> > Yes that is what it looks like here as well. I always test ax25 by >using a ping to nodes on both ports. In 2.2.1 the ping and later call >applications ran fine, but nothing went out... > > The impression I get is that the writers of the ax25-utils used >some functions of the kernel that are changing toward other uses. > Ok we are talking about the very same thing. I did a dumb test with kissattach, first attaching a console (/dev/ttyx) instead a real serial, then attaching a phone modem on the serial instead a tnc. In both cases I saw something getting out the serial, now I am going to use a laptop connected with a null modem cable. ..... I did it, and the output on the serial port is correct (sequence of C0 ... C0) I used a good 2.0.36 kernel and ax25 utils in the first case (for comparation) then I recorded a beacon with the laptop. Second case is kernel 2.2.1 ad ax25 utils recompiled. The kiss sequence is the same, so why in the first case the tx is switched but not in the second case ? Perhaps hardware handshaking has something to do with that ? Bye , Martino.