[dpdk-dev] [PATCH v3 1/4] ip_frag: ensure minimum v4 fragmentation length

[Ananyev, Konstantin]

> 
> The IPv4 specification says that each fragment must at least the size of
> an IP header plus 8 octets.  When attempting to run ipfrag using a
> smaller size, the fragment library will return successful completion,
> even though it is a violation of RFC791 (and updates).
> 
> Signed-off-by: Aaron Conole <aconole at redhat.com>
> ---
>  lib/librte_ip_frag/rte_ipv4_fragmentation.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/lib/librte_ip_frag/rte_ipv4_fragmentation.c b/lib/librte_ip_frag/rte_ipv4_fragmentation.c
> index 9e9f986cc5..4baaf6355c 100644
> --- a/lib/librte_ip_frag/rte_ipv4_fragmentation.c
> +++ b/lib/librte_ip_frag/rte_ipv4_fragmentation.c
> @@ -76,6 +76,12 @@ rte_ipv4_fragment_packet(struct rte_mbuf *pkt_in,
>  	uint16_t fragment_offset, flag_offset, frag_size;
>  	uint16_t frag_bytes_remaining;
> 
> +	/*
> +	 * Ensure the IP fragmentation size is at least iphdr length + 8 octets
> +	 */
> +	if (unlikely(mtu_size < (sizeof(struct rte_ipv4_hdr) + 8*sizeof(char))))
> +		return -EINVAL;
> +

Same comment as for ipv6: ipv4 min MTU is 68B.
Why do we need extra checking here?

>  	/*
>  	 * Ensure the IP payload length of all fragments is aligned to a
>  	 * multiple of 8 bytes as per RFC791 section 2.3.
> --
> 2.25.1