mm/gup: follow_pfn_pte(): -EEXIST cleanup
authorJohn Hubbard <jhubbard@nvidia.com>
Tue, 22 Mar 2022 21:39:40 +0000 (14:39 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 22 Mar 2022 22:57:01 +0000 (15:57 -0700)
Remove a quirky special case from follow_pfn_pte(), and adjust its
callers to match.  Caller changes include:

__get_user_pages(): Regardless of any FOLL_* flags, get_user_pages() and
its variants should handle PFN-only entries by stopping early, if the
caller expected **pages to be filled in.  This makes for a more reliable
API, as compared to the previous approach of skipping over such entries
(and thus leaving them silently unwritten).

move_pages(): squash the -EEXIST error return from follow_page() into
-EFAULT, because -EFAULT is listed in the man page, whereas -EEXIST is
not.

Link: https://lkml.kernel.org/r/20220204020010.68930-3-jhubbard@nvidia.com
Signed-off-by: John Hubbard <jhubbard@nvidia.com>
Suggested-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Jan Kara <jack@suse.cz>
Cc: Peter Xu <peterx@redhat.com>
Cc: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Claudio Imbrenda <imbrenda@linux.ibm.com>
Cc: Alex Williamson <alex.williamson@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/gup.c
mm/migrate.c

index 41da0bd..43ad2bc 100644 (file)
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -464,10 +464,6 @@ static struct page *no_page_table(struct vm_area_struct *vma,
 static int follow_pfn_pte(struct vm_area_struct *vma, unsigned long address,
                pte_t *pte, unsigned int flags)
 {
-       /* No page to get reference */
-       if (flags & (FOLL_GET | FOLL_PIN))
-               return -EFAULT;
-
        if (flags & FOLL_TOUCH) {
                pte_t entry = *pte;
 
@@ -1205,8 +1201,15 @@ retry:
                } else if (PTR_ERR(page) == -EEXIST) {
                        /*
                         * Proper page table entry exists, but no corresponding
-                        * struct page.
+                        * struct page. If the caller expects **pages to be
+                        * filled in, bail out now, because that can't be done
+                        * for this page.
                         */
+                       if (pages) {
+                               ret = PTR_ERR(page);
+                               goto out;
+                       }
+
                        goto next_page;
                } else if (IS_ERR(page)) {
                        ret = PTR_ERR(page);
index c7da064..be0d5ae 100644 (file)
@@ -1762,6 +1762,13 @@ static int do_pages_move(struct mm_struct *mm, nodemask_t task_nodes,
                }
 
                /*
+                * The move_pages() man page does not have an -EEXIST choice, so
+                * use -EFAULT instead.
+                */
+               if (err == -EEXIST)
+                       err = -EFAULT;
+
+               /*
                 * If the page is already on the target node (!err), store the
                 * node, otherwise, store the err.
                 */