* AW: Need support J1939 protocol support to my kernel (v3.8 on BeagleBone Black)
@ 2013-11-11 16:14 Michael Luxen
2013-11-12 1:10 ` swapnil
0 siblings, 1 reply; 9+ messages in thread
From: Michael Luxen @ 2013-11-11 16:14 UTC (permalink / raw)
To: linux-can@vger.kernel.org, patilswa@gmail.com
Hello Swapnil,
I've success with building the J1939 stack on my BBB based on Debian using these steps:
git clone git://gitorious.org/linux-can/can-utils.git
cd can-utils
--- get J1939 stuff ---
git remote add j1939 git://gitorious.org/~kurt-vd/linux-can/linux-can-j1939.git
git remote update j1939
--- depending on kernel ---
git merge j1939/v3.6
...or…
git merge j1939/v3.8
Regards
Mike
-----Ursprüngliche Nachricht-----
Von: Swapnil Patil <patilswa@gmail.com>
Gesendet: Mo 11.11.2013 16:46
Betreff: Need support J1939 protocol support to my kernel (v3.8 on BeagleBone Black)
An: linux-can@vger.kernel.org;
> Hi Kurt,
>
> Thanks for your reply.
> I am using Begalbone Black with Ubuntu 13.04
> I am able to transmit the RAW and extended can message over the can
> bus using can utility
> Now I want to use J1939 stack.
> I downloaded the
> Linux-can-can-j1939-modules and Linux-can-iproute2-j1939 and tried to
> build them but it failed
>
> Log of iproute2-j1939$ make
>
> make[1]: Entering directory `/home/ubuntu/BBB/iproute2-j1939/misc'
> gcc -D_GNU_SOURCE -O2 -Wstrict-prototypes -Wall -I../include
> -DRESOLVE_HOSTNAM -DLIBDIR=\"/usr/lib/\" -I -o arpdarpd.c
> ../lib/libnetlink.a ../lib/libutil.-ldb -lpthread
> gcc: error: arpd: No such file or directory
>
> Log of can-j1939-modules: make
> /home/ubuntu/BBB/can-j1939-modules/net/can/j1939/rtnl.c:232:25: error:
> âstruct netlink_skb_parmsâ has no member namedâpidâ
> /
>
> If someone already build this on Begalbone black with ubuntu
> 13.04 or angstrom kernel Please add the steps
>
> Thanks
> Swapnil
> --
> To unsubscribe from this list: send the line "unsubscribe linux-can" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: AW: Need support J1939 protocol support to my kernel (v3.8 on BeagleBone Black) 2013-11-11 16:14 AW: Need support J1939 protocol support to my kernel (v3.8 on BeagleBone Black) Michael Luxen @ 2013-11-12 1:10 ` swapnil 2013-11-12 10:02 ` Kurt Van Dijck 0 siblings, 1 reply; 9+ messages in thread From: swapnil @ 2013-11-12 1:10 UTC (permalink / raw) To: linux-can Michael Luxen <mluxen <at> gmx.net> writes: > > Hello Swapnil, > > I've success with building the J1939 stack on my BBB based on Debian using these steps: > > git clone git://gitorious.org/linux-can/can-utils.git > cd can-utils > --- get J1939 stuff --- > git remote add j1939 git://gitorious.org/~kurt-vd/linux-can/linux-can- j1939.git > git remote update j1939 > --- depending on kernel --- > git merge j1939/v3.6 > ...or… > git merge j1939/v3.8 git merge j1939/j1939-v3.8 after this step ubuntu@arm:~/BBB/can-utils$ git merge j1939/j1939-v3.8 error: The following untracked working tree files would be overwritten by merge: .mailmap COPYING CREDITS Documentation/.gitignore Documentation/00-INDEX Documentation/ABI/README Documentation/ABI/obsolete/sysfs-bus-usb Documentation/ABI/obsolete/sysfs-class-rfkill Documentation/ABI/obsolete/sysfs-driver-hid-roccat-koneplus Documentation/ABI/removed/devfs Documentation/ABI/removed/dv1394 Documentation/ABI/removed/o2cb Documentation/ABI/removed/raw1394 Documentation/ABI/removed/video1394 Documentation/ABI/stable/firewire-cdev Documentation/ABI/stable/o2cb Documentation/ABI/stable/syscalls Documentation/ABI/stable/sysfs-acpi-pmprofile Documentation/ABI/stable/sysfs-bus-firewire Documentation/ABI/stable/sysfs-bus-xen-backend Documentation/ABI/stable/sysfs-class-backlight Documentation/ABI/stable/sysfs-class-rfkill Documentation/ABI/stable/sysfs-class-ubi Documentation/ABI/stable/sysfs-devices-node Documentation/ABI/stable/sysfs-devices-system-xen_memory Documentation/ABI/stable/sysfs-driver-qla2xxx Documentation/ABI/stable/sysfs-driver-usb-usbtmc Documentation/ABI/stable/sysfs-firmware-efi-vars Documentation/ABI/stable/sysfs-module Documentation/ABI/stable/thermal-notification Documentation/ABI/stable/vdso Documentation/ABI/testing/configfs-spear-pcie-gadget Documentation/ABI/testing/debugfs-ec Documentation/ABI/testing/debugfs-ideapad Documentation/ABI/testing/debugfs-pktcdvd Documentation/ABI/testing/evm Documentation/ABI/testing/ima_policy Documentation/ABI/testing/procfs-diskstats Documentation/ABI/testing/pstore Documentation/ABI/testing/sysfs-ata Documentation/ABI/testing/sysfs-block Documentation/ABI/testing/sysfs-block-zram Documentation/ABI/testing/sysfs-bus-bcma Documentation/ABI/testing/sysfs-bus-css Documentation/ABI/testing/sysfs-bus-i2c-devices-fsa9480 