On Sep 18, 2010, at 7:35 PM, Eric Decker wrote:
> I'm sure this is pretty old but ....
>
> I'm currently debugging a serial protocol problem. I'm sending a well formed
> Serial AM packet to my mote but for some reason it is getting
> dropped by dispatch. The AM type isn't being seen correctly (dispatch seems
> to be looking at the wrong byte). I've probably got something
> wired wrong.
>
>
> But in the course of looking at the Serial driver (SerialP.nc), I noticed the
> following:
>
>
> bool valid_rx_proto(uint8_t proto){
> switch (proto){
> case SERIAL_PROTO_PACKET_ACK:
> return TRUE;
> case SERIAL_PROTO_ACK:
> case SERIAL_PROTO_PACKET_NOACK:
> default:
> return FALSE;
> }
> }
>
>
>
> Is there anything that describes the protocol for serial acking in more
> detail?
>
> eric
TEP 113 is probably the best resource. There were a lot of discussions back in
2004-6, I think, particularly with respect to high throughput. Ben Greenstein
ran a lot of profiling tests to see where the bottlenecks where. Sometime in
2007? Razvan examined the issue of incoming packets being dropped due to high
interrupt load (for Deluge) and so introduced a bit of delay on serial tools to
give the mote side a bit of wiggle room. These discussions were all on the
mailing lists.
Phil
_______________________________________________
Tinyos-help mailing list
[email protected]
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
