[Trac #968] Mesh parameters need to be tuned

[Zarro Boogs per Child]

#968: Mesh parameters need to be tuned
-------------------------+--------------------------------------------------
 Reporter:  jcardona     |        Owner:  jcardona 
     Type:  enhancement  |       Status:  assigned 
 Priority:  low          |    Milestone:  Untriaged
Component:  wireless     |   Resolution:           
 Keywords:               |  
-------------------------+--------------------------------------------------
Changes (by jcardona):

 * cc: mbletsas at laptop.org, rchoki at marvell.com (added)
  * status:  new => assigned

Comment:

 > [mbletsas at laptop.org said]
 >
 > I think that "loosing a few ICMP packets" during an initial ping
 exchange
 > (when there is no route for the destination) bug is back.
 > Subsequent sessions seem to complete without any loss

 I'm not sure about this.  The bug that you mention was caused by not
 adding unresolved frames to the pending queue.  The symptoms were that we
 would *always* lose the first packet or two until the route was
 established.

 And we do not see this now:

 {{{
 # iwpriv msh0 fwt_reset; iwpriv msh1 fwt_reset; ping -I msh0 192.168.2.81
 PING 192.168.2.81 (192.168.2.81) from 192.168.2.80 msh0: 56(84) bytes of
 data.
 64 bytes from 192.168.2.81: icmp_seq=1 ttl=64 time=45.6 ms
 64 bytes from 192.168.2.81: icmp_seq=2 ttl=64 time=3.92 ms
 64 bytes from 192.168.2.81: icmp_seq=3 ttl=64 time=2.66 ms
 }}}

 Packets in the pending queue expire, currently this time is set to ~2
 seconds.
 So if the route discovery takes longer than that, those initial packets
 will be dropped.  On the OLPC-testbed, the time to discover a route can be
 on that order.

 The tests below show one case when the route discover happened before the
 pending queue expiration, and one in which it did not.  Note that in the
 successful test, the rtt is right around the pending queue expiration
 time.

 {{{
 -bash-3.1# iwpriv msh0 fwt_reset; ping mshy5
 PING mshy5 (10.2.51.50) 56(84) bytes of data.
 64 bytes from mshy5 (10.2.51.50): icmp_seq=2 ttl=64 time=14.1 ms
 ...
 }}}
 {{{
 -bash-3.1# iwpriv msh0 fwt_reset; ping mshy5
 PING mshy5 (10.2.51.50) 56(84) bytes of data.
 64 bytes from mshy5 (10.2.51.50): icmp_seq=1 ttl=64 time=2007 ms
 ...
 }}}

-- 
Ticket URL: <http://dev.laptop.org/ticket/968#comment:1>
One Laptop Per Child <http://laptop.org/>