[c-nsp] CPU interrupt utilization

[shmapty at foureleven.org]

> The easiest way is to post the configuration and let people
> comment on features that cause packets to be punted on 
> that platform.

6509 sup2/MSFC2 running native mode 12.1(23)E2 (pk2o3sv-mz).  it has a 
single BGP session to our service provider, static routes internally, 
both static and dynamic NAT (inside/outside interfaces below)

we deployed new ingres and egress ACLs w/o any log keywords and the 
6500 still exhibits the problem.

we've got a similarly deployed sup1a/MSFC2 at another site.  both 
primarily host Internet-facing web services.  at their peak they both 
handle ~170Mbps/40Kpps.  the sup1a MSFC CPU is typically 25%/16%, the 
sup2 is 70%/50%

mls rp ip
mls flow ip destination
mls flow ipx destination
mls aclmerge algorithm odm
mls aclmerge odm optimizations
!
interface GigabitEthernet1/1
 ip address x.x.x.x 255.255.255.248
 ip access-group x-in-56 in
 ip access-group x-eg-01 out
 no ip redirects
 no ip unreachables
 no ip proxy-arp
 ip nat outside
 no ip mroute-cache
 keepalive 3
 flowcontrol send off
 tcam priority high
 no cdp enable
!
interface Vlan100
 ip address x.x.x.x 255.255.255.0
 no ip redirects
 no ip proxy-arp
 ip nat inside
 no ip mroute-cache
 mls rp ip
!