[Intel-gfx] [PATCH 04/22] drm/i915/gemfs: enable THP

[Chris Wilson]

Quoting Matthew Auld (2017-09-25 19:47:19)
> Enable transparent-huge-pages through gemfs by mounting with
> huge=within_size.
> 
> v2: prefer kern_mount_data
> 
> Signed-off-by: Matthew Auld <matthew.auld at intel.com>
> Cc: Joonas Lahtinen <joonas.lahtinen at linux.intel.com>
> Cc: Chris Wilson <chris at chris-wilson.co.uk>
> Cc: Dave Hansen <dave.hansen at intel.com>
> Cc: Kirill A. Shutemov <kirill at shutemov.name>
> Cc: Andrew Morton <akpm at linux-foundation.org>
> Cc: Hugh Dickins <hughd at google.com>
> Cc: linux-mm at kvack.org
> ---
>  drivers/gpu/drm/i915/i915_gemfs.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_gemfs.c b/drivers/gpu/drm/i915/i915_gemfs.c
> index 168d0bd98f60..dc35719814f0 100644
> --- a/drivers/gpu/drm/i915/i915_gemfs.c
> +++ b/drivers/gpu/drm/i915/i915_gemfs.c
> @@ -24,6 +24,7 @@
>  
>  #include <linux/fs.h>
>  #include <linux/mount.h>
> +#include <linux/pagemap.h>
>  
>  #include "i915_drv.h"
>  #include "i915_gemfs.h"
> @@ -32,12 +33,17 @@ int i915_gemfs_init(struct drm_i915_private *i915)
>  {
>         struct file_system_type *type;
>         struct vfsmount *gemfs;
> +       char within_size[] = "huge=within_size";
> +       char *options = NULL;
>  
>         type = get_fs_type("tmpfs");
>         if (!type)
>                 return -ENODEV;
>  

/*
 * Enable hugepages that fit into the object's size so that
 * we do not overallocate, as the excess will not be reused and
 * our objects may be extremely long lived.
 */

> -       gemfs = kern_mount(type);
> +       if (has_transparent_hugepage())
> +               options = within_size;
> +
> +       gemfs = kern_mount_data(type, options);

The alternative to the previous two patches would be to export a
function to set the hugepage policy on the shmem_sb_info.

>         if (IS_ERR(gemfs))
>                 return PTR_ERR(gemfs);

Reviewed-by: Chris Wilson <chris at chris-wilson.co.uk>
-Chris