Binder interface for wpa_supplicant?

[Jouni Malinen]

On Wed, Dec 30, 2015 at 01:04:30PM -0800, Christopher Wiley wrote:
> The WiFi group in Android is interested in developing a control interface
> for wpa_supplicant based on binder (Android’s native IPC system). This is
> chiefly desirable so that we can get a reasonable RPC interface with
> generated marshaling logic (no string parsing or handwritten boilerplate on
> either side).
> 
> We would love to send patches upstream if there is interest in keeping this
> implementation in-repository. The intent is to keep IPC and communication
> logic separate from core logic by using existing abstractions or
> introducing new ones where appropriate. The tentative plan is to parallel
> the existing pair of DBus interface implementations, but thoughts on how to
> best express this are appreciated.

Based on the experience I have had with the D-Bus interface, I'd very
much like to make sure this could be done in a way that can be included
in the upstream repository and automated test setup (tests/hwsim/* and
use of mac80211_hwsim driver). There were number of significant issues
that showed up when I finally wrote some test cases for the D-Bus
interface a year ago and it used to be way too easy to break things that
were not automatically tested.

There are already three supported control interface options and each one
of those is of pretty significant amount of code. While I hope we can
eventually get rid of the older D-Bus interface (though, it doesn't
really get any extensions now, so its maintenance is at least relatively
small effort), having three significant interfaces provide more or less
the same capabilities does come with a price.. As such, any direction
that would help re-use some of the design or even implementation would
certainly be welcome.

One more challenge with the D-Bus interface has been that it is still
far from the capabilities of the ctrl_iface implementation. The latter
happens to be the one that I use as the first target for everything new
and in practice, it will be superset of all control interface in what
kind of functionality gets exposed. Having active contributors to help
maintain any new interface would obviously be important to avoid issues
with an interface not being sufficient for something.

-- 
Jouni Malinen                                            PGP id EFC895FA