f2fs: cover free_nid management with spin_lock
authorJaegeuk Kim <jaegeuk.kim@samsung.com>
Tue, 7 May 2013 11:47:40 +0000 (20:47 +0900)
committerJaegeuk Kim <jaegeuk.kim@samsung.com>
Wed, 8 May 2013 10:54:22 +0000 (19:54 +0900)
commit59bbd474abb9dd6a0c1a74df758ec29c7a8b150f
tree0e1f7f74108b0e9cc4a7aa8a9d80e333344a366a
parent23d38844276680abbf33624f56b6779d43f53633
f2fs: cover free_nid management with spin_lock

After build_free_nids() searches free nid candidates from nat pages and
current journal blocks, it checks all the candidates if they are allocated
so that the nat cache has its nid with an allocated block address.

In this procedure, previously we used
    list_for_each_entry_safe(fnid, next_fnid, &nm_i->free_nid_list, list).
But, this is not covered by free_nid_list_lock, resulting in null pointer bug.

This patch moves this checking routine inside add_free_nid() in order not to use
the spin_lock.

Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
fs/f2fs/node.c