Documentation/ABI/testing/sysfs-bus-i2c-devices-hm6352 Documentation/ABI/testing/sysfs-bus-media Documentation/ABI/testing/sysfs-bus-pci Documentation/ABI/testing/sysfs-bus-pci-devices-cciss Documentation/ABI/testing/sysfs-bus-pci-drivers-ehci_hcd Documentation/ABI/testing/sysfs-bus-rbd Documentation/ABI/testing/sysfs-bus-umc Documentation/ABI/testing/sysfs-bus-usb Documentation/ABI/testing/sysfs-bus-usb-devices-usbsevseg Documentation/ABI/testing/sysfs-c2port Documentation/ABI/testing/sysfs-class Documentation/ABI/testing/sysfs-class-backlight-driver-adp8870 Documentation/ABI/testing/sysfs-class-bdi Documentation/ABI/testing/sysfs-class-devfreq Documentation/ABI/testing/sysfs-class-lcd Documentation/ABI/testing/sysfs-class-led Documentation/ABI/testing/sysfs-class-mtd Documentation/ABI/testing/sysfs-class-net-batman-adv Documentation/ABI/testing/sysfs-class-net-mesh Documentation/ABI/testing/sysfs-class-pktcdvd Documentation/ABI/testing/sysfs-class-power Documentation/ABI/testing/sysfs-class-regulator Documentation/ABI/testing/sysfs-class-rtc-rtc0-device- rtc_calibration Documentation/ABI/testing/sysfs-class-scsi_host Documentation/ABI/testing/sysfs-class-uwb_rc Documentation/ABI/testing/sysfs-class-uwb_rc-wusbhc Documentation/ABI/testing/sysfs-dev Documentation/ABI/testing/sysfs-devices Documentation/ABI/testing/sysfs-devices-memory Documentation/ABI/testing/sysfs-devices-mmc Documentation/ABI/testing/sysfs-devices-platform-_UDC_-gadget Documentation/ABI/testing/sysfs-devices-platform-docg3 Documentation/ABI/testing/sysfs-devices-power Documentation/ABI/testing/sysfs-devices-system-cpu Documentation/ABI/testing/sysfs-devices-system-ibm-rtl Documentation/ABI/testing/sysfs-driver-hid Documentation/ABI/testing/sysfs-driver-hid-logitech-lg4ff Documentation/ABI/testing/sysfs-driver-hid-multitouch Documentation/ABI/testing/sysfs-driver-hid-picolcd Documentation/ABI/testing/sysfs-driver-hid-prodikeys Documentation/ABI/testing/sysfs-driver-hid-roccat-arvo Documentation/ABI/testing/sysfs-driver-hid-roccat-isku Documentation/ABI/testing/sysfs-driver-hid-roccat-kone Documentation/ABI/testing/sysfs-driver-hid-roccat-koneplus Documentation/ABI/testing/sysfs-driver-hid-roccat-kovaplus Docume Aborting didn't understand ?? what is the next step > > Regards > Mike > > -----Ursprüngliche Nachricht----- > Von: Swapnil Patil <patilswa <at> gmail.com> > Gesendet: Mo 11.11.2013 16:46 > Betreff: Need support J1939 protocol support to my kernel (v3.8 on BeagleBone Black) > An: linux-can <at> vger.kernel.org; > > Hi Kurt, > > > > Thanks for your reply. > > I am using Begalbone Black with Ubuntu 13.04 > > I am able to transmit the RAW and extended can message over the can > > bus using can utility > > Now I want to use J1939 stack. > > I downloaded the > > Linux-can-can-j1939-modules and Linux-can-iproute2-j1939 and tried to > > build them but it failed > > > > Log of iproute2-j1939$ make > > > > make[1]: Entering directory `/home/ubuntu/BBB/iproute2-j1939/misc' > > gcc -D_GNU_SOURCE -O2 -Wstrict-prototypes -Wall -I../include > > -DRESOLVE_HOSTNAM -DLIBDIR=\"/usr/lib/\" -I -o arpdarpd.c > > ../lib/libnetlink.a ../lib/libutil.-ldb -lpthread > > gcc: error: arpd: No such file or directory > > > > Log of can-j1939-modules: make > > /home/ubuntu/BBB/can-j1939-modules/net/can/j1939/rtnl.c:232:25: error: > > âstruct netlink_skb_parmsâ has no member namedâpidâ > > / > > > > If someone already build this on Begalbone black with ubuntu > > 13.04 or angstrom kernel Please add the steps > > > > Thanks > > Swapnil > > -- > > To unsubscribe from this list: send the line "unsubscribe linux-can" in > > the body of a message to majordomo <at> vger.kernel.org > > More majordomo info at http://vger.kernel.org/majordomo-info.html > > > -- > To unsubscribe from this list: send the line "unsubscribe linux-can" in > the body of a message to majordomo <at> vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > > ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: AW: Need support J1939 protocol support to my kernel (v3.8 on BeagleBone Black) 2013-11-12 1:10 ` swapnil @ 2013-11-12 10:02 ` Kurt Van Dijck 2013-11-28 4:49 ` swapnil 0 siblings, 1 reply; 9+ messages in thread From: Kurt Van Dijck @ 2013-11-12 10:02 UTC (permalink / raw) To: swapnil, Michael Luxen; +Cc: linux-can Hey all, Maybe Michael was a bit too fast. I suspect you tried to merge the j1939 linux kernel into can-utils ? On Tue, Nov 12, 2013 at 01:10:04AM +0000, swapnil wrote: > Michael Luxen <mluxen <at> gmx.net> writes: > > > > > Hello Swapnil, > > > > I've success with building the J1939 stack on my BBB based on Debian using > these steps: > > > > git