> > I'm fine with that, provided that libdlpi is enhanced to have some way to > > report DL_CURR_DEST_ADDR indications. > > Okay, that's fine, and I can work on that separately.
The interim libdlpi changes don't look quite right to me. In particular, i_dlpi_notifyind_process() needs to silently ignore properly-formed messages it doesn't understand, rather than passing that error onto e.g. the innocent caller issuing the dlpi_recv(). -- meem
