[PATCH] jffs2: fix potential deadlock in jffs2_do_setattr

[Liu Song]

In jffs2_do_setattr, we hold two mutexes which are 
A) c->alloc_sem 
B) f->sem 
If new_metadata is error, then we release mutexes in bad sequence 
which could cause ABAB deadlock. 
This patch adjusts release sequence which could avoid deadlock. 

Signed-off-by: Liu Song <liu.son...@zte.com.cn>
Reviewed-by: Jiang Biao <jiang.bi...@zte.com.cn>
---
 fs/jffs2/fs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/jffs2/fs.c b/fs/jffs2/fs.c
index 76fa814..f998126 100644
--- a/fs/jffs2/fs.c
+++ b/fs/jffs2/fs.c
@@ -140,9 +140,9 @@ int jffs2_do_setattr (struct inode *inode, struct iattr 
*iattr)
                kfree(mdata);
 
        if (IS_ERR(new_metadata)) {
-               jffs2_complete_reservation(c);
                jffs2_free_raw_inode(ri);
                mutex_unlock(&f->sem);
+               jffs2_complete_reservation(c);
                return PTR_ERR(new_metadata);
        }
        /* It worked. Update the inode */
-- 
2.1.0.GIT