Created attachment 349301 [details]
suggested patch

Description of Problem:

 It is assumed that the router with two interfaces (IF0 and IF1) exists.
 When router's IF0 received a packet, and forwarding this packet with IF1,
 the router will check the dst address. If the dst address is unreachable,
 router will send a DestUnreachs packet back to sender using the IF0.
 But the IF0's counter of OutDestUnreachs can not be increased,
 the IF1's counter of OutDestUnreachs will be increased instead.

 The counter of Icmp6OutPktTooBigs and Icmp6OutTimeExcds have the same problem.

Version-Release number of selected component:

 Red Hat Enterprise Linux Version Number: RHEL4
 Release Number:                          4.8
 Architecture:                            x86_64
 Kernel Version:                          kernel-2.6.9-89.EL
 Related Package Version:                 none
 Related Middleware / Application:        none

Drivers or hardware or architecture dependency:
 none

How reproducible:
 every time

Step to Reproduce:
 The topology as follow for example:

  IF0: eth0
  IF1: eth0.100
                               -------|------------------2002::/64
                                      |[eth0.100]
                                      |fe80::2zz:zzff:fezz:zzzz/64
    -------                       --------
   |tester |                     | Router |(RHEL4.8)
    -------                       --------
       |[eth0]                        |[eth0]
       |3ffe:501:ffff:100::xxxx/64    |3ffe:501:ffff:100::yyyy/64
 ------|------------------------------|------------fe80::/64

 Step1
   Config the tester
   #ip -6 addr add 3ffe:501:ffff:100::xxxx/64 dev eth0
   #ip -6 route add 2002::/64 via 3ffe:501:ffff:100::yyyy dev eth0

 Step2
   Config the router
   #sysctl net.ipv6.conf.all.forwarding=1
   #ip -6 addr add 3ffe:501:ffff:100::yyyy/64 dev eth0
   #ip -6 route add  2002::/64 via fe80::2zz:zzff:fezz:zzzz dev eth0.100

 Step3
   router: record the value of Icmp6OutDestUnreachs of eth0 and eth0.100
   #cat /proc/net/dev_snmp6/eth0 | grep Icmp6OutDestUnreachs
   #cat /proc/net/dev_snmp6/eth0.100 | grep Icmp6OutDestUnreachs

 Step4
   host: send icmp packets to unknown dst addr of 2002::2.
   #ping6 -I eth0 -c 1 2002::2

   (note) the ping6 command is failed by address unreachable.

 Step5
   router: record the value of Icmp6OutDestUnreachs of eth0 and eth1
   #cat /proc/net/dev_snmp6/eth0 | grep Icmp6OutDestUnreachs
   #cat /proc/net/dev_snmp6/eth0.100 | grep Icmp6OutDestUnreachs

Actual Results:
 The eth0's counter of Icmp6OutDestUnreachs can not be increased,
 but the eth0.100's counter of Icmp6OutDestUnreachs can be increased 1 count instead.

Expected Results:
 The eth0's counter of Icmp6OutDestUnreachs can be increased 1 count,
 and the eth0.100's counter of Icmp6OutDestUnreachs can not be increased.

Hardware configuration:
 Model:        PRIMERGY TX150 S5
 CPU Info:     Intel(R) Xeon(R) CPU   3040  @ 1.86GHz
 Memory Info:  6GB

Business Impact:
  Target Release: 4.9
  Errata Request: No
  Hotfix Request: No

Additional Info:
 The sosreport file is attached:
   file: sosreport-ueki.001-617398-8adccb.tar.bz2
   md5sum: 3e52026401fa6a71d43dcb132e8adccb

 The patch file is attached:
   file: net-fix-counter-of-OutDestUnreachs.patch

   This problem has already been solved in latest community's kernel.