JFS is missing a memory barrier

JFS is missing a memory barrier needed to close the critical section before
clearing the lock bit. Use lock bitops for this.

unlock_page() has a second barrier after clearing the lock, which is
required because it checks whether the waitqueue is active without locks.
Such a barrier is not required here because the waitqueue spinlock is
always taken (something to think about if performance is an issue).

Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>

diff --git a/fs/jfs/jfs_metapage.c b/fs/jfs/jfs_metapage.c
index b27fa117f229b86a636769b8de6d9f2ee3dfe080..1dfaae5adf1bae79c314f53c5f49979327ab680d 100644 (file)
--- a/fs/jfs/jfs_metapage.c
+++ b/fs/jfs/jfs_metapage.c
@@ -39,11 +39,11 @@ static struct {
 #endif
 
 #define metapage_locked(mp) test_bit(META_locked, &(mp)->flag)
-#define trylock_metapage(mp) test_and_set_bit(META_locked, &(mp)->flag)
+#define trylock_metapage(mp) test_and_set_bit_lock(META_locked, &(mp)->flag)
 
 static inline void unlock_metapage(struct metapage *mp)
 {
-       clear_bit(META_locked, &mp->flag);
+       clear_bit_unlock(META_locked, &mp->flag);
        wake_up(&mp->wait);
 }