[PATCH] mm/gup: might_lock_read(mmap_sem) in get_user_pages_fast()

[John Hubbard]

Instead of scattering these assertions across the drivers,
do this assertion inside the core of get_user_pages_fast*()
functions. That also includes pin_user_pages_fast*()
routines.

Add a might_lock_read(mmap_sem) call to internal_get_user_pages_fast().

Suggested-by: Matthew Wilcox <willy@xxxxxxxxxxxxx>
Cc: Michel Lespinasse <walken@xxxxxxxxxx>
Cc: Jason Gunthorpe <jgg@xxxxxxxx>
Signed-off-by: John Hubbard <jhubbard@xxxxxxxxxx>
---

Hi Andrew,

This applies on top of [1], which in turn applies to
today's (20200521) linux-next.

As noted in the discussion [2], this will need changing from
mmap_sem to mmap_lock, after Michel Lespinasse's patchset
arrives.

[1]
https://lore.kernel.org/r/20200521233841.1279742-1-jhubbard@xxxxxxxxxx

[2] https://lore.kernel.org/linux-mm/20200520124817.GG31189@xxxxxxxx/

thanks,
John Hubbard
NVIDIA

 mm/gup.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/mm/gup.c b/mm/gup.c
index ada6aa79576dc..3462c076e8ecf 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -2728,6 +2728,9 @@ static int internal_get_user_pages_fast(unsigned long start, int nr_pages,
 				       FOLL_FAST_ONLY)))
 		return -EINVAL;
 
+	if (!(gup_flags & FOLL_FAST_ONLY))
+		might_lock_read(&current->mm->mmap_sem);
+
 	start = untagged_addr(start) & PAGE_MASK;
 	addr = start;
 	len = (unsigned long) nr_pages << PAGE_SHIFT;
-- 
2.26.2