Here are a couple of fixes for autofs.
The first is fixing an incorrect log message.
The second is a bit more interesting.
When people want to use a global instance of automount in a container
they bind mount in the container as a propagation slave mount so the
automounted mounts propagate to the container (actually really only
useful for indirect mounts). But there are other cases, for example,
using ushare(1) results in propagation private mount present in the
cloned file system mounts. I'm not sure about excluding such mounts
in these cases as that would prevent a "mount --make-slave" or the like
which would reduce flexability or possibly cause a regression. So I've
elected to simply check and return a permission denied error. Note some
action is needed becuase if the kernel sends a mount request the daemon
will mount it in the init mount namespace which we also don't want in
this case.

Ian Kent (2):
  autofs: fix per-dentry timeout warning
  autofs: dont trigger mount if it cant succeed

 fs/autofs/autofs_i.h  |  4 ++++
 fs/autofs/dev-ioctl.c | 22 ++++++++++++----------
 fs/autofs/inode.c     |  1 +
 fs/autofs/root.c      |  8 ++++++++
 fs/namespace.c        |  6 ++++++
 include/linux/fs.h    |  1 +
 6 files changed, 32 insertions(+), 10 deletions(-)

-- 
2.51.1