clone git://gitorious.org/linux-can/can-utils.git > > cd can-utils > > --- get J1939 stuff --- So you're in the can-utils directory now. Merge can-utils-j1939 $ git remote add j1939 git://gitorious.org/~kurt-vd/linux-can/can-j1939-utils.git $ git remote update j1939 $ git merge j1939/master build can-utils now. And then: $ cd .. $ git clone <your kernel> $ cd linux (or your linux kernel directory) and then these steps apply. > > git remote add j1939 git://gitorious.org/~kurt-vd/linux-can/linux-can-j1939.git > > git remote update j1939 > > --- depending on kernel --- > > git merge j1939/v3.6 > > ...or… > > git merge j1939/v3.8 Maybe this will work better. Kind regards, Kurt ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: AW: Need support J1939 protocol support to my kernel (v3.8 on BeagleBone Black) 2013-11-12 10:02 ` Kurt Van Dijck @ 2013-11-28 4:49 ` swapnil 2013-11-28 8:52 ` Kurt Van Dijck 0 siblings, 1 reply; 9+ messages in thread From: swapnil @ 2013-11-28 4:49 UTC (permalink / raw) To: linux-can Kurt Van Dijck <kurt.van.dijck <at> eia.be> writes: > > Hey all, > > Maybe Michael was a bit too fast. > I suspect you tried to merge the j1939 linux kernel into can-utils ? > > On Tue, Nov 12, 2013 at 01:10:04AM +0000, swapnil wrote: > > Michael Luxen <mluxen <at> gmx.net> writes: > > > > > > > > Hello Swapnil, > > > > > > I've success with building the J1939 stack on my BBB based on Debian using > > these steps: > > > > > > git clone git://gitorious.org/linux-can/can-utils.git > > > cd can-utils > > > --- get J1939 stuff --- > > So you're in the can-utils directory now. Merge can-utils-j1939 > > $ git remote add j1939 git://gitorious.org/~kurt-vd/linux-can/can-j1939- utils.git > $ git remote update j1939 > $ git merge j1939/master > > build can-utils now. > And then: > > $ cd .. > $ git clone <your kernel> > $ cd linux (or your linux kernel directory) > > and then these steps apply. > > > > git remote add j1939 git://gitorious.org/~kurt-vd/linux-can/linux-can- j1939.git > > > git remote update j1939 > > > --- depending on kernel --- > > > git merge j1939/v3.6 > > > ...or… > > > git merge j1939/v3.8 > > Maybe this will work better. > > Kind regards, > Kurt > -- > To unsubscribe from this list: send the line "unsubscribe linux-can" in > the body of a message to majordomo <at> vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > > Hi Kurt , Thanks for adding the information on elinux.org/J1939 I am able to build the kernal for begabone black using j1939 but when i tried to build iproute2-j1939 root@arm:/home/ubuntu/OBDscan/iproute2-j1939# make make[1]: Entering directory `/home/ubuntu/OBDscan/iproute2-j1939/misc' gcc -D_GNU_SOURCE -O2 -Wstrict-prototypes -Wall -I../include - DRESOLVE_HOSTNAMES -DLIBDIR=\"/usr/lib/\" -I/usr//usr/include -o arpd arpd.c ../lib/libnetlink.a ../lib/libutil.a -ldb -lpthread arpd.c:19:20: fatal error: db_185.h: No such file or directory #include <db_185.h> ^ compilation terminated. make[1]: *** [arpd] Error 1 make[1]: Leaving directory `/home/ubuntu/OBDscan/iproute2-j1939/misc' make: *** [all] Error 2 i am getting above error. i tried to install sudo apt-get install db4.7-util but did not worked if y >ip link set can0 j1939 on Error: either "dev" is duplicate, or "j1939" is a garbage. can you please help me. I am really trying hard to build it on begalbone black Thank You Swapnil ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: AW: Need support J1939 protocol support to my kernel (v3.8 on BeagleBone Black) 2013-11-28 4:49 ` swapnil @ 2013-11-28 8:52 ` Kurt Van Dijck 2013-12-04 22:10 ` swapnil 0 siblings, 1 reply; 9+ messages in thread From: Kurt Van Dijck @ 2013-11-28 8:52 UTC (permalink / raw) To: swapnil; +Cc: linux-can Hey, > Hi Kurt , > Thanks for adding the information on elinux.org/J1939 > I am able to build the kernal for begabone black using j1939 > but when i tried to build iproute2-j1939 > > root@arm:/home/ubuntu/OBDscan/iproute2-j1939# make > make[1]: Entering directory `/home/ubuntu/OBDscan/iproute2-j1939/misc' > gcc -D_GNU_SOURCE -O2 -Wstrict-prototypes -Wall -I../include - > DRESOLVE_HOSTNAMES -DLIBDIR=\"/usr/lib/\" -I/usr//usr/include -o arpd > arpd.c ../lib/libnetlink.a ../lib/libutil.a -ldb -lpthread > arpd.c:19:20: fatal error: db_185.h: No such file or directory > #include <db_185.h> > ^ > compilation terminated. > make[1]: *** [arpd] Error 1 > make[1]: Leaving directory `/home/ubuntu/OBDscan/iproute2-j1939/misc' > make: *** [all] Error 2 > > i am getting above error. > i tried to install sudo apt-get install db4.7-util > but did not worked you'd need some sort of db-devel package, as explained in README of iprout2. Another trick I