Hello.

Next to PIM networking is the second big missing piece.

After looking at NetworkManager, exalt and ConnMan we decided for the later one.

I don't want to got into all the details, this sometimes seems to be a emotional
topic, but give you the core points why we think that this is the best
alternative.

* Modular system. Almost all functionality can be realised as plugin. Make it
  easy to fit it out needs.
* Easy to use and expandable dbus API
* Connecting to open, wep128 and WPA2-PSK networks worked out of the box on a
  laptop.
* No need to write a full agent just to connect to wep and wpa networks
* Nice to embedded as it only has a few dependencies
* Driven by moblin, means it can use intel resources

We also had a brainstorming how to integrate it best with FSO. Given that
connman has it's own mature dbus API there is not much we would add on top of
it. Later some aggregation of available devices could be nice, but for now we
prefer to make use of the wifi chip at all.

A pragmatic approach with putting connman in the build and poking it and the
wifi driver as long as needed to be able to make it work good seems the best
appraoch. So that's what I will do in the next time.

I already have connman 0.1 in OE, but need real testing on the device. For sure
we will see problems with the wifi driver and perhaps other components. What I'm
aiming for is to be able to connect to open, wep and wpa networks. There will
come more, but this will already be some work I bet.

To have a good setup that allows me to work together with the kernel people I'll
do the work on an andy-tracking kernel. It get's into better shape and from the
timing it should be save to use when I reach the status that wifi get's usable
on the device.

regards
Stefan Schmidt

_______________________________________________
smartphones-standards mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/smartphones-standards

Reply via email to