[c-nsp] Giants reported in 12.0(30)S1 , bug?

[David Freedman]

Hiya.


We're running 12.0(30)S1 on a GSR12012 box, using mostly 3GE-GBIC-SC 
(E2) linecards.

Ever since the upgrade to this release, the oversize MPLS frames that 
the ports have been receiving are being marked as giants.


#sh int g4/1 | in giant
      Received 0 broadcasts, 0 runts, 1955149209 giants, 0 throttles




Both this interface, and its partner have the same configuration,
being:


!
  tag-switching mtu 1526
  tag-switching ip
!


Plus any non-GSR boxes connected to this GSR with the same interface 
configuration (mpls mtu 1526) show no giants ever having been received.



As well, funnily enough, the controller statistics don't seem to think 
anything is wrong:

#execute-on slot 4 show controllers | in giant
========= Line Card (Slot 4) =========
    0 tgiant, 987193 mcast, 0 bcast, 0 tpause, 0 tisl, 0 tiq
    1368673326 r1023, 3266521914 r1518, 2947791614 r1548, 0 rgiant, 0 rcode
    giant_pkt_ctr        0x00000000 (0)
    0 tgiant, 2540991 mcast, 1 bcast, 0 tpause, 0 tisl, 0 tiq
    904669826 r1023, 1348098506 r1518, 11663679292 r1548, 0 rgiant, 0 rcode
    giant_pkt_ctr        0x00000000 (0)
    0 tgiant, 843625 mcast, 1 bcast, 0 tpause, 0 tisl, 0 tiq
    827632139 r1023, 1923703721 r1518, 0 r1548, 0 rgiant, 0 rcode
    giant_pkt_ctr        0x00000000 (0)



The data seems to pass through the box unharmed, no dropping occurs
when passing in a two label deep frame filled to maximum payload.


I'm sure this is just a cosmetic bug, but bugtool brings up nothing!


Is there anybody else here running the same (or near same) code release
and seeing the same symptoms?


Thanks in advance for any help,


Kind Regards,


-- 
David Freedman
Network Engineering Department
Claranet UK Limited
http://www.clara.net