[PATCH 0/7 v3] wpa_supplicant P2P DBus API improvements

Jouni Malinen j at w1.fi
Sun Jun 29 11:03:52 EDT 2014


On Wed, Jun 25, 2014 at 11:30:40AM +0300, Tomasz Bursztyka wrote:

> Something tells me no-one has been using P2P DBus API yet. Besides
> this patch,
> I fixed already quite some issues (already upstream). Without all
> those it's basically
> impossible to get a consistent P2P experience through this API. So
> people would have
> complained already before I guess.

That kind of points towards a different question: What's the point of
maintaining an interface that has not been used for three years? Since
you are working on this now, I'd hope that there is a real use case
behind this. Otherwise, I'd just rather remove the interface if no one
is really using it.

Since there has been no comments from others on the changes, I applied
these now. There were couple of forgotten updates on the documentation
side (e.g., change from P2P Interface Address to P2P Device Address)
that I fixed in the patches. Anyway, the externally visible behavior
should be identical.

> >As a friendly hint, it would also make it easier and likely faster for
> >me to accept D-Bus changes if someone were to contribute hwsim test
> >cases (see tests/hwsim/README) to exercise the D-Bus interface..:-)
> 
> Sure, why not.
> I noticed some python script there implementing a test for a signal
> which is not implemented
> for instance (P2PStateChanged). So there is probably both sides to
> work on yhere.
> (Actually I don't know what's the point of this signal)

I'd assume you are referring to the wpa_supplicant/examples/p2p/*
scripts. These are not used in any way in hwsim test cases, but they
could be updated to match the current D-Bus API. Anyway, I'd still like
to see the hwsim scripts extended to test the D-Bus interface. Since I
don't really use this myself almost at all, it would be especially
useful to have automated test cases that would catch regressions (and
well, also areas that don't work at all in the first place, if that was
indeed the case with the earlier P2P interface design).

-- 
Jouni Malinen                                            PGP id EFC895FA


More information about the HostAP mailing list