Re: [PATCH 1/2] autofs4: allow autofs to work outside the initial PID namespace

[Eric W. Biederman]

Miklos Szeredi <miklos@xxxxxxxxxx> writes:

> On Mon, Nov 26, 2012 at 3:29 AM, Ian Kent <raven@xxxxxxxxxx> wrote:
>>> > > MS_UNBINDABLE says:  skip this mount when copying a mount tree, such
>>> > > as when the mount namespace is cloned.
>>> > >
>>> > > If you set MS_UNBINDABLE on autofs mounts then they will simply not
>>> > > appear in a cloned namespace.  Which sounds like a good idea,  no?
>>> >
>>> > Good point.  If the desire is for a mount to be managed by autofs
>>> > setting MS_UNBINDABLE seems required.
>>>
>>> Arrgh, I know that's something I should have looked into long ago.
>>> The fact is that autofs mounts are directly related to a specific path
>>> defined by automount maps that are associated with the daemon so bind
>>> mounting them elsewhere makes no sense.
>>
>> Except, AFAICS, they do appear in the clone.
>
> Hmm, yes, apparently the semantics of MS_UNBINDABLE only apply to
> actual bind mounts not to namespace cloning. Even though the two
> operations are closely related.  Not sure why this is so, but it is
> probably not a good idea to change the semantics at this point.

And for whatever reason this appears deliberate.

CL_COPY_ALL in copy_tree allows the copy.

The selected semantics of namespace sharing tend to mystify me.

So I don't know how much MS_UNBINDABLE helps over MS_PRIVATE.  Both
prevent propogation of changes to other namespaces.  I don't know how
much using MS_UNBINDABLE to also prevent bind mounts helps.

Eric

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/