[Devel] [PATCH rh7] fuse: allow to mount fuse filesystems inside CT

[Stanislav Kinsburskiy]

11.11.2015 15:40, Konstantin Khorenko пишет:
> On 11/10/2015 05:56 PM, Andrey Ryabinin wrote:
>> On 11/10/2015 05:52 PM, Stanislav Kinsburskiy wrote:
>>>>            goto err;
>>>> +    if (get_exec_env()->init_cred)
>>>> +        root_user_ns = get_exec_env()->init_cred->user_ns;
>>>> +    else
>>>> +        root_user_ns = &init_user_ns;
>>>> +
>>>
>>> We have some helper for this, if I'm not mistaken.
>>> Probably this one:
>>>
>>> current_user_ns_initial()
>>
>> current_user_ns_initial() check if current user_ns is initial for 
>> current ve.
>> We need to check if file's user_ns is initial for current ve, so it 
>> doesn't fit here.
>
> Stas, agree? Ack?
>

Sorry for the delay.
I definitely experience some problems with IMAP, because received this 
mails only an hour ago.
Basically yes, I have nothing against. And I was mistaken: we don't have 
such a helper.
But I would like to note, that from my POW it worth to create a helper 
similar to this:

+ struct user_namespace *ve_init_user_ns(void)
+ {

+	if (get_exec_env()->init_cred)
+		return get_exec_env()->init_cred->user_ns;
+	return &init_user_ns;

+}

And use it here and also in do_new_mount(), where we have the same chunk.
I expect, that we will have more places, where it could be useful.

>
>>>>        if ((file->f_op != &fuse_dev_operations) ||
>>>> -        (file->f_cred->user_ns != &init_user_ns))
>>>> +        (file->f_cred->user_ns != root_user_ns))
>>>>            goto err_fput;
>>>>          fc = kmalloc(sizeof(*fc), GFP_KERNEL);
>>>
>>