[Tinyos-devel] Re: [net2-wg] bug in LqiForwardingEngineP.nc
gnawali at usc.edu
Tue Jan 1 20:10:59 PST 2008
On Dec 28, 2007 12:17 PM, Philip Levis <pal at cs.stanford.edu> wrote:
> On Dec 24, 2007, at 12:44 PM, Omprakash Gnawali wrote:
> > I was looking through some lqi traces. Once in a while (in three
> > experiments, 1.6%, 0.6%, and 0.5% of retransmission sequence), the
> > number of transmissions was larger than the max retransmissions. It
> > turns out fail_count can get reset two different ways - forwarding the
> > pkts or while sending locally sent pkts. If such reset occurs during
> > retransmissions, the number of retransmissions will be larger than the
> > max retransmissions. Do people buy this?
> This definitely looks like a bug. The 1.x MLQI didn't have two
> separate sends underneath, such that a local send could fail due to
> forwarding. So when I ported it I gave it two sends, but there is
> only one retx count.
Committed a fix that uses two variables to keep track of
retransmissions for those two sends.
More information about the Tinyos-devel