On 2015/1/13 13:28, Al Viro wrote:
> On Tue, Jan 13, 2015 at 01:08:06PM +0800, Jiang Liu wrote:
>> Simplify function commit_tree() in fs/namespace.c, it also improves code
>> readability.
>
> Huh?
>
>> static void commit_tree(struct mount *mnt, struct mount *shadows)
>> {
>> struct mount *parent = mnt->mnt_parent;
>> - struct mount *m;
>> - LIST_HEAD(head);
>> struct mnt_namespace *n = parent->mnt_ns;
>>
>> BUG_ON(parent == mnt);
>> -
>> - list_add_tail(&head, &mnt->mnt_list);
>> - list_for_each_entry(m, &head, mnt_list)
>> - m->mnt_ns = n;
>> -
>> - list_splice(&head, n->list.prev);
>> -
>> + mnt->mnt_ns = n;
>> + list_add_tail(&mnt->mnt_list, &n->list);
>
> And who will set ->mnt_ns on other mounts reachable via ->mnt_list from mnt?
> Not to mention putting them on n->list...
>
> What you seem to have missed is that list_add_tail(&head, &mnt->mnt_list)
> is not the same thing as list_add_tail(&mnt->mnt_list, &head). We are
> not ripping mnt->mnt_list from whatever list it was on and insert it into
> a previously empty one in head. We take an unanchored cyclic list (all
> elements being some_mnt->mnt_list) and anchor it at head, then go through
> all elements of the original list and move them all at once into the
> end of n->list.
Hi Al,
Sorry for the noise, didn't realized that mnt->mnt_list anchors
a whole list:(
Regards!
Gerry
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
