>> To use the kernel-based CAN routing, you need to build the can-gw protocol >> which >> is based on the PF_CAN network layer stuff and currently only sits in the >> Berlios SVN: >> >> http://svn.berlios.de/wsvn/socketcan/trunk/kernel/2.6/net/can/gw.c?op=log&; >> > > That was patched into my kernel within minutes.
Fine :-) > Oliver Hartkopp wrote: >> >> You also need the configuration tool cangw which is in >> trunk/can-utils/cangw.c >> > > That should be easy too, I thought... > I have those can-utils in my target code so I just copied cangw.c there and > added it to the Makefile.... > Then I rewrote the Makefile to be less talking and more informative... > Then I hardened the compiler flags and then I was busy to remove all those > double- and shadow-declarations from all of the tools, including cangw.c... ??? - what do you mean in detail? Something to improve? > Oliver Hartkopp wrote: >> >> The can-gw is pretty fast and efficient, so that you won't see any >> relevant timing >> differences. >> >> If you get any problems using can-gw or the cangw userspacetool please ask >> me. >> >> I'm always interested in feedback ;-) >> > It's actually flashing and we see.... Ready! > > Normally, as I just like to mirror vcan0<->can0, my rule should be easy like > #cangw -A -s can0 -d vcan0 > #cangw -A -s vcan0 -d can0 -e > > and I run slave and master on vcan0. > I don't see the need for any filtering, are you? No - not in your use-case. It's just possible to do. E.g. if you have several single CAN-IDs to route, you can just create a script with several cangw -A ... lines and you're done. > A first simple compare of candump -ta can0 and candump -ta vcan0 looks > pretty fine! > Disconnected plug from can0... vcan0 still active -> Good! > Reconnecting the debugger... can0 active again -> Good! > CanOPEN LEDs constantly showing green... -> Not so perfect but that's one > for the next release :) You can probable route the error msgs too, or query the controller state via netlink. > > So I guess this is the correct setup? Yep! > Oliver Hartkopp wrote: >> >> Even if can-gw is using netlink i created a separate tool as all the >> CAN-stuff >> and filtering is so different to IP and other networking. >> > > And you did a pretty good job with that :) Thanks :-) > As kernel 2.6.39 has some different include paths I quick hacked them into > the code. After cleaning it up I can send you the code or a diff if you > like. Yes, of course. I'll probably put some #if KERNELVERSION stuff around that in the SVN. My plan is to check the gw.c code for the possibility to make Ethernet/IP routings from/to CAN before i post it on netdev ML. Currently only CAN_CAN mappings are possible. > A big thanks! Ditto! It's always nice to see, that things are requested i spent my time for ;-) Best regards, Oliver _______________________________________________ Socketcan-users mailing list [email protected] https://lists.berlios.de/mailman/listinfo/socketcan-users
