fs: Establish locking order for unrelated directories

Currently the locking order of inode locks for directories that are not
in ancestor relationship is not defined because all operations that
needed to lock two directories like this were serialized by
sb->s_vfs_rename_mutex. However some filesystems need to lock two
subdirectories for RENAME_EXCHANGE operations and for this we need the
locking order established even for two tree-unrelated directories.
Provide a helper function lock_two_inodes() that establishes lock
ordering for any two inodes and use it in lock_two_directories().

CC: stable@vger.kernel.org
Signed-off-by: Jan Kara <jack@suse.cz>
Message-Id: <20230601105830.13168-4-jack@suse.cz>
Signed-off-by: Christian Brauner <brauner@kernel.org>

diff --git a/fs/inode.c b/fs/inode.c
index 577799b7855f6f91c75a6f100877811e3a853d08..b9d4980322700257066618563be707fc72bbbceb 100644 (file)
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -1103,6 +1103,48 @@ void discard_new_inode(struct inode *inode)
 }
 EXPORT_SYMBOL(discard_new_inode);
 
+/**
+ * lock_two_inodes - lock two inodes (may be regular files but also dirs)
+ *
+ * Lock any non-NULL argument. The caller must make sure that if he is passing
+ * in two directories, one is not ancestor of the other.  Zero, one or two
+ * objects may be locked by this function.
+ *
+ * @inode1: first inode to lock
+ * @inode2: second inode to lock
+ * @subclass1: inode lock subclass for the first lock obtained
+ * @subclass2: inode lock subclass for the second lock obtained
+ */
+void lock_two_inodes(struct inode *inode1, struct inode *inode2,
+                    unsigned subclass1, unsigned subclass2)
+{
+       if (!inode1 || !inode2) {
+               /*
+                * Make sure @subclass1 will be used for the acquired lock.
+                * This is not strictly necessary (no current caller cares) but
+                * let's keep things consistent.
+                */
+               if (!inode1)
+                       swap(inode1, inode2);
+               goto lock;
+       }
+
+       /*
+        * If one object is directory and the other is not, we must make sure
+        * to lock directory first as the other object may be its child.
+        */
+       if (S_ISDIR(inode2->i_mode) == S_ISDIR(inode1->i_mode)) {
+               if (inode1 > inode2)
+                       swap(inode1, inode2);
+       } else if (!S_ISDIR(inode1->i_mode))
+               swap(inode1, inode2);
+lock:
+       if (inode1)
+               inode_lock_nested(inode1, subclass1);
+       if (inode2 && inode2 != inode1)
+               inode_lock_nested(inode2, subclass2);
+}
+
 /**
  * lock_two_nondirectories - take two i_mutexes on non-directory objects
  *

diff --git a/fs/internal.h b/fs/internal.h
index bd3b2810a36b6030585e809d026e61a8ad25e760..377030a50aca6db105a0e8c909916f4775eb26cf 100644 (file)
--- a/fs/internal.h
+++ b/fs/internal.h
@@ -152,6 +152,8 @@ extern long prune_icache_sb(struct super_block *sb, struct shrink_control *sc);
 int dentry_needs_remove_privs(struct mnt_idmap *, struct dentry *dentry);
 bool in_group_or_capable(struct mnt_idmap *idmap,
                         const struct inode *inode, vfsgid_t vfsgid);
+void lock_two_inodes(struct inode *inode1, struct inode *inode2,
+                    unsigned subclass1, unsigned subclass2);
 
 /*
  * fs-writeback.c

diff --git a/fs/namei.c b/fs/namei.c
index e4fe0879ae5539d3e95dc3b7eec988b16db2a25f..148570aabe7422e03809350873106f77a0eda841 100644 (file)
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -3028,8 +3028,8 @@ static struct dentry *lock_two_directories(struct dentry *p1, struct dentry *p2)
                return p;
        }
 
-       inode_lock_nested(p1->d_inode, I_MUTEX_PARENT);
-       inode_lock_nested(p2->d_inode, I_MUTEX_PARENT2);
+       lock_two_inodes(p1->d_inode, p2->d_inode,
+                       I_MUTEX_PARENT, I_MUTEX_PARENT2);
        return NULL;
 }