[ovs-dev] [PATCH] conntrack: increment coverage counter for all bad checksum cases

[Tonghao Zhang]

On Mon, Mar 8, 2021 at 7:24 AM Paolo Valerio <pvalerio at redhat.com> wrote:
>
> conntrack_l4csum_err gets incremented only when corrupted icmp
> pass through conntrack.
> Increase it for the remaining bad checksum cases including when
> checksum is offloaded.
This patch looks good to me, and one question, do you have plan to
support l3 csum counters, such as "conntrack_l3csum_err".
Reviewed-by: Tonghao Zhang <xiangxia.m.yue at gmail.com>

> Signed-off-by: Paolo Valerio <pvalerio at redhat.com>
> ---
>  lib/conntrack.c |   15 ++++++++++++---
>  1 file changed, 12 insertions(+), 3 deletions(-)
>
> diff --git a/lib/conntrack.c b/lib/conntrack.c
> index 99198a601..30277f6fd 100644
> --- a/lib/conntrack.c
> +++ b/lib/conntrack.c
> @@ -1669,15 +1669,22 @@ static inline bool
>  checksum_valid(const struct conn_key *key, const void *data, size_t size,
>                 const void *l3)
>  {
> +    bool valid;
> +
>      if (key->dl_type == htons(ETH_TYPE_IP)) {
>          uint32_t csum = packet_csum_pseudoheader(l3);
> -        return csum_finish(csum_continue(csum, data, size)) == 0;
> +        valid = (csum_finish(csum_continue(csum, data, size)) == 0);
>      } else if (key->dl_type == htons(ETH_TYPE_IPV6)) {
> -        return packet_csum_upperlayer6(l3, data, key->nw_proto, size) == 0;
> +        valid = (packet_csum_upperlayer6(l3, data, key->nw_proto, size) == 0);
>      } else {
> +        valid = false;
> +    }
> +
> +    if (!valid) {
>          COVERAGE_INC(conntrack_l4csum_err);
> -        return false;
>      }
> +
> +    return valid;
>  }
>
>  static inline bool
> @@ -2076,6 +2083,8 @@ conn_key_extract(struct conntrack *ct, struct dp_packet *pkt, ovs_be16 dl_type,
>                  ctx->hash = conn_key_hash(&ctx->key, ct->hash_basis);
>                  return true;
>              }
> +        } else {
> +            COVERAGE_INC(conntrack_l4csum_err);
>          }
>      }
>
>
> _______________________________________________
> dev mailing list
> dev at openvswitch.org
> https://mail.openvswitch.org/mailman/listinfo/ovs-dev



-- 
Best regards, Tonghao