rmdir(2) should return ENOTEMPTY

Under Solaris the behavior for rmdir(2) is to return EEXIST when
a directory still contains entries.  However, on Linux ENOTEMPTY
is the expected return value with EEXIST being technically allowed.
According to rmdir(2):

ENOTEMPTY
   pathname contains entries other than . and .. ; or, pathname has
   ..  as its final component.  POSIX.1-2001 also allows EEXIST for
   this condition.

Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #895

diff --git a/module/zfs/zfs_dir.c b/module/zfs/zfs_dir.c
index 6cd9c8508a1ec0e1f29c4a8215c65c6fb2736de4..670e31398d80d8acfdfc0eb02ab0890073746c70 100644 (file)
--- a/module/zfs/zfs_dir.c
+++ b/module/zfs/zfs_dir.c
@@ -837,8 +837,8 @@ zfs_dropname(zfs_dirlock_t *dl, znode_t *zp, znode_t *dzp, dmu_tx_t *tx,
 }
 
 /*
- * Unlink zp from dl, and mark zp for deletion if this was the last link.
- * Can fail if zp is a mount point (EBUSY) or a non-empty directory (EEXIST).
+ * Unlink zp from dl, and mark zp for deletion if this was the last link. Can
+ * fail if zp is a mount point (EBUSY) or a non-empty directory (ENOTEMPTY).
  * If 'unlinkedp' is NULL, we put unlinked znodes on the unlinked list.
  * If it's non-NULL, we use it to indicate whether the znode needs deletion,
  * and it's the caller's job to do it.
@@ -865,7 +865,7 @@ zfs_link_destroy(zfs_dirlock_t *dl, znode_t *zp, dmu_tx_t *tx, int flag,
 
                if (zp_is_dir && !zfs_dirempty(zp)) {
                        mutex_exit(&zp->z_lock);
-                       return (EEXIST);
+                       return (ENOTEMPTY);
                }
 
                /*