initially did is disable arpd from building: in misc/Makefile, drop 'arpd' from TARGETS. Kind regards, Kurt ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: AW: Need support J1939 protocol support to my kernel (v3.8 on BeagleBone Black) 2013-11-28 8:52 ` Kurt Van Dijck @ 2013-12-04 22:10 ` swapnil 2013-12-05 19:59 ` Kurt Van Dijck 0 siblings, 1 reply; 9+ messages in thread From: swapnil @ 2013-12-04 22:10 UTC (permalink / raw) To: linux-can Kurt Van Dijck <kurt.van.dijck <at> eia.be> writes: > > Hey, > > > Hi Kurt , > > Thanks for adding the information on elinux.org/J1939 > > I am able to build the kernal for begabone black using j1939 > > but when i tried to build iproute2-j1939 > > > > root <at> arm:/home/ubuntu/OBDscan/iproute2-j1939# make > > make[1]: Entering directory `/home/ubuntu/OBDscan/iproute2-j1939/misc' > > gcc -D_GNU_SOURCE -O2 -Wstrict-prototypes -Wall -I../include - > > DRESOLVE_HOSTNAMES -DLIBDIR=\"/usr/lib/\" -I/usr//usr/include -o arpd > > arpd.c ../lib/libnetlink.a ../lib/libutil.a -ldb -lpthread > > arpd.c:19:20: fatal error: db_185.h: No such file or directory > > #include <db_185.h> > > ^ > > compilation terminated. > > make[1]: *** [arpd] Error 1 > > make[1]: Leaving directory `/home/ubuntu/OBDscan/iproute2-j1939/misc' > > make: *** [all] Error 2 > > > > i am getting above error. > > i tried to install sudo apt-get install db4.7-util > > but did not worked > > you'd need some sort of db-devel package, as explained in README of iprout2. > Another trick I initially did is disable arpd from building: > in misc/Makefile, drop 'arpd' from TARGETS. > > Kind regards, > Kurt > -- > To unsubscribe from this list: send the line "unsubscribe linux-can" in > the body of a message to majordomo <at> vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > > Hi Kurt, I commented out arpd part from building and it worked after that i replaced current IP with the new one. After that i tried the sample program jspy it worked as expected on beaglebone black. I have few questions 1.I want to transmit more than 8 bytes on can and want to check RTS->CTS->EOM flow 2.I a able to receive 900 messages/sec on begalbone with 250k baud rate actually it should receive 1600 messages ? i will try to figure it out? I think JSR program is used to transmit and receive j1939 packet but i am not able to figure out how to send TP message can u give me simple example to send and receive J1939 tp message I have canAlyzer. Also i want to post the document how I build j1939 for begalbone black. Thank you for your support ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: AW: Need support J1939 protocol support to my kernel (v3.8 on BeagleBone Black) 2013-12-04 22:10 ` swapnil @ 2013-12-05 19:59 ` Kurt Van Dijck 0 siblings, 0 replies; 9+ messages in thread From: Kurt Van Dijck @ 2013-12-05 19:59 UTC (permalink / raw) To: swapnil; +Cc: linux-can [-- Attachment #1: Type: text/plain, Size: 1708 bytes --] 2nd try, without HTML guide Hey, > > After that i tried the sample program jspy it worked as expected on > beaglebone black. > I have few questions > 1.I want to transmit more than 8 bytes on can and want to check > RTS->CTS->EOM flow One of the major reasons of putting j1939 in kernel is that multiple transport sessions must be serialized due to CAN-id reuse. As such, userspace will not have to worry about the RTS->CTS->EOM flow. Instead, userspace does sendto() of more than 8 bytes, and kernel will activate the transport protocol (fragmentation). If kernel is able to attempt delivery of the packet, the call to sendto() succeeds, as needed for DGRAM sockets. If transport protocol aborts later in the process, userspace is not alerted. > 2.I a able to receive 900 messages/sec on begalbone with 250k baud rate > actually it should receive 1600 messages ? i will try to figure it out? You're welcome to post your throughput measurements. I'm actually very interested. > > > I think JSR program is used to transmit and receive j1939 packet > but i am not able to figure out how to send TP message > can u give me simple example to send and receive J1939 tp message sendto(sock, dat, 9, 0, &peername, sizeof(peername)); will need to engage transport protocol with 2 datapackets :-) It does not need to be difficult. I added my preliminary 'kickstarter guide to j1939' and accompanying test program. You could give it a try. I'm still looking how to host it on elinux.org. The syntax is markdown :-( > I have canAlyzer. > Also i want to post the document how I build j1939 for begalbone > black. Euhm? You mean, adding the steps on elinux.org? You're welcome. Kurt [-- Attachment #2: testj1939.c --] [-- Type: text/x-csrc, Size: 