Hi om_p, I have the same doubt, so I set the basestation ID to -1 for trasparent bridging (could I send a custumized ACK??) and ID>= 0 for packets generator, and I have added a check in Send.send (CC2420CsmaP.nc, if( call AMPacket.address() != -1) header->src = call AMPacket.address(); )
I agree with you, but I think that the instruction (line 274) in BaseStationP is misleading. Thanks, Salvo > On Thu, Mar 6, 2008 at 4:15 PM, Salvo Fiduccia <[EMAIL PROTECTED]> wrote: > > Hi all, > > > > in SerialActiveMessageP there is a note in the AMSend.send command: > > "Do not set the source address or group, as doing so > > prevents transparent bridging" > > > > I think this is correct, but the same don't happen in the send command of > > ActiveMessageC: in the micaz implementation, in CC2420csmaP, the Send.send > > command sets the header source field with the node id, so these istructions > > (BaseStationP) don't have effect: > > > > task void radioSendTask() { > > ............. > > 273 call RadioPacket.clear(msg); > > 274 call RadioAMPacket.setSource(msg, source); > > 275 > > 276 if (call RadioSend.send[id](addr, msg, len).... > > .... > > } > > > > I am not sure if there is a straight forward resolution to this topic. > The question is if the BaseStation application should provide for > transparent serial to radio bridging. Some people use a BaseStation > and some PC application to generate packets. One example is how Tenet > uses CTP - a PC software generates a CTP beacon, encapsulates it as a > serial packet and sends it to the base station. I believe there were > some protocols like that in the core distribution of TinyOS 1.x as > well. Our experience is if the bridging is not transparent, then it is > easy to make an error - program the BaseStation as node y and have the > PC software generate packet as node x and you have a problem when the > children of the root start building the neighbor and routing tables. > On the other hand, a transparent bridging would make node id for the > BaseStation not useful at all which prevents from using ACKs for > unicast transmissions - so how useful is it? > > - om_p > _______________________________________________ Tinyos-help mailing list Tinyos-help@millennium.berkeley.edu https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help