_spec *parameters;
> diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h
> index 36bfc9d855bb..c6db74b5e62f 100644
> --- a/include/linux/pagemap.h
> +++ b/include/linux/pagemap.h
> @@ -116,6 +116,11 @@ static inline void mapping_set_gfp_mask(struct address_space *m, gfp_t mask)
> m->gfp_mask = mask;
> }
>
> +static inline bool mapping_large_pages(struct address_space *mapping)
> +{
> + return mapping->host->i_sb->s_type->fs_flags & FS_LARGE_PAGES;
> +}

If you've got to dereference 4 layers deep to check a behaviour
flag, the object needs it's own flag. Can you just propagate this
to the address space when the inode is instantiated and the address
space initialised?

Cheers,

Dave.
--
Dave Chinner
david@xxxxxxxxxxxxx