6596 bytes --] /* * Copyright (c) 2013 EIA Electronics * * Authors: * Kurt Van Dijck <kurt.van.dijck@eia.be> * * This program is free software; you can redistribute it and/or modify * it under the terms of the version 2 of the GNU General Public License * as published by the Free Software Foundation */ #include <signal.h> #include <time.h> #include <inttypes.h> #include <errno.h> #include <string.h> #include <stdlib.h> #include <stdio.h> #include <unistd.h> #include <getopt.h> #include <error.h> #include <sys/time.h> #include <sys/socket.h> #include <net/if.h> #include <linux/can.h> #include <linux/can/j1939.h> static const char help_msg[] = "testj1939: demonstrate j1939 use\n" "Usage: testj1939 FROM TO\n" " FROM / TO - or [IFACE][:[SA][,[PGN][,NAME]]]\n" "Options:\n" " -v Print relevant API calls\n" " -s[=LEN] Initial send of LEN bytes dummy data\n" " -r Receive (and print) data\n" " -e Echo incoming packets back\n" " This atually receives packets\n" " -c Issue connect()\n" " -p=PRIO Set priority to PRIO\n" " -n Emit 64bit NAMEs in output\n" "\n" "Example:\n" "testj1939 can1 20\n" "\n" ; static const char optstring[] = "?vs::rep:cn"; static void parse_canaddr(char *spec, struct sockaddr_can *paddr) { char *str; str = strsep(&spec, ":"); if (strlen(str)) paddr->can_ifindex = if_nametoindex(str); str = strsep(&spec, ","); if (str && strlen(str)) paddr->can_addr.j1939.addr = strtoul(str, NULL, 0); str = strsep(&spec, ","); if (str && strlen(str)) paddr->can_addr.j1939.pgn = strtoul(str, NULL, 0); str = strsep(&spec, ","); if (str && strlen(str)) paddr->can_addr.j1939.name = strtoul(str, NULL, 0); } static const char *canaddr2str(const struct sockaddr_can *paddr) { static char buf[128]; char *str = buf; char ifname[IF_NAMESIZE]; if (paddr->can_ifindex) str += sprintf(str, "%s", if_indextoname(paddr->can_ifindex, ifname)); *str++ = ':'; if (paddr->can_addr.j1939.addr != J1939_NO_ADDR) str += sprintf(str, "%02x", paddr->can_addr.j1939.addr); *str++ = ','; if (paddr->can_addr.j1939.pgn != J1939_NO_PGN) str += sprintf(str, "%05x", paddr->can_addr.j1939.pgn); *str++ = ','; if (paddr->can_addr.j1939.name != J1939_NO_NAME) str += sprintf(str, "%016llx", paddr->can_addr.j1939.name); *str++ = 0; return buf; } /* main */ int main(int argc, char *argv[]) { int ret, sock, opt, j, verbose; socklen_t peernamelen; struct sockaddr_can sockname = { .can_family = AF_CAN, .can_addr.j1939 = { .addr = J1939_NO_ADDR, .name = J1939_NO_NAME, .pgn = J1939_NO_PGN, }, }, peername = { .can_family = AF_CAN, .can_addr.j1939 = { .addr = J1939_NO_ADDR, .name = J1939_NO_NAME, .pgn = J1939_NO_PGN, }, }; uint8_t dat[128]; int valid_peername = 0; int todo_send = 0, todo_recv = 0, todo_echo = 0, todo_prio = -1; int todo_connect = 0, todo_names = 0; /* argument parsing */ while ((opt = getopt(argc, argv, optstring)) != -1) switch (opt) { case 'v': verbose = 1; break; case 's': todo_send = strtoul(optarg ?: "8", NULL, 0); break; case 'r': todo_recv = 1; break; case 'e': todo_echo = 1; break; case 'p': todo_prio = strtoul(optarg, NULL, 0); break; case 'c': todo_connect = 1; break; case 'n': todo_names = 1; break; default: fputs(help_msg, stderr); exit(1); break; } if (argv[optind]) { if (strcmp("-", argv[optind])) parse_canaddr(argv[optind], &sockname); ++optind; } if (argv[optind]) { if (strcmp("-", argv[optind])) { parse_canaddr(argv[optind], &peername); valid_peername = 1; } ++optind; } /* open socket */ if (verbose) fprintf(stderr, "- socket(PF_CAN, SOCK_DGRAM, CAN_J1939);\n"); sock = ret = socket(PF_CAN, SOCK_DGRAM, CAN_J1939); if (ret < 0) error(1, errno, "socket(j1939)"); if (todo_prio >= 0) { if (verbose) fprintf(stderr, "- setsockopt(, SOL_CAN_J1939, SO_J1939_SEND_PRIO, &%i);\n", todo_prio); ret = setsockopt(sock, SOL_CAN_J1939, SO_J1939_SEND_PRIO, &todo_prio, sizeof(todo_prio)); if (ret < 0) error(1, errno, "set priority %i", todo_prio); } if (verbose) fprintf(stderr, "- bind(, %s, %li);\n", canaddr2str(&sockname), sizeof(sockname)); ret = bind(sock, (void *)&sockname, sizeof(sockname)); if (ret < 0) error(1, errno, "bind()"); if (todo_connect) { if (!valid_peername) error(1, 0, "no peername supplied"); if (verbose) fprintf(stderr, "- connect(, %s, %li);\n", canaddr2str(&peername), sizeof(peername)); ret = connect(sock, (void *)&peername, sizeof(peername)); if (ret < 0) error(1, errno, "connect()"); } if (todo_send) { /* initialize test vector */ for (j = 0; j < sizeof(dat); ++j) dat[j] = ((2*j) << 4) + ((2*j+1) & 0xf); /* send data */ /* * when using connect, do not provide additional * destination information and use send() */ if (valid_peername && !todo_connect) { if (verbose) fprintf(stderr, "- sendto(, <dat>, %i, 