[8/8] mm/gup: make get_user_pages_fast_only() return the common return value

Commit Message

Jason Gunthorpe Jan. 17, 2023, 3:58 p.m. UTC

There are only two callers, both can handle the common return code:

- get_user_page_fast_only() checks == 1

- gfn_to_page_many_atomic() already returns -1, and the only caller
  checks for -ve

Remove the restriction against returning negative values.

Signed-off-by: Jason Gunthorpe <[e-mail]>
---
 mm/gup.c | 17 +----------------
 1 file changed, 1 insertion(+), 16 deletions(-)

Comments

John Hubbard Jan. 20, 2023, 7:27 p.m. UTC | #1

On 1/17/23 07:58, Jason Gunthorpe wrote:
> There are only two callers, both can handle the common return code:
> 
> - get_user_page_fast_only() checks == 1
> 
> - gfn_to_page_many_atomic() already returns -1, and the only caller
>   checks for -ve

Maybe "checks for negative return values" ?

> 
> Remove the restriction against returning negative values.
> 
> Signed-off-by: Jason Gunthorpe <[e-mail]>
> ---
>  mm/gup.c | 17 +----------------
>  1 file changed, 1 insertion(+), 16 deletions(-)
> 

Reviewed-by: John Hubbard <[e-mail]>


thanks,

David Hildenbrand Jan. 23, 2023, 11:33 a.m. UTC | #2

On 17.01.23 16:58, Jason Gunthorpe wrote:
> There are only two callers, both can handle the common return code:
> 
> - get_user_page_fast_only() checks == 1
> 
> - gfn_to_page_many_atomic() already returns -1, and the only caller
>    checks for -ve
> 
> Remove the restriction against returning negative values.
> 
> Signed-off-by: Jason Gunthorpe <[e-mail]>
> ---

Reviewed-by: David Hildenbrand <[e-mail]>

Patch

diff --git a/mm/gup.c b/mm/gup.c
index b260c182135587..20ebc4d1f0d719 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -3002,8 +3002,6 @@  static int internal_get_user_pages_fast(unsigned long start,
  *
  * Like get_user_pages_fast() except it's IRQ-safe in that it won't fall back to
  * the regular GUP.
- * Note a difference with get_user_pages_fast: this always returns the
- * number of pages pinned, 0 if no pages were pinned.
  *
  * If the architecture does not support this function, simply return with no
  * pages pinned.
@@ -3015,7 +3013,6 @@  static int internal_get_user_pages_fast(unsigned long start,
 int get_user_pages_fast_only(unsigned long start, int nr_pages,
 			     unsigned int gup_flags, struct page **pages)
 {
-	int nr_pinned;
 	/*
 	 * Internally (within mm/gup.c), gup fast variants must set FOLL_GET,
 	 * because gup fast is always a "pin with a +1 page refcount" request.
@@ -3027,19 +3024,7 @@  int get_user_pages_fast_only(unsigned long start, int nr_pages,
 			       FOLL_GET | FOLL_FAST_ONLY))
 		return -EINVAL;
 
-	nr_pinned = internal_get_user_pages_fast(start, nr_pages, gup_flags,
-						 pages);
-
-	/*
-	 * As specified in the API description above, this routine is not
-	 * allowed to return negative values. However, the common core
-	 * routine internal_get_user_pages_fast() *can* return -errno.
-	 * Therefore, correct for that here:
-	 */
-	if (nr_pinned < 0)
-		nr_pinned = 0;
-
-	return nr_pinned;
+	return internal_get_user_pages_fast(start, nr_pages, gup_flags, pages);
 }
 EXPORT_SYMBOL_GPL(get_user_pages_fast_only);