Re: CFT: vr(4)

At 12:34 AM 2/26/2008, Pyun YongHyeon wrote:
>  >
>
>Thanks again for your testing!
>An unresolved issue for Milan Obuch's router board still prevent me
>from committing the overhauled one. Unfortunately I still have no
>clue why his hardware does not work even though it has the same
>hardware revision(0x96).

I found another bug that your version of the driver fixes.  Its 
fairly easy for me to reproduce now and other people seem to run into 
it as well.

I hooked up 2 boxes with the NICs to a cisco switch with the 2 
interfaces on the same vlan. I then start a continuous ping either 
from the other box, or on the box itself.

boxA# ping boxB
PING 192.168.7.23 (192.168.7.23): 56 data bytes
64 bytes from 192.168.7.23: icmp_seq=0 ttl=64 time=0.974 ms
64 bytes from 192.168.7.23: icmp_seq=1 ttl=64 time=0.420 ms
64 bytes from 192.168.7.23: icmp_seq=2 ttl=64 time=0.427 ms
64 bytes from 192.168.7.23: icmp_seq=3 ttl=64 time=0.416 ms
64 bytes from 192.168.7.23: icmp_seq=4 ttl=64 time=0.455 ms
64 bytes from 192.168.7.23: icmp_seq=5 ttl=64 time=0.405 ms
64 bytes from 192.168.7.23: icmp_seq=6 ttl=64 time=0.423 ms
64 bytes from 192.168.7.23: icmp_seq=7 ttl=64 time=0.422 ms


^C
--- 192.168.7.23 ping statistics ---
66 packets transmitted, 20 packets received, 69.7% packet loss
round-trip min/avg/max/stddev = 0.396/0.476/0.974/0.132 ms


While in the middle of the ping, I change the media speed of the port 
of the box that is doing the pinging to 10 and then back to auto
boxA# ifconfig vr2
vr2: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
         options=b<RXCSUM,TXCSUM,VLAN_MTU>
         ether 00:00:24:c9:34:8a
         inet 192.168.7.21 netmask 0xffffff00 broadcast 192.168.7.255
         media: Ethernet autoselect (100baseTX <full-duplex>)
         status: active

The nic is now wedged on boxA. I then have to do a ifconfig vr2 down 
and ifconfig vr2 up to un wedge it, and in the logs I see the forced reset

vr2: link state changed to DOWN
vr2: link state changed to UP
vr2: link state changed to DOWN
vr2: link state changed to UP
vr2: Using force reset command.

..... HOWEVER, using the updated drivers on your web page, the box 
survives this exercise.

There is the occasional "shutdown error", but I guess thats to be 
expected as the physical layer is bouncing up and down.  But the main 
thing is that the box recovers on its own.

vr2: link state changed to DOWN
vr2: link state changed to UP
vr2: link state changed to DOWN
vr2: link state changed to UP
vr2: link state changed to DOWN
vr2: link state changed to UP
vr2: vr_link_task: Tx/Rx shutdown error -- resetting
vr2: link state changed to DOWN
vr2: restarting
vr2: vr_stop: Rx shutdown error
vr2: Using force reset command.
vr2: link state changed to UP

         ---Mike