Re: [PATCH 2/3] net: packet: fix information leak to userland

[walter harms]

Am 07.11.2010 13:06, schrieb Vasiliy Kulikov:
> On Sun, Nov 07, 2010 at 12:37 +0100, walter harms wrote:
>> Am 06.11.2010 15:39, schrieb Vasiliy Kulikov:
>>> On Mon, Nov 01, 2010 at 10:14 +0100, walter harms wrote:
>>>> Vasiliy Kulikov schrieb:
>>>>> @@ -1719,7 +1719,7 @@ static int packet_getname_spkt(struct socket *sock, struct sockaddr *uaddr,
>>>>>  	rcu_read_lock();
>>>>>  	dev = dev_get_by_index_rcu(sock_net(sk), pkt_sk(sk)->ifindex);
>>>>>  	if (dev)
>>>>> -		strlcpy(uaddr->sa_data, dev->name, 15);
>>>>> +		strncpy(uaddr->sa_data, dev->name, 14);
>>>>>  	else
>>>>>  		memset(uaddr->sa_data, 0, 14);
>>>>
>>>> if i understand the code correcly the max size for dev->name is IFNAMSIZ.
>>>
>>> For dev->name - IFNAMSIZ, for uaddr->sa_data - 14.
>>>
>>
>>
>> did not notice, since uaddr->sa_data should take dev->name this does no look very
>> clever. How is the size of  sa_data defined ?
> 
> Magic size...
> 
> ~/linux/include/linux/socket.h:
> 
> struct sockaddr {
> 	sa_family_t	sa_family;	/* address family, AF_xxx	*/
> 	char		sa_data[14];	/* 14 bytes of protocol address	*/
> };
> 
> 
>> Would it hurt when some uses IFNAMSIZ here ?
> 

so i should be more direct. the idea was :
char		sa_data[IFNAMSIZ];



> If copy _to_ sa_data string of maximum IFNAMSIZ bytes - yes.
> 
> 
> In packet_getname_spkt() the output buffer is 128 bytes, so it doesn't
> really overflows anything.  I don't think that *_getname() implementations
> don't know this.
> 
>> Perhaps someone who know more about the network stack can figure out what is actualy done
>> with uaddr->sa_data.
> 
> Yeah, also I wonder whether it is always NULL-terminated string or not.
> 
>> looks like a can of worms.
>>
>>
>> In packet_bind_spkt() they will copy a char[15], obviously it is a real problem.
> 
> No, packet_bind_spkt() copies 14-byte string into array of 15 bytes.
> The vice versa would be a bug.
> 

ups your are right, wrong way around. it still does not look clever.
I have the feeling that the basic idea what to store the string with out \0.

according to this:
http://www.gnu.org/s/libc/manual/html_node/Address-Formats.html

It should work.

re,
 wh
--
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/