"Niels Reedijk" <n.reedijk@xxxxxxxxx> wrote: > > Sure. But I am quite anal about our coding style, so if you have > > any > > worries, you better send it over to someone else - of course I am > > kidding (except for the anal part ;-)) > I'll change the code style as soon as I'm in the process of cleaning > up > (I like the way it is coded now, but I am sure there are tons of > people > who don't agree). I'll send the driver to Philippe to be on the safe > side ;-) That's fine with me :-)) > > > > But which new interface API are you referring to? > > > I assumed that the kernel network stack would use a different > > > API, > > > because the API between the driver and the current net_server > > > addon > > > is private > > > and thus undefined. > > It's not private - it's defined and documented in the sample > > network > > driver from Be Inc. > > We will continue to use it, at least for a start; we may > > enhance/change > > it a bit over time, though - but we'll always keep backwards > > compatibility. > I think I don't really agree there. For example, my driver wouldn't > work > with the realtek addon Be provided. The read hook wouldn't quite work > together with what the add-on wanted. So I've written my own very > simple > addon. Now I understand what you mean! But that's still not really correct; the API to the net_server is public. But you can use an internal API to your driver if you wish to - but since Be's Realtek driver also works with BONE which don't use the add-ons anymore, it's safe to say that also this driver will use the standard interface. I would guess it's something else. Adios... Axel.