Hi Thiago, Thanks for the feedback.
On Tue, Jul 12, 2016 at 4:26 PM, Thiago Macieira <thiago.macieira at intel.com> wrote: > Em ter?a-feira, 12 de julho de 2016, ?s 12:35:22 PDT, Gregg Reynolds > escreveu: > > The Edison defaults to IPv4 on wlan0. Shouldn't that work? I finally > > figured out how to make it use IPv6, but naturally cannot find a way to > > make it go back to V4, or to run a dual stack. > > The default behaviour should be to use dual stack and turn IPv6 on. > > IPv4 requires active configuration. Turning IPv6 off requires active > configuration. > Still not sure what's going on, but at this point I do have the client (on the gateway) successfully discovering platform, device, and services on the server (on the Edison) using IPv4. But CRUDN ops still don't work which leads me to suspect the Java/JNI bits of the stack. > > To complicate things, the gateway is connected by ethernet wire to my > > wireless router, which is connected to my cable modem. So my devhost and > my > > Edison can connect wirelessly to the gateway, or to the router. So > there's > > a whole bunch of networking going on and I'm a long way from being a > > network expert. > > Unfortunately, you need to do some network debugging to figure out what's > happening. Please install wireshark or tcpdump on both your edison and your > devhost and watch the packets. Do packets sent from one arrive on the > other? > Are they on the same subnet? > I was kinda afraid somebody would say that. ;) Haven't done it yet. I did manage to get some of the sample programs talking to each other across the network, which again leads me to suspect the Java stuff (my code is actually Clojure, but it translates pretty directly into Java). > > Is your wireless router configured for "client isolation"? > I don't believe so. In any case, I use the gateway as the AP: my mac and the edison are both on its network, so I think the router is out of the picture. The mac and the edison can talk to each other; I ssh into the edison from my mac. At this point I know that the server does respond to discovery requests and to GET/POST/etc. on resources from some of the sample programs, so the problem must be on the client side. I just rebuilt the stack with logging so I'm hoping I'll see something in the log. My suspicion is that somehow e.g. a "get" call on an OcResource (in the Java API) is not getting properly translated down to the right call in the c stack. Or something like that, who knows. It doesn't error out so maybe it just gets stuck somewhere. ( I note in passing that the Java examples do not really exercise the CRUDN stuff very much.) Or itt might be something as simple as bad linking - the Edison comes with Iotivity (1.0.0?) preinstalled, and I've been bitten hard a few times by forgetting to ldconfig stuff correctly. I'll let you know what I come up with, thanks, gregg -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.iotivity.org/pipermail/iotivity-dev/attachments/20160713/3aa161a8/attachment.html>