0, %s, %li);\n", todo_send, canaddr2str(&peername), sizeof(peername)); ret = sendto(sock, dat, todo_send, 0, (void *)&peername, sizeof(peername)); } else { /* * we may do sendto(sock, dat, todo_send, 0, NULL, 0) * as well, but using send() demonstrates the API better */ if (verbose) fprintf(stderr, "- send(, <dat>, %i, 0);\n", todo_send); ret = send(sock, dat, todo_send, 0); } if (ret < 0) error(1, errno, "sendto"); } /* main loop */ if ((todo_echo || todo_recv) && verbose) fprintf(stderr, "- while (1)\n"); while (todo_echo || todo_recv) { /* * re-use peername for storing the sender's peername of * received packets */ if (verbose) fprintf(stderr, "- recvfrom(, <dat>, %li, 0, &<peername>, %li);\n", sizeof(peername), sizeof(peername)); peernamelen = sizeof(peername); ret = recvfrom(sock, dat, sizeof(dat), 0, (void *)&peername, &peernamelen); if (ret < 0) { if (EINTR == errno) { if (verbose) fprintf(stderr, "-\t<interrupted>\n"); continue; } error(1, errno, "recvfrom()"); } if (todo_echo) { if (verbose) fprintf(stderr, "- sendto(, <dat>, %i, 0, %s, %i);\n", ret, canaddr2str(&peername), peernamelen); ret = sendto(sock, dat, ret, 0, (void *)&peername, peernamelen); if (ret < 0) error(1, errno, "sendto"); } if (todo_recv) { if (todo_names && peername.can_addr.j1939.name) printf("%016llx ", peername.can_addr.j1939.name); printf("%02x %05x:", peername.can_addr.j1939.addr, peername.can_addr.j1939.pgn); for (j = 0; j < ret; ++j) printf(" %02x", dat[j]); printf("\n"); } } return 0; } [-- Attachment #3: j1939.page --] [-- Type: text/plain, Size: 7638 bytes --] # CAN-J1939 on linux ## CAN on linux See [Wikipedia:socketcan](http://en.wikipedia.org/wiki/Socketcan) ## J1939 networking in short * Add addressing on top of CAN (destination address & broadcast) * Any (max 1780) length packets. Packets of 9 or more use **Transport Protocol** (fragmentation) Such packets use different CANid for the same PGN. * only **29**bit, non-**RTR** CAN frames * CAN id is composed of * 0..8: SA (source address) * 9..26: * PDU1: PGN+DA (destionation address) * PDU2: PGN * 27..29: PRIO * SA / DA may be dynamically assigned via j1939-81 Fixed rules of precedence in Specification, no master necessary ## J1939 on SocketCAN J1939 is *just another protocol* that fits in the Berkely sockets. socket(AF_CAN, SOCK_DGRAM, CAN_J1939) ## differences from CAN_RAW ### addressing SA, DA & PGN are used, not CAN id. Berkeley socket API is used to communicate these to userspace: * SA+PGN is put in sockname ([getsockname](http://man7.org/linux/man-pages/man2/getsockname.2.html)) * DA+PGN is put in peername ([getpeername](http://man7.org/linux/man-pages/man2/getpeername.2.html)) PGN is put in both structs PRIO is a datalink property, and irrelevant for interpretation Therefore, PRIO is not in *sockname* or *peername*. The *data* that is [recv][recvfrom] or [send][sendto] is the real payload. Unlike CAN_RAW, where addressing info is data. ### Packet size J1939 handles packets of 8+ bytes with **Transport Protocol** fragmentation transparently. No fixed data size is necessary. send(sock, data, 8, 0); will emit a single CAN frame. send(sock, data, 9, 0); will use fragementation, emitting 1+ CAN frames. # Enable j1939 CAN has no protocol id field. Enabling protocols must be done manually ### netlink ip link set can0 j1939 on ### procfs for legacy kernel (2.6.25) This API is dropped for kernels with netlink support! echo can0 > /proc/net/can-j1939/net # Using J1939 ## BSD socket implementation * socket * bind / connect * recvfrom / sendto * getsockname / getpeername ## Modified *struct sockaddr_can* struct sockaddr_can { sa_family_t can_family; int can_ifindex; union { struct { __u64 name; __u32 pgn; __u8 addr; } j1939; } can_addr; } * *can_addr.j1939.pgn* is PGN * *can_addr.j1939.addr* & *can_addr.j1939.name* determine the ECU * receiving address information, *addr* is always set, *name* is set when available. * When providing address information, *name* != 0 indicates dynamic addressing ## iproute2 ### Static addressing ip addr add j1939 0x80 dev can0 ### Dynamic addressing ip addr add j1939 name 0x012345678abcdef dev can0 # Kickstart guide to j1939 on linux ## Prepare using VCAN You may skip this step entirely if you have a functional **can0** bus on your system. Load module, when *vcan* is not in-kernel modprobe vcan Create a virtual can0 device and start the device ip link add can0 type vcan ip link set can0 up ## First steps with j1939 Use [testj1939](testj1939.c) When *can-j1939* is compiled as module, load it. modprobe can-j1939 Enable the j1939 protocol stack on the CAN device ip link set can0 j1939 on Most of the subsequent examples will