Hi Omprakash,

Thanks for the prompt response. We use:

-I$(TOSDIR)/lib/net/ctp  -I$(TOSDIR)/lib/net/le

I've checked the getLinkQuality and it seems different from what you sent.
We have been using the latest standard distribution (2.0.2) rather than the
CVS checkout:

command uint8_t LinkEstimator.getLinkQuality(am_addr_t neighbor) {
    uint8_t idx;
    idx = findIdx(neighbor);
    if (idx == INVALID_RVAL) {
      return INFINITY;
    } else {
      return NeighborTable[idx].eetx;
    };
  }

I'll test our application with the latest CVS version and let you know about
the results.
Thanks,
Bulut


On Mon, Apr 21, 2008 at 11:41 PM, Omprakash Gnawali <[EMAIL PROTECTED]> wrote:

> 2008/4/21 Bulut ERSAVAS <[EMAIL PROTECTED]>:
> > Hi Omprakash,
> >
> > Concerning this issue, we have had chance to reproduce the problem. I am
> > attaching 3 files including beacons snooped from the network. We believe
> > that the problem is caused by nodes selecting parents who had direct
> > connection to the root in the past. When such parents are moved away
> from
> > the root, it looks like they don't remove root from their neighborhood
> table
> > and keep them as their parent.
>
> It seems like you have a bug that was fixed in December/January. The
> bug was getLinkQuality was not returning a large ETX value for links
> that are not mature. Which link estimator are you using? Does your
> getLinkQuality(...) look something like this:
>
>  443   command uint16_t LinkEstimator.getLinkQuality(am_addr_t neighbor) {
>  444     uint8_t idx;
>  445     idx = findIdx(neighbor);
>  446     if (idx == INVALID_RVAL) {
>  447       return VERY_LARGE_EETX_VALUE;
>  448     } else {
>  449       if (NeighborTable[idx].flags & MATURE_ENTRY) {
>  450   return NeighborTable[idx].eetx;
>  451       } else {
>  452   return VERY_LARGE_EETX_VALUE;
>  453       }
>  454     }
>  455   }
>
> It also seems like you are not using the latest BaseStation because
> the source addresses are not preserved when they are sent over the
> UART. This makes it quite difficult to know which lines correspond to
> which packets from which node - although I used elimination to
> determine what packets were sent by which nodes.
>
> - om_p
>
_______________________________________________
Tinyos-help mailing list
[email protected]
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

Reply via email to