net: Check for the minimum IP fragmented datagram size

Nicolas Bidron and Nicolas Guigo reported the two bugs below:

"
----------BUG 1----------

In compiled versions of U-Boot that define CONFIG_IP_DEFRAG, a value of
`ip->ip_len` (IP packet header's Total Length) higher than `IP_HDR_SIZE`
and strictly lower than `IP_HDR_SIZE+8` will lead to a value for `len`
comprised between `0` and `7`. This will ultimately result in a
truncated division by `8` resulting value of `0` forcing the hole
metadata and fragment to point to the same location. The subsequent
memcopy will overwrite the hole metadata with the fragment data. Through
a second fragment, this can be exploited to write to an arbitrary offset
controlled by that overwritten hole metadata value.

This bug is only exploitable locally as it requires crafting two packets
the first of which would most likely be dropped through routing due to
its unexpectedly low Total Length. However, this bug can potentially be
exploited to root linux based embedded devices locally.

```C
static struct ip_udp_hdr *__net_defragment(struct ip_udp_hdr *ip, int *lenp)
{
     static uchar pkt_buff[IP_PKTSIZE] __aligned(PKTALIGN);
     static u16 first_hole, total_len;
     struct hole *payload, *thisfrag, *h, *newh;
     struct ip_udp_hdr *localip = (struct ip_udp_hdr *)pkt_buff;
     uchar *indata = (uchar *)ip;
     int offset8, start, len, done = 0;
     u16 ip_off = ntohs(ip->ip_off);

     /* payload starts after IP header, this fragment is in there */
     payload = (struct hole *)(pkt_buff + IP_HDR_SIZE);
     offset8 =  (ip_off & IP_OFFS);
     thisfrag = payload + offset8;
     start = offset8 * 8;
     len = ntohs(ip->ip_len) - IP_HDR_SIZE;
```

The last line of the previous excerpt from `u-boot/net/net.c` shows how
the attacker can control the value of `len` to be strictly lower than
`8` by issuing a packet with `ip_len` between `21` and `27`
(`IP_HDR_SIZE` has a value of `20`).

Also note that `offset8` here is `0` which leads to `thisfrag = payload`.

```C
     } else if (h >= thisfrag) {
         /* overlaps with initial part of the hole: move this hole */
         newh = thisfrag + (len / 8);
         *newh = *h;
         h = newh;
         if (h->next_hole)
             payload[h->next_hole].prev_hole = (h - payload);
         if (h->prev_hole)
             payload[h->prev_hole].next_hole = (h - payload);
         else
             first_hole = (h - payload);

     } else {
```

Lower down the same function, execution reaches the above code path.
Here, `len / 8` evaluates to `0` leading to `newh = thisfrag`. Also note
that `first_hole` here is `0` since `h` and `payload` point to the same
location.

```C
     /* finally copy this fragment and possibly return whole packet */
     memcpy((uchar *)thisfrag, indata + IP_HDR_SIZE, len);
```

Finally, in the above excerpt the `memcpy` overwrites the hole metadata
since `thisfrag` and `h` both point to the same location. The hole
metadata is effectively overwritten with arbitrary data from the
fragmented IP packet data. If `len` was crafted to be `6`, `last_byte`,
`next_hole`, and `prev_hole` of the `first_hole` can be controlled by
the attacker.

Finally the arbitrary offset write occurs through a second fragment that
only needs to be crafted to write data in the hole pointed to by the
previously controlled hole metadata (`next_hole`) from the first packet.

 ### Recommendation

Handle cases where `len` is strictly lower than 8 by preventing the
overwrite of the hole metadata during the memcpy of the fragment. This
could be achieved by either:
* Moving the location where the hole metadata is stored when `len` is
lower than `8`.
* Or outright rejecting fragmented IP datagram with a Total Length
(`ip_len`) lower than 28 bytes which is the minimum valid fragmented IP
datagram size (as defined as the minimum fragment of 8 octets in the IP
Specification Document:
[RFC791](https://datatracker.ietf.org/doc/html/rfc791) page 25).

----------BUG 2----------

In compiled versions of U-Boot that define CONFIG_IP_DEFRAG, a value of
`ip->ip_len` (IP packet header's Total Length) lower than `IP_HDR_SIZE`
will lead to a negative value for `len` which will ultimately result in
a buffer overflow during the subsequent `memcpy` that uses `len` as it's
`count` parameter.

This bug is only exploitable on local ethernet as it requires crafting
an invalid packet to include an unexpected `ip_len` value in the IP UDP
header that's lower than the minimum accepted Total Length of a packet
(21 as defined in the IP Specification Document:
[RFC791](https://datatracker.ietf.org/doc/html/rfc791)). Such packet
would in all likelihood be dropped while being routed to its final
destination through most routing equipment and as such requires the
attacker to be in a local position in order to be exploited.

```C
static struct ip_udp_hdr *__net_defragment(struct ip_udp_hdr *ip, int *lenp)
{
     static uchar pkt_buff[IP_PKTSIZE] __aligned(PKTALIGN);
     static u16 first_hole, total_len;
     struct hole *payload, *thisfrag, *h, *newh;
     struct ip_udp_hdr *localip = (struct ip_udp_hdr *)pkt_buff;
     uchar *indata = (uchar *)ip;
     int offset8, start, len, done = 0;
     u16 ip_off = ntohs(ip->ip_off);

     /* payload starts after IP header, this fragment is in there */
     payload = (struct hole *)(pkt_buff + IP_HDR_SIZE);
     offset8 =  (ip_off & IP_OFFS);
     thisfrag = payload + offset8;
     start = offset8 * 8;
     len = ntohs(ip->ip_len) - IP_HDR_SIZE;
```

The last line of the previous excerpt from `u-boot/net/net.c` shows
where the underflow to a negative `len` value occurs if `ip_len` is set
to a value strictly lower than 20 (`IP_HDR_SIZE` being 20). Also note
that in the above excerpt the `pkt_buff` buffer has a size of
`CONFIG_NET_MAXDEFRAG` which defaults to 16 KB but can range from 1KB to
64 KB depending on configurations.

```C
     /* finally copy this fragment and possibly return whole packet */
     memcpy((uchar *)thisfrag, indata + IP_HDR_SIZE, len);
```

In the above excerpt the `memcpy` overflows the destination by
attempting to make a copy of nearly 4 gigabytes in a buffer that's
designed to hold `CONFIG_NET_MAXDEFRAG` bytes at most which leads to a DoS.

 ### Recommendation

Stop processing of the packet if `ip_len` is lower than 21 (as defined
by the minimum length of a data carrying datagram in the IP
Specification Document:
[RFC791](https://datatracker.ietf.org/doc/html/rfc791) page 34)."

Add a check for ip_len lesser than 28 and stop processing the packet
in this case.

Such a check covers the two reported bugs.

Reported-by: Nicolas Bidron <nicolas.bidron@nccgroup.com>
Signed-off-by: Fabio Estevam <festevam@denx.de>

1 files changed, 2 insertions, 0 deletions

diff --git a/include/net.h b/include/net.h
index 675bf4171b..e3889a0bc8 100644
--- a/include/net.h
+++ b/include/net.h
@@ -391,6 +391,8 @@ struct ip_hdr {
 
 #define IP_HDR_SIZE		(sizeof(struct ip_hdr))
 
+#define IP_MIN_FRAG_DATAGRAM_SIZE	(IP_HDR_SIZE + 8)
+
 /*
  *	Internet Protocol (IP) + UDP header.
  */