CAN Problems on BBB rev C

GoogleGroups
1

I am working with a BeagleBone Black rev C for a project using CAN. I am using the CBB-Serial-r02 cape because we need the UART features too. The problem is that CAN is not working. I did a lot of things to test it on a 3.8-rt kernel from RobertCNelson (which is only PREEMPT), but it didn’t worked at all. Some things that I did:

  1. Always modprobe the can modules (can, can-dev, can-raw)
  2. Compiled and installed canutils, always using the “cangen can0” to test the can output
  3. Always run “ip link set can0 up type can bitrate 125000; ifconfig can0 up” to activate can0
  4. Installed the CBB-Serial-r02 dtbs from the official repository, even that the debian image that I used already includes it
  5. Test the output with an osciloscope. I did the same test with an EzDSP 28335 and it worked, so it is not a problem with this part of the method
  6. I tried to do a candump can0 with the DSP generating random data, but the candump didn’t show any data

I recompiled 3.14 kernel using the dtb-rebuilder (instructions on this tutorial: https://groups.google.com/d/msg/beagleboard/_9u1B6ZkgCU/K2ARgwfC490J) and enabling the DCAN1 (because DCAN0 disables the I2C used for capemgr) and it didn’t worked too.

Well, I ran out of options here. Some one is having problems with the rev C and CAN too or that is some problem with my method?

2

No reason to double post to the group. :wink:

can0/can1 works on 3.14/4.1.x

Right now i'd recommend you use v4.1.x and utilze cape overlays..

Step 1: update kernel:

sudo apt-get update
sudo apt-get install linux-image-4.1.0-rc6-bone5
sudo reboot

Step 2: clone overlay repo:

git clone https://github.com/beagleboard/bb.org-overlays
cd ./bb.org-overlays

Update dtc:
./dtc-overlay.sh

Install overlays:
./install.sh

reboot..

the CBB-Serial-r02 cape should be auto-detected.. if not, please reply with:

dmesg | grep cape

So i can add it..

Regards,

3

According to the Logic Supply serial cape manual . . .

Important:

You cannot just test the CAN Bus software without connecting the CBB-Serial hardware to another CