Hi Johny,

I have looked briefly at your patch. There are a lot of changes, and
it will take time to sort out all these to my satisfaction. What we
could do is the following, if you are willing to go along:

I will create a new branch on my git repository, this is going to be
based on the current tinyos.main SVN repo. I will incorporate some of
your changes and prepare the commits. Each change will not break any
other code. I would like to go as slowly as possibly. After each
change I would like to look at the remaining issues in a separate
diff. Alternatively, it would be nice if you have a public git repo I
can look at (and some instructions on what to use and what to ignore).

The RF212 code base is fine, once that is done, it will be trivial to
support other rfxlink stacks.  Maybe I will do the RF230 synchronously
so I can compile the changes myself.

I cannot compile the current tinyos-main for telosb, it gives some
compiler errors. Do you see this too? Can you compile your version for
telosb? (Of course I assume it compiles with cc2420x.

Best,
Miklos

On Wed, Nov 30, 2011 at 2:50 AM, Johny Mattsson <jmatts...@dius.com.au> wrote:
> On Wed, Nov 30, 2011 at 8:17 AM, Miklos Maroti <mmar...@math.u-szeged.hu>
> wrote:
>>
>> I will update the IRIS code to work with BLIP after the current
>> release. I do not want to mess with the code before that to keep as
>> stable as possible for the release.
>
>
> Cool. Let me pull together a fresh patch for you then - we've been doing a
> fair bit of internal BLIP-related work here that I'd love to see adopted
> upstream. We're using RF212 chips here, but I just got a bit of time
> allocated to make our changes work with the CC2420X driver as well, so I'm
> hoping that means it'll give you a good starting point if nothing else.
>
> Improvements we've got running here include:
> - BLIP on RF212
> - BLIP using extended IEEE 15.4 addresses (we can't assign static node IDs
> due to scale, so need DHCPv6 to work)
> - Support for both short and extended IEEE 15.4 addresses in rfxlink (I
> submitted most of this over in issue 30 already)
> - Support for on-chip frame filtering for short and extended addresses
> (needed to avoid packet storms when meshing, we discovered)
> - Support for larger IPv6 packet sizes (up to 1280)
>
>
> As I mentioned when I posted my first patch to the list, there are a few
> things that I'm definitely not happy about with it:
> - It doesn't resolve the confusion around the Send, BareSend and Ieee154Send
> interfaces - in many locations each of them gets renamed to another!
> - Both the BLIP and the Ieee154 packet layers hook into the
> TinyosNetworkLayerC's Ieee154Receive.receive, which is a very bad idea if
> the latter was ever used (it works for our cases, which is why I haven't
> been able to find time to fix it)
> - There should be a more elegant way of structuring this to support the AM,
> Ieee154 and BLIP stacks neatly next to each other. Again, I haven't had the
> time to even look at it, much less get a discussion happening on -dev.
>
>
> ** goes diffing in the internal repo **
>
>
> Okay, I've attached a patch here (it will get stripped for the list, I know
> - if anyone else wants a copy, just contact me off-list). This patch
> contains all the BLIP/RF212 related changes we have running here, and also
> dry-coded changes to make PppRouter and CoapBlip apps build for telosb using
> the cc2420x driver with our mods. The cc2420x changes should make it
> functionally equivalent to the RF212, but I have no hardware to verify it,
> sorry.
>
>
> To avoid any potential confusion: We're not using IRIS motes here at DiUS.
> We have custom hardware based around the LSR siflex02 chip (ATxmega256a3 +
> rf212), and also use LSR's modflex dev kit board. We're not yet ready to
> submit the x256a3/siflex02 platform support upstream however - there are
> some embarrassing issues there we still need to resolve :)
>
>
> Best regards,
> /Johny
>
> --
> Johny Mattsson
> Senior Software Engineer
>
> DiUS Computing Pty. Ltd.
> where ideas are engineered
>
_______________________________________________
Tinyos-help mailing list
Tinyos-help@millennium.berkeley.edu
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

Reply via email to