On 10/14, Paulo Alcantara wrote:
The crash occurred because we were calling memzero_explicit() on an
already freed sess_data::iov[1] (ntlmsspblob) in sess_free_buffer().
Fix this by not calling memzero_explicit() on sess_data::iov[1] as
it's already by handled by callers.
Fixes: d867d4ae29c7 ("cifs: replace kfree() with kfree_sensitive() for sensitive data")
Signed-off-by: Paulo Alcantara (SUSE) <pc@xxxxxx>
Good catch! Thanks.
Reviewed-by: Enzo Matsumiya <ematsumiya@xxxxxxx>
---
fs/cifs/sess.c | 16 +++++++++-------
1 file changed, 9 insertions(+), 7 deletions(-)
diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c
index f1c3c6d9146c..d33a27c1af4e 100644
--- a/fs/cifs/sess.c
+++ b/fs/cifs/sess.c
@@ -1213,16 +1213,18 @@ sess_alloc_buffer(struct sess_data *sess_data, int wct)
static void
sess_free_buffer(struct sess_data *sess_data)
{
- int i;
+ struct kvec *iov = sess_data->iov;
- /* zero the session data before freeing, as it might contain sensitive info (keys, etc) */
- for (i = 0; i < 3; i++)
- if (sess_data->iov[i].iov_base)
- memzero_explicit(sess_data->iov[i].iov_base, sess_data->iov[i].iov_len);
+ /*
+ * Zero the session data before freeing, as it might contain sensitive info (keys, etc).
+ * Note that iov[1] is already freed by caller.
+ */
+ if (sess_data->buf0_type != CIFS_NO_BUFFER && iov[0].iov_base)
+ memzero_explicit(iov[0].iov_base, iov[0].iov_len);
- free_rsp_buf(sess_data->buf0_type, sess_data->iov[0].iov_base);
+ free_rsp_buf(sess_data->buf0_type, iov[0].iov_base);
sess_data->buf0_type = CIFS_NO_BUFFER;
- kfree(sess_data->iov[2].iov_base);
+ kfree_sensitive(iov[2].iov_base);
}
static int
--
2.37.3