use 2 sockets programs (in 2 terminals). One will use CAN_J1939 sockets using *testj1939*, and the other will use CAN_RAW sockets using cansend+candump. testj1939 can be told to print the used API calls by adding **-v** program argument. ### receive without source address Do in terminal 1 ./testj1939 -r can0: Send raw CAN in terminal 2 cansend can0 1823ff40#0123 You should have this output in terminal 1 40 02300: 01 23 This means, from NAME 0, SA 40, PGN 02300 was received, with 2 databytes, *01* & *02*. now emit this CAN message: cansend can0 18234140#0123 In J1939, this means that ECU 0x40 sends directly to ECU 0x41 Since we did not bind to address 0x41, this traffic is not meant for us and *testj1939* does not receive it. ### Use source address ./testj1939 can0:0x80 will say ./testj1939: bind(): Cannot assign requested address Since J1939 maintains addressing, **0x80** has not yet been assigned as an address on **can0** . This behaviour is very similar to IP addressing: you cannot bind to an address that is not your own. Now tell the kernel that we *own* address 0x80. It will be available from now on. ip addr add j1939 0x80 dev can0 ./testj1939 can0:0x80 now succeeds. ### receive with source address Terminal 1: ./testj1939 -r can0:0x80 Terminal 2: cansend can0 18238040#0123 Will emit this output 40 02300: 01 23 This is because the traffic had destination address __0x80__ . ### send Open in terminal 1: candump -L can0 And to these test in another terminal ./testj1939 -s can0:0x80 This produces **1BFFFF80#0123456789ABCDEF** on CAN. ./testj1939 -s can0: will produce exactly the same because **0x80** is the only address currently assigned to **can0:** and is used by default. ### Multiple source addresses on 1 CAN device ip addr add j1939 0x90 dev can0 ./testj1939 -s can0:0x90 produces **1BFFFF90#0123456789ABCDEF** , ./testj1939 -s can0: still produces **1BFFFF80#0123456789ABCDEF** , since **0x80** is the default _source address_. Check ip addr show can0 emits X: can0: <NOARP,UP,LOWER_UP> mtu 16 qdisc noqueue state UNKNOWN link/can can-j1939 0x80 scope link can-j1939 0x90 scope link 0x80 is the first address on can0. ### Use specific PGN ./testj1939 -s can0:,0x12345 emits **1923FF80#0123456789ABCDEF** . Note that the real PGN is **0x12300**, and destination address is **0xff**. ### Emit destination specific packets The destination field may be set during sendto(). *testj1939* implements that like this ./testj1939 -s can0:,0x12345 can0:0x40 emits **19234080#0123456789ABCDEF** . The destination CAN iface __must__ always match the source CAN iface. Specifing one during bind is therefore sufficient. ./testj1939 -s can0:,0x12300 :0x40 emits the very same. ### Emit different PGNs using the same socket The PGN is provided in both __bind( *sockname* )__ and __sendto( *peername* )__ , and only one is used. *peername* PGN has highest precedence. For broadcasted transmissions ./testj1939 -s can0:,0x12300 :,0x32100 emits **1B21FF80#0123456789ABCDEF** rather than 1923FF80#012345678ABCDEF Desitination specific transmissions ./testj1939 -s can0:,0x12300 :0x40,0x32100 emits **1B214080#0123456789ABCDEF** . It makes sometimes sense to omit the PGN in __bind( *sockname* )__ . ### Larger packets J1939 transparently switches to *Transport Protocol* when packets do not fit into single CAN packets. ./testj1939 -s20 can0:0x80 :,0x12300 emits: 18ECFF80#20140003FF002301 18EBFF80#010123456789ABCD 18EBFF80#02EF0123456789AB 18EBFF80#03CDEF01234567 The fragments for broadcasted *Transport Protocol* are seperated __50ms__ from each other. Destination specific *Transport Protocol* applies flow control and may emit CAN packets much faster. ./testj1939 -s20 can0:0x80 :0x90,0x12300 emits: 18EC9080#1014000303002301 18EC8090#110301FFFF002301 18EB9080#010123456789ABCD 18EB9080#02EF0123456789AB 18EB9080#03CDEF01234567 18EC8090#13140003FF002301 The flow control causes a bit overhead. This overhead scales very good for larger J1939 packets. ## Advanced topics with j1939 ### Change priority of J1939 packets ./testj1939 -s can0:0x80,0x0100 ./testj1939 -s -p3 can0:0x80,0x0200 emits 1801FF80#0123456789ABCDEF 0C02FF80#0123456789ABCDEF ### using connect ### advanced filtering ## dynamic addressing ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: AW: Need support J1939 protocol support to my kernel (v3.8 on BeagleBone Black)
@ 2014-01-15 16:14 swapnil
2014-01-16 21:43 ` Kurt Van Dijck
0 siblings, 1 reply; 9+ messages in thread
From: swapnil @ 2014-01-15 16:14 UTC (permalink / raw)
To: linux-can
Hi Kurt,
Using the test code given by you I have developed an app but I am
facing some issues i don't know its j1939 limitation or the stack.
