On ma, 12 nov 2018 11:01:36 +0100, Oleksij Rempel wrote: > On 04.11.18 21:49, Kurt Van Dijck wrote: > >On za, 03 nov 2018 09:24:42 +0100, Oleksij Rempel wrote: > >>On Fri, Nov 02, 2018 at 07:40:48PM +0100, Kurt Van Dijck wrote: > >>>On vr, 02 nov 2018 12:32:19 +0100, Oleksij Rempel wrote: > >>>> > >>>>we will drop this packets soon or later... so drop it better ASAP. > >>> > >>>This is true, unless you monitor all j1939 traffic in 'promiscuous' > >>>mode. I'm not sure if that hasn't been thrown away yet, and is something > >>>designed to be used on a PC anyway. You could extend the check a bit. > >> > >>Same like with ipv4/ipv6/tcp/udp stacks, we do not process unrelated > >>traffic. We capture RAW Ethernet traffic. In this case, in promiscuous mode we > >>should capture RAW CAN traffic, not decoded J1939 traffic. > > > >That means a userspace decoder for j1939 promiscuous spying. > >That makes sense. > > @Kurt, to confirm, you have nothing against this patches? Sorry for not being very clear. Yes, you convinced me, decoding j1939 traffic between other hosts does not necessarily belong in kernel, and since it complicates things here and there, I agree to remove it. Kind regards, Kurt
