Hi, Javier and Thomas

> Yes, but you are assuming that Thomas' patch does successfully filter
> out the RA mismatch frames.
> Is it possible for you to paste the frame dumps of the duplicated
> frames?  Are they identical frames (same DA, SA, TA, RA and mesh
> seqnum)?  Do they have different TA?  Something else?
Ok, I have put some log information to check for the RA, TA, DA, SA
and also seqnum,

I have doing a Ping from 11th node to 1st node for about an hour. Ping
results are as follow:
355556 packets transmitted, 349013 received, 1% packet loss, time 4166648ms
rtt min/avg/max/mdev = 2.717/11.554/3160.335/35.995 ms, pipe 10

Here are amount of duplicate unicast frame detected.
Node 1: 0
Node 2: 0
Node 3: 651
Node 4: 0
Node 5: 1681
Node 6: 0
Node 7: 205
Node 8: 413
Node 9: 424
Node 10: 202
Node 11: 0

Node 3:
[ 3937.630000] mesh0: [1] ra dc:9f:db:00:31:90 ta 00:27:22:ec:fb:1c da
dc:9f:db:00:31:56 sa 00:27:22:ec:fb:1c seq -1426128896
[ 3937.630000] mesh0: [2] ra dc:9f:db:00:31:90 ta dc:9f:db:00:32:da da
dc:9f:db:00:31:56 sa 00:27:22:ec:fb:1c seq -1426128896

Node 5:
[ 6218.200000] mesh0: [1] ra dc:9f:db:00:32:da ta 00:27:22:ec:fb:1c da
dc:9f:db:00:31:56 sa 00:27:22:ec:fb:1c seq 1324745472
[ 6218.200000] mesh0: [2] ra dc:9f:db:00:32:da ta dc:9f:db:00:31:3e da
dc:9f:db:00:31:56 sa 00:27:22:ec:fb:1c seq 1324745472
[ 6752.030000] mesh0: [1] ra dc:9f:db:00:32:da ta 00:27:22:ec:fa:fa da
dc:9f:db:00:31:56 sa 00:27:22:ec:fb:1c seq 598475776
[ 6752.030000] mesh0: [2] ra dc:9f:db:00:32:da ta dc:9f:db:00:31:3e da
dc:9f:db:00:31:56 sa 00:27:22:ec:fb:1c seq 598475776

Node 9:
[ 5841.650000] mesh0: [1] ra 00:27:22:ec:fb:50 ta dc:9f:db:00:31:56 da
00:27:22:ec:fb:1c sa dc:9f:db:00:31:56 seq -395050240
[ 5841.650000] mesh0: [2] ra 00:27:22:ec:fb:50 ta 00:27:22:ec:fb:c9 da
00:27:22:ec:fb:1c sa dc:9f:db:00:31:56 seq -395050240
[ 5841.670000] mesh0: [1] ra 00:27:22:ec:fb:50 ta dc:9f:db:00:31:56 da
00:27:22:ec:fb:1c sa dc:9f:db:00:31:56 seq -378273024
[ 5841.670000] mesh0: [2] ra 00:27:22:ec:fb:50 ta 00:27:22:ec:fb:c9 da
00:27:22:ec:fb:1c sa dc:9f:db:00:31:56 seq -378273024

[1] is the non-duplicate unicast frame and [2] is the duplicate
unicast frame. Just show a few records since most of them have similar
pattern. SA and DA are both MAC addresses from the Ping nodes.

Based on the observation, the duplicate frame [2] has the same
sequence number, RA, DA and SA, but TA is different. So it should be
due to the path refresh since the detected duplicate frame is coming
from different TA.

---
Chun-Yeow

P.S: Change the topic subject since it may not due to the same problem
reported by MingAnn
_______________________________________________
Devel mailing list
[email protected]
http://lists.open80211s.org/cgi-bin/mailman/listinfo/devel

Reply via email to