f2fs: bugfix for alloc_nid_failed()

Directly drop the free_nid cache when nm_i->fcnt > 2 * MAX_FREE_NIDS

Since there is NOT nmi->free_nid_list_lock spinlock protection between
a sequential calling of alloc_nid() and alloc_nid_failed(), some other
threads may already add new free_nid to the free_nid_list during this
period.

We need to make sure nmi->fcnt is never > 2 * MAX_FREE_NIDS.

Signed-off-by: Haicheng Li <haicheng.li@linux.intel.com>
[Jaegeuk Kim: fit the coding style]
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>

diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index 7209d637f942d532dd018b4b6f06ef746c91a810..d682f34ee0f855413c759221ce54f735985d8fb1 100644 (file)
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -1439,8 +1439,12 @@ void alloc_nid_failed(struct f2fs_sb_info *sbi, nid_t nid)
        spin_lock(&nm_i->free_nid_list_lock);
        i = __lookup_free_nid_list(nid, &nm_i->free_nid_list);
        BUG_ON(!i || i->state != NID_ALLOC);
-       i->state = NID_NEW;
-       nm_i->fcnt++;
+       if (nm_i->fcnt > 2 * MAX_FREE_NIDS) {
+               __del_from_free_nid_list(i);
+       } else {
+               i->state = NID_NEW;
+               nm_i->fcnt++;
+       }
        spin_unlock(&nm_i->free_nid_list_lock);
 }