[apparmor] [PATCH 02/24] apparmor: convert profile lists to RCU based locking

Seth Arnold seth.arnold at canonical.com
Sat Mar 2 01:57:02 UTC 2013


> @@ -546,17 +534,40 @@ static void __ns_list_release(struct list_head *head);
>   */
>  static void destroy_namespace(struct aa_namespace *ns)
>  {
> +	struct aa_profile *unconfined;
> +
>  	if (!ns)
>  		return;
>  
> -	write_lock(&ns->lock);
> +	mutex_lock(&ns->lock);
>  	/* release all profiles in this namespace */
>  	__profile_list_release(&ns->base.profiles);
>  
>  	/* release all sub namespaces */
>  	__ns_list_release(&ns->sub_ns);
>  
> -	write_unlock(&ns->lock);
> +	unconfined = ns->unconfined;
> +	/*
> +	 * break the ns, unconfined profile cyclic reference and forward
> +	 * all new unconfined profiles requests to the parent namespace
> +	 * This will result in all confined tasks that have a profile
> +	 * being removed, inheriting the parent->unconfined profile.
> +	 */
> +	if (ns->parent)
> +		ns->unconfined = aa_get_profile(ns->parent->unconfined);

Incidentally, the comment in alloc_namespace() is incorrect now that
this code block have moved functions:

	/*
	 * released by free_namespace, however __remove_namespace breaks
	 * the cyclic references (ns->unconfined, and unconfined->ns) and
	 * replaces with refs to parent namespace unconfined
	 */


It might not actually matter much that the name has changed, since
__remove_namespace() calls destroy_namespace().. but it does bug me a
touch that the outer function has double-leading-underscore and the inner
function doesn't have the double-leading-underscore. Do the names need
some cleanup? (I'd love to see the split removed entirely, but removing
the root namespace is a special-case operation...)

Thanks
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 490 bytes
Desc: Digital signature
URL: <https://lists.ubuntu.com/archives/apparmor/attachments/20130301/468fa02a/attachment.pgp>


More information about the AppArmor mailing list