[net2-wg] results from Tmote testbed

[Rodrigo Fonseca]

Kyle,

I tested this for a bit, and it was pretty useless otherwise: with
only one buffer, many messages were not boing logged. The way it is
right now seems to be working, as I haven't missed a single uart debug
message submitted to the component.

I would leave it as is, and check the last (new) field of the debug
messages: for each node it should never skip. If it doesn't skip a
number, there is no problem.

Rodrigo


On 6/16/06, Kyle Jamieson <jamieson at csail.mit.edu> wrote:
> Actually, I'd like to propose not using the MessagePool in
> UartDebugSenderP.  If there are any bugs in the MessagePool or its use
> in the FE, this would complicate tracking them down.  Instead I'd like
> to have UartDebugSenderP just be a simple SerialAMSender.
>
> Would this be okay with everyone?  If we agree, I can make the changes.
>
> Kyle
>
> Kyle Jamieson wrote:
> >> A message pool entry report means that there are no free message
> >> buffers with which to forward. This can be due to 1 of 2 things:
> >>
> >> 1) The send queue is actually full (there is no route, incoming rate
> >> is higher than outgoing rate).
> >> 2) There is a memory leak.
> >
> > I've just instrumented the forwarding engine to tell us when calls to
> > either of the Pools' get or put methods fail.
> >
> > Kyle
> > _______________________________________________
> > net2-wg mailing list
> > net2-wg at millennium.berkeley.edu
> > https://mail.millennium.berkeley.edu/cgi-bin/mailman/listinfo/net2-wg
> _______________________________________________
> net2-wg mailing list
> net2-wg at millennium.berkeley.edu
> https://mail.millennium.berkeley.edu/cgi-bin/mailman/listinfo/net2-wg
>