mm: migrate: hugetlb: putback destination hugepage to active list
authorNaoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Tue, 22 Sep 2015 21:59:14 +0000 (14:59 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 22 Sep 2015 22:09:53 +0000 (15:09 -0700)
Since commit bcc54222309c ("mm: hugetlb: introduce page_huge_active")
each hugetlb page maintains its active flag to avoid a race condition
betwe= en multiple calls of isolate_huge_page(), but current kernel
doesn't set the f= lag on a hugepage allocated by migration because the
proper putback routine isn= 't called.  This means that users could
still encounter the race referred to by bcc54222309c in this special
case, so this patch fixes it.

Fixes: bcc54222309c ("mm: hugetlb: introduce page_huge_active")
Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: <stable@vger.kernel.org> [4.1.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/migrate.c

index c3cb566af3e273a92e8353835b1cd6d03d64c7e3..7452a00bbb50c134b529c1d024dfc53fcfca093b 100644 (file)
@@ -1075,7 +1075,7 @@ out:
        if (rc != MIGRATEPAGE_SUCCESS && put_new_page)
                put_new_page(new_hpage, private);
        else
-               put_page(new_hpage);
+               putback_active_hugepage(new_hpage);
 
        if (result) {
                if (rc)