While 485dbd1fda (from #12175) was right
in assuming that the for most ICMPv6 error messages the originating
packet's destination address must not be a multicast, this is not the
case for _all_ ICMPv6 error messages (see [RFC 4443], section 2.4(e.3)).
Additionally, 485dbd1fda removed the
check for the source address ([RFC 4443], section 2.4(e.6)), which this
PR re-adds.
[RFC 4443]: https://tools.ietf.org/html/rfc4443#section-2.4
Rather than dispatching the packet automatically once it is complete,
`gnrc_sixlowpan_frag_rb_add()` now only returns success, and leaves it
to the caller to dispatch the packet.
While it is correct to not use an invalid address as a source address,
it is incorrect to assume that addresses not assigned to the interface
(`idx == -1` in the respective piece of code) are invalid: Other than
classic forwarding via a FIB, forwarded packets utilizing a IPv6
routing header will pass this check, like any other packet sent by this
node. The source address for these is not on the given node, so e.g.
source routing is not possible at the moment.
The IPv6 (extension) headers of the first fragment received are re-used
for the reassembled packet, so when receiving a subsequent packet we
need to distinguish, if we just want to release the payload or all of
the packet after the packet data was added to the reassembly buffer.
Due to some changes to the minimal forwarding draft and in preparation
for Selective Fragment Recovery some changes to the VRB API were
needed. Now the index of a VRB entry is only (L2 src, tag) not as
before (L2 src, L2 dst, length, tag).
I know that the current `rbuf_base` causes waste, as all the fields not
used by the new index are effectively not used by the VRB. I'd like to
fix that however in a later change, since that also requires some
modifications of the classic reassembly buffer, and thus would
complicate the review and testing of the change.
Sources for the index change:
- https://tools.ietf.org/html/draft-ietf-6lo-minimal-fragment-04#section-1
- https://mailarchive.ietf.org/arch/browse/6lo/?gbt=1&index=DLCTxC2X4bRNtYPHhtEkavMWlz4
before this commit the src address was checked for multicast, but the dst address should be checked. Therefore udp multicast packets would be flooded back to the src as ICMPv6 error, as not all nodes had a UDP receiver registered.
If an address was pre-configured by the upper layer its validity is
currently ignored. It is neither checked if the address is on the
interface at all nor is it checked if it is valid.
This change provides a fix for that by checking both facts.
When reworking the reception of IPv6 packets I reset a previously set
`ipv6` snip as follows when the IPv6 extension handler returns a
packet (see first hunk of this commit):
```C
ipv6 = pkt->next->next
```
With `gnrc_ipv6_ext` this makes *somewhat* sense, `pkt->next` was
previously equal to `ipv6` and after the function call `pkt->next`
is the marked extension header, while `pkt->next->next` is the IPv6
header. However, since `ipv6` is already write-protected i.e.
`ipv6->users == 1` (see ll. 665-675), any additional call of
`gnrc_pktbuf_start_write()` [won't][start-write-doc] duplicate the
packet. In fact, the only `gnrc_pktbuf_start_write()` in
`gnrc_ipv6_ext` is used to send the *result* to the subscribers of that
extension header type, leaving the original packet unchanged for the
caller. As such `ipv6` remains the pointer to the IPv6 header whether
we set it in the line above or not. So we actually don't need that
line.
However, the extension header handling also returns a packet when
`gnrc_ipv6_ext` is not compiled in. In that case it is just a dummy
define that returns the packet you give provide it which means that
this still holds true: `pkt->next == ipv6`.
So setting `ipv6` in this case is actually harmful, as `ipv6` now
points to the NETIF header [following the IPv6 header][pkt-structure]
in the packet and this causes the `user` counter of that NETIF header
`hdr` to be decremented if `hdr->users > 1` in the write-protection I
removed in hunk 2 of this commit:
```C
/* pkt might not be writable yet, if header was given above */
ipv6 = gnrc_pktbuf_start_write(ipv6);
if (ipv6 == NULL) {
DEBUG("ipv6: unable to get write access to packet: dropping it\n");
gnrc_pktbuf_release(pkt);
return;
}
```
But as we already established, `ipv6->users` is already 1, so we don't
actually need the write protection here either.
Since the packet stays unchanged after the `ipv6` snip, we also don't
need to re-search for `netif_hdr` after the other two lines are
removed.
[start-write-doc]: https://doc.riot-os.org/group__net__gnrc__pktbuf.html#ga640418467294ae3d408c109ab27bd617
[pkt-structure]: https://doc.riot-os.org/group__net__gnrc__pkt.html#ga278e783e56a5ee6f1bd7b81077ed82a7
