Re: [PATCH 2/3] net/unix: hook unix_socketpair() into LSM

[Paul Moore]

On Tue, Apr 24, 2018 at 1:56 PM, David Miller <davem@xxxxxxxxxxxxx> wrote:
> From: Paul Moore <paul@xxxxxxxxxxxxxx>
> Date: Tue, 24 Apr 2018 13:55:31 -0400
>
>> On Mon, Apr 23, 2018 at 9:30 AM, David Herrmann <dh.herrmann@xxxxxxxxx> wrote:
>>> Use the newly created LSM-hook for unix_socketpair(). The default hook
>>> return-value is 0, so behavior stays the same unless LSMs start using
>>> this hook.
>>>
>>> Signed-off-by: David Herrmann <dh.herrmann@xxxxxxxxx>
>>> ---
>>>  net/unix/af_unix.c | 5 +++++
>>>  1 file changed, 5 insertions(+)
>>>
>>> diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c
>>> index 68bb70a62afe..bc9705ace9b1 100644
>>> --- a/net/unix/af_unix.c
>>> +++ b/net/unix/af_unix.c
>>> @@ -1371,6 +1371,11 @@ static int unix_stream_connect(struct socket *sock, struct sockaddr *uaddr,
>>>  static int unix_socketpair(struct socket *socka, struct socket *sockb)
>>>  {
>>>         struct sock *ska = socka->sk, *skb = sockb->sk;
>>> +       int err;
>>> +
>>> +       err = security_unix_stream_socketpair(ska, skb);
>>> +       if (err)
>>> +               return err;
>>
>> I recognize that AF_UNIX is really the only protocol that supports
>> socketpair(2) at the moment, but I like to avoid protocol specific LSM
>> hooks whenever possible.  Unless someone can think of a good
>> objection, I would prefer to see the hook placed in __sys_socketpair()
>> instead (and obviously drop the "unix_stream" portion from the hook
>> name).
>
> The counterargument is that after 30 years no other protocol has grown
> usage of this operation. :-)

Call me a an optimist ;)

-- 
paul moore
www.paul-moore.com