Re: A possible case of TCP server no-response, was: Re: bufs lost in TCP connection establishment

Paul Sokolovsky


On Thu, 15 Sep 2016 02:51:49 -0000
"Flavio Santes" <flavio.santes(a)> wrote:

One particularly naughty conditions I experienced is that while my
usual testcase usually proceeded beyond connect() call, sometimes I
got time "strips" when connect() call just hanged (I don't use

This is a hardware-agnostic issue, see: The issue description
is not really useful, although comments will help. BTW, we apply the
same workaround: restart the server :)
Thanks, I read it, but not sure it's the same issue as I described
above - indeed, there're many different issues. I checked into Z source,
and found that it actually tries to make a random local port number (if
user passed one as 0). But Wireshark still shows that after a restart,
1025 is what gets actually sent on wire, and the situation got pretty
unpleasant when I moved from local servers to knocking on AWS IoT ones,
which, well, I can't restart.

So, digging further, turned that Zephyr networking layer and uIP live
their own separate lives wrt to local port numbering. makes them talk together,
though on uIP's terms (using global variables). As patch description
says, I'm open to suggestions to redo it via adding function arguments.


On Wed, 7 Sep 2016 15:26:23 +0000
Rohit Grover <Rohit.Grover(a); wrote:

Best Regards,
Paul | Open source software for ARM SoCs
Follow Linaro:!/linaroorg -

Join to automatically receive all group messages.