mm/migrate.c: handle freed page at the first place
authorYang Shi <yang.shi@linux.alibaba.com>
Sun, 1 Dec 2019 01:57:12 +0000 (17:57 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Sun, 1 Dec 2019 20:59:09 +0000 (12:59 -0800)
When doing migration if the freed page is met, we just return without
migrating it since it is pointless to migrate a freed page.  But, the
current code allocates target page unconditionally before handling freed
page, if the page is freed, the newly allocated will be just freed.  It
doesn't make too much sense and is just a waste of time although
migrating freed page is rare.

So, handle freed page at the before that to avoid unnecessary page
allocation and free.

Link: http://lkml.kernel.org/r/1573755869-106954-1-git-send-email-yang.shi@linux.alibaba.com
Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/migrate.c

index 4fe45d1428c8676133cf17d4a84e9cbbd6d27bbb..a8f87cb432516b427b97aa10668edc8655b59751 100644 (file)
@@ -1168,15 +1168,11 @@ static ICE_noinline int unmap_and_move(new_page_t get_new_page,
                                   enum migrate_reason reason)
 {
        int rc = MIGRATEPAGE_SUCCESS;
-       struct page *newpage;
+       struct page *newpage = NULL;
 
        if (!thp_migration_supported() && PageTransHuge(page))
                return -ENOMEM;
 
-       newpage = get_new_page(page, private);
-       if (!newpage)
-               return -ENOMEM;
-
        if (page_count(page) == 1) {
                /* page was freed from under us. So we are done. */
                ClearPageActive(page);
@@ -1187,13 +1183,13 @@ static ICE_noinline int unmap_and_move(new_page_t get_new_page,
                                __ClearPageIsolated(page);
                        unlock_page(page);
                }
-               if (put_new_page)
-                       put_new_page(newpage, private);
-               else
-                       put_page(newpage);
                goto out;
        }
 
+       newpage = get_new_page(page, private);
+       if (!newpage)
+               return -ENOMEM;
+
        rc = __unmap_and_move(page, newpage, force, mode);
        if (rc == MIGRATEPAGE_SUCCESS)
                set_page_owner_migrate_reason(newpage, reason);