[mc1322x] Radio demos

[Mariano Alvira]

On Wed, Oct 07, 2009 at 08:59:19AM -0400, Jon Smirl wrote:
> I can compile now using the OE toolchain. I don't know what is wrong
> with the Phytec one.
> 
> Which radio demos work?

rftest-tx and rftest-rx in mc1322x-tests are the "standard" radio test
programs. You can use rftest-rx and rftestrx2pcap.pl to make pcaps
that you can load into wireshark to examine (this is only helpful if
you are sending packets in an 802.15.4 frame).

As for contiki, I've used nullmac and sicslowmac with success. xmac
and lpp seem to have problems, but it's difficult to separate issues
in contiki with issues in my radio driver. I believe the root of the
problem is that there are assumptions in xmac and lpp regarding the
radio driver blocking until the bytes are transmitted (since that's
how a lot of the radio's contiki uses work). With the mc1322x radio
you fill a buffer and DMA takes care of the rest and so it often
returns before the bytes are transmitted. What I would like to do is
change the radio driver from a contiki process over to an interrupt
driven thing, but I haven't had a chance to work on that yet...

I'd use sicslowmac (which is what is currently used in my repo). I've
had success with example-mesh and with example-collect. Both of these
are higher level RIME protocols and test a lot of the RIME
primitives. example-abc is the most basic RIME test. I program the
RIME address into NVRAM at 0x1E000, as in the example on the website:

./mc1322x-load.pl -f flasher.bin -s example-mesh.bin -t /dev/ttyUSB0
0x1e000,0x01020304,0x05060708


> What's needed to get IPV6 going?

That is a good question. I haven't gotten into that yet. I do plan to
work on that soon though (over the next few months).

-Mar.