mm/gup.c: allow CMA migration to propagate errors back to caller
authorzhong jiang <zhongjiang@huawei.com>
Sun, 1 Dec 2019 01:49:50 +0000 (17:49 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Sun, 1 Dec 2019 14:29:18 +0000 (06:29 -0800)
check_and_migrate_cma_pages() was recording the result of
__get_user_pages_locked() in an unsigned "nr_pages" variable.

Because __get_user_pages_locked() returns a signed value that can
include negative errno values, this had the effect of hiding errors.

Change check_and_migrate_cma_pages() implementation so that it uses a
signed variable instead, and propagates the results back to the caller
just as other gup internal functions do.

This was discovered with the help of unsigned_lesser_than_zero.cocci.

Link: http://lkml.kernel.org/r/1571671030-58029-1-git-send-email-zhongjiang@huawei.com
Signed-off-by: zhong jiang <zhongjiang@huawei.com>
Suggested-by: John Hubbard <jhubbard@nvidia.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/gup.c

index 8f236a3..c2b3e11 100644 (file)
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1443,6 +1443,7 @@ static long check_and_migrate_cma_pages(struct task_struct *tsk,
        bool drain_allow = true;
        bool migrate_allow = true;
        LIST_HEAD(cma_page_list);
+       long ret = nr_pages;
 
 check_again:
        for (i = 0; i < nr_pages;) {
@@ -1504,17 +1505,18 @@ check_again:
                 * again migrating any new CMA pages which we failed to isolate
                 * earlier.
                 */
-               nr_pages = __get_user_pages_locked(tsk, mm, start, nr_pages,
+               ret = __get_user_pages_locked(tsk, mm, start, nr_pages,
                                                   pages, vmas, NULL,
                                                   gup_flags);
 
-               if ((nr_pages > 0) && migrate_allow) {
+               if ((ret > 0) && migrate_allow) {
+                       nr_pages = ret;
                        drain_allow = true;
                        goto check_again;
                }
        }
 
-       return nr_pages;
+       return ret;
 }
 #else
 static long check_and_migrate_cma_pages(struct task_struct *tsk,