Correct RTU aging bug
From Maciej email:
following the bug of GSI, I decided to make some testing of
the RTU (SW+GW) in a "larger" WR network (3 switches) ...
something actually never done before.
The results are not-that-optimistic
1. Aging of MAC entries by RTUd (SW bug)
It seems that RTUd does age entries but it takes more
then the claimed 20s... I measured with a stopper the aging
time and it's ~ 5min.... I remember having some problems
in implementing timers for the (crappa-code) PTPd. I would
guess that it might be the same problem
2. Port MACs (SW bug)
In the log of Dietrich I've noticed something strange, which is:
>adding static route for port wr8 index 8 [mac 00:00:00:00:00:00]
>adding static route for port wr6 index 6 [mac 00:00:00:00:00:00]
> [...]
WTF are MACs 00:00:00:00:00:00 ??? These are MAC entries signed
for each of WR ports.. this means that the RTUd is wrongly reading
the MAC addresses assigned to switch's ports and provides wrong
entries in RTU. This is not very critical for switch's normal functions
but
this is a clear bug.
3. Simple ping test
sometimes I'm not able to ping a PC through the WR network. Or the
ping
stops working after some time.... not sure yet what's wrong.