On 2/5/16, 11:27 AM, Robert Shearman wrote:

Having MPLS packet stats is useful for observing network operation and
for diagnosing network problems. In the absence of anything better,
use RFCs for MIBs defining MPLS stats for guidance on the semantics of
the stats to expose. RFC3813 details two per-interface packet stats
that should be provided (label lookup failures and fragmented packets)
and also provides interpretation of RFC2863 for other per-interface
stats (in/out ucast, mcast and bcast, in/out discards and errors and
in unknown protos).

Multicast, fragment and broadcast packet counters are printed, but not
stored to allow for future implementation of current standards or
future standards without user-space having to change.

All the introduced fields are 64-bit, even error ones, to ensure no
overflow with long uptimes. Per-CPU counters are used to avoid
cache-line contention on the commonly used fields. The other fields
have also been made per-CPU for code to avoid performance problems in
error conditions on the assumption that on some platforms the cost of
atomic operations could be more pexpensive than sending the packet
(which is what would be done in the success case). If that's not the
case, we could instead not use per-CPU counters for these fields.

The IPv6 proc code was used as an inspiration for the proc code
here, both in terms of the implementation as well as the location of
the per-device stats proc files: /proc/net/dev_snmp_mpls/<dev>.

Signed-off-by: Robert Shearman <redacted>

Robert, any interest in moving this to the new stats api ?.

I had done some work for AF_<family> stats. Did not eventually end up including it in the
final version. The AF_<family> infra patch is here:
https://github.com/CumulusNetworks/net-next/commits/mpls-stats

Thanks!.