Hi Kishen

Thanks for the follow-up. Yes, I will submit it although I would like to
have a more generic fix. This fix uses the hard-wired address in the
prj.conf file. Would like to have an option of using auto-address
generation if the network has a router (e.g. radvd) distributing prefixes.

Khaled


On Tue, Sep 19, 2017 at 4:22 PM, Maloor, Kishen <kishen.mal...@intel.com>
wrote:

> Hi Khaled,
>
> Its quite simple - constrained OSes may
> be configured in different ways that are
> highly dependent on the application in
> question. You might set up nodes that take
> on one or the other role in an IPv6 mesh,
> for e.g., or set different names to
> advertise over BLE, etc., not to mention the
> memory allocation sizes. This is why
> the port/<OS>/* adaptations for constrained
> OSes are to strictly be taken as samples
> and the port/* interfaces have been
> (by design) made simple enough for anyone
> to understand.
>
> As you noticed, this works differently for
> the rich OS (Linux, Windows, etc.)
> adaptations that can be generic as they’re
> able to consume higher-level APIs from those
> OSes.
>
> The Zephyr sample used to work as-is and its
> likely that its network stack has changed
> a bit since. IoTivity-Constrained is
> currently being updated for OCF 1.3
> compliance (security, etc.) so if you
> have found a fix for this issue that works
> please just directly submit it to gerrit and
> I could merge it in.
>
> Thanks,
> -Kishen.
>
>
>
> -
> Kishen Maloor
> Intel Open Source Technology Center
>
>
>
>
> From:  <iotivity-dev-boun...@lists.iotivity.org> on behalf of Khaled
> Elsayed <khaledi...@gmail.com>
> Reply-To:  "khaledi...@gmail.com" <khaledi...@gmail.com>
> Date:  Tuesday, September 19, 2017 at 1:45 AM
> To:  "Macieira, Thiago" <thiago.macie...@intel.com>
> Cc:  iotivity-dev <iotivity-dev@lists.iotivity.org>
> Subject:  Re: [dev] IoTivity constrained zephyr port connectivity issues
>
>
> Hi,
>
> Actually, in theory this is correct. However, as far as I understand the
> way the iotivity-constrained ports work is that this is done in the
> function oc_connectivity_init as a way to help developers call a single
> platform-independent function to initialize
>  the interfaces. This is the case for RIOT and Contiki. For RIOT
> oc_connectivity_init does eventually call gnrc_ipv6_netif_init_by_dev
> which does initialization of the IP address. In Contiki, it starts the
> ip_adapter_process which also sets IP address. For
>  zephyr it binds the multicast address and does not set a link local or
> global address for the interface and this is where the bug seems to be.
>
> The exception is the linux port which should of course starts with
> everything in place via the linux box setup.
>
>
>
>
> On Mon, Sep 18, 2017 at 4:38 PM, Thiago Macieira
> <thiago.macie...@intel.com> wrote:
>
> On Monday, 18 September 2017 01:29:20 PDT Khaled Elsayed wrote:
> > es, but you closed it as invalid. This is part of the release code for
> > zephyr in the file port/zephyr/src/ipadpater.c.
> >
> > It was not in my code :-)
> >
> > It should be corrected in the iotivity-constrained zephyr port source
> >code.
> > I can file an issue or make a new pull request on github.
> >
> > Please advise
>
> Correct, because setting up the IP addresses is out of scope. You must do
> that
> in your application.
>
> --
> Thiago Macieira - thiago.macieira (AT)
> intel.com <http://intel.com>
>   Software Architect - Intel Open Source Technology Center
>
>
>
>
>
>
>
>
>
_______________________________________________
iotivity-dev mailing list
iotivity-dev@lists.iotivity.org
https://lists.iotivity.org/mailman/listinfo/iotivity-dev

Reply via email to