[PATCH 5.15.y] ovl: take a snapshot of the name for copy up

[Sasha Levin <sashal@kernel.org>]

From: Chris Mason <mason@kernel.org>

[ Upstream commit 65bf000962a33ec9d2488e4e1797b6cf630abcbb ]

The copy up of a directory reads a freed dentry name when a lookup
moves the directory at the same time.

A pointer to the name of a dentry is only valid while something stops
a move of the dentry. ovl_do_copy_up() keeps a plain copy of
c->dentry->d_name in c->destname, and ovl_copy_up_start() only keeps
ovl_rename() away. When a lookup finds the same directory under another
name after a change to the lower layer, d_splice_alias() moves the
dentry and frees its external name.

With a delay in ovl_do_copy_up(), a chmod of a lower-only directory with
a 60 byte name raced with a rename of that directory in the lower layer
and a stat of the new name, and KASAN reported a use-after-free in
lookup_noperm_common().

Fix this by using take_dentry_name_snapshot() for c->destname.

Fixes: 829c28be9bb9 ("ovl: use d_splice_alias() in place of d_add() in lookup")
Cc: stable@vger.kernel.org
Signed-off-by: Chris Mason <mason@kernel.org>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
[ sashal: adapted name snapshot handling to copy-up code with earlier destname initialization and no file-handle cleanup ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/overlayfs/copy_up.c | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

diff --git a/fs/overlayfs/copy_up.c b/fs/overlayfs/copy_up.c
index 4c7ef60dbaac0..71df585385f1e 100644
--- a/fs/overlayfs/copy_up.c
+++ b/fs/overlayfs/copy_up.c
@@ -791,6 +791,7 @@ static int ovl_do_copy_up(struct ovl_copy_up_ctx *c)
 {
 	int err;
 	struct ovl_fs *ofs = OVL_FS(c->dentry->d_sb);
+	struct name_snapshot name;
 	bool to_index = false;
 
 	/*
@@ -819,13 +820,23 @@ static int ovl_do_copy_up(struct ovl_copy_up_ctx *c)
 		/* Disconnected dentry must be copied up to index dir */
 		return -EIO;
 	} else {
+		/*
+		 * ovl_copy_up_start() keeps ovl_rename() away from c->dentry,
+		 * because ovl_rename() calls ovl_copy_up() on source and target
+		 * dentry.  It does not stop d_splice_alias() from moving a
+		 * directory that a lookup found under another name after a
+		 * change to the lower layer.  That move frees an external name,
+		 * so take a snapshot.
+		 */
+		take_dentry_name_snapshot(&name, c->dentry);
+		c->destname = name.name;
 		/*
 		 * Mark parent "impure" because it may now contain non-pure
 		 * upper
 		 */
 		err = ovl_set_impure(c->parent, c->destdir);
 		if (err)
-			return err;
+			goto out;
 	}
 
 	/* Should we copyup with O_TMPFILE or with workdir? */
@@ -857,6 +868,8 @@ static int ovl_do_copy_up(struct ovl_copy_up_ctx *c)
 out:
 	if (to_index)
 		kfree(c->destname.name);
+	else
+		release_dentry_name_snapshot(&name);
 	return err;
 }
 
@@ -982,7 +995,6 @@ static int ovl_copy_up_one(struct dentry *parent, struct dentry *dentry,
 	if (parent) {
 		ovl_path_upper(parent, &parentpath);
 		ctx.destdir = parentpath.dentry;
-		ctx.destname = dentry->d_name;
 
 		err = vfs_getattr(&parentpath, &ctx.pstat,
 				  STATX_ATIME | STATX_MTIME,
-- 
2.53.0