1.I want to transmit and single frame message RACM message
canId--18EAFEFF data 00EE00FFFFFFFFFFFF it asks to bind address with 0xFE
note -can i transmit single frame message like raw can from any source
address
to any dest address.
I think address claim is needed for only TP message
2. In case of multiple address claims like 0x80 and 0x90 if i want to
transmit the data from both address parallel y do i need to create different
socket for each address?
3.It don't reply for RACM message if i send from different adapter
example: 18EAFEFF#00EE00FFFFFFFFFFFF j1939 stack should reply with claimed
address Like 80 and 90
I was busy in creating the app to test the linux j1939 stack on
begalbone Black. thank you very much for the test programs those helped me
a lot
Thank
Swapnil
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: AW: Need support J1939 protocol support to my kernel (v3.8 on BeagleBone Black) 2014-01-15 16:14 swapnil @ 2014-01-16 21:43 ` Kurt Van Dijck 0 siblings, 0 replies; 9+ messages in thread From: Kurt Van Dijck @ 2014-01-16 21:43 UTC (permalink / raw) To: swapnil; +Cc: linux-can Hi Swapnil, > Using the test code given by you I have developed an app but I am > facing some issues i don't know its j1939 limitation or the stack. > > 1.I want to transmit and single frame message RACM message RACM? I don't know the RACM acronym. Can you explain a bit? > canId--18EAFEFF data 00EE00FFFFFFFFFFFF it asks to bind address with 0xFE I guess you want 18eafffe#00ee (PGN 0ea00 is per J1939 only 3 bytes long). Because FF is not allowed as source, and 0xfe is fuzzy to use as destination. Likewise, 0xfe as source is common thing to do for requesting others address claims, and FF as destination is, well, a broadcast. > > note -can i transmit single frame message like raw can from any source > address > to any dest address. That is true, but not every combination is a 'legal' withing J1939 context. > I think address claim is needed for only TP message Address claim is needed for dynamic addressing (j1939-81). This can be combined, but it really unrelated to TP (transport protocol). Your story is valid without address claims also, using static addressing. > > 2. In case of multiple address claims like 0x80 and 0x90 if i want to > transmit the data from both address parallel y do i need to create different > socket for each address? s/multiple address claims/multiple addresses/ If you want to transmit both 18000080#xxxx & 18000090#xxxx, you need a different socket for each address. Note that a single source address indicates a logical node. It's quite uncommon to combine different logical nodes into 1 program, but if you'd want to, different sockets is the way to go ... > > 3.It don't reply for RACM message if i send from different adapter > example: 18EAFEFF#00EE00FFFFFFFFFFFF j1939 stack should reply with claimed > address Like 80 and 90 I see what you mean. 18eafffe#00ee00 from a remote should be replied with the address claims. You should really consider using jacd then. That will deal with j1939-81 (1 instance for each SA). Do you need dynamic addresses? Static addresses make things a lot simpler. And static addresses allow Transport Protocol :-). > > I was busy in creating the app to test the linux j1939 stack on > begalbone Black. thank you very much for the test programs those helped me > a lot That's cool! It feels fine that it helps others. -- Kurt Van Dijck ^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2014-01-16 21:43 UTC | newest] Thread overview: 9+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2013-11-11 16:14 AW: Need support J1939 protocol support to my kernel (v3.8 on BeagleBone Black) Michael Luxen 2013-11-12 1:10 ` swapnil 2013-11-12 10:02 ` Kurt Van Dijck 2013-11-28 4:49 ` swapnil 2013-11-28 8:52 ` Kurt Van Dijck 2013-12-04 22:10 ` swapnil 2013-12-05 19:59 ` Kurt Van Dijck -- strict thread matches above, loose matches on Subject: below -- 2014-01-15 16:14 swapnil 2014-01-16 21:43 ` Kurt Van Dijck
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox