lonworks: MIP DPS lockup (MIP from echelon)

[lroy@xxxxxxxxxxx (lroy)]

I am looking for information on the MIP/DPS from echelon to fix the
problem described below. In particular,  a web page on the MIP: "MIP
Link Layer Timing" by Darron Fick.

We have built a driver for a power PC to interface to a MIP. After as
little as 1h30, the neuron stops transmitting. The length of time it
lasts is proportional to application/network buffers. The number of
messages that does not get back completion events seems to be
equivalent to the number of application/network buffers. Normally, one
message every 50ms is sent to the MIP. However, for the test reproduce
the problem faster, the messages are sent has fast as the driver/MIP
can take it (8 to 10 ms). The MIP does not receive any messages. The
neuron is only plugged with a lonworks protocol analyzer. There are
two threads in the driver one that reads and one that writes, however
reads and writes are mutually exclusive. After the MIP stops
transmitting, the driver seems to be able to talk to the driver and
nodeutil can access the chip (for this particular test).

If you have had a similar problem, it would be appreciated if you
could tell me what you have done to fix. If you have any suggestion on
what should be looked into, it will also be appreciated.

Thank you in advance

lroy@xxxxxxxxxxx

Related subjects: MIP, DPS, MIP/DPS, echelon, lonworks, LON.