[Tinyos-devel] Re: [Tinyos Core WG] Proposal for VariableSizedQueueC component

[Philip Levis]

On Sep 20, 2007, at 1:12 AM, Jan Hauer wrote:

> I think it would be helpful to have async versions of QueueC (and
> PoolC). They could ease the potential timing-problem on the receive
> path when a second packet is received, before the first packet's
> task-handler has finished - e.g. CC2420ReceiveP has only one buffer
> for incoming packets (has this ever been an issue) ? Then, using a
> common pool for the low-level components and app would allow to
> utilize the memory better.

This is not an issue for a single packet (unless you make the packet  
payloads larger than the default): the CC2420 has a receive memory  
buffer, and the CC2420 stack spools out packets from this memory.  Is  
this a potential problem or a real problem? I'm leery of adding  
complexity unless there are documented cases. Since acquiring the SPI  
bus goes through a task and resource scheduling latency regardless,  
adding a receive queue might not help at all. A receive queue would  
only help if the latency appears only between the RXFIFO.readDone and  
its posted task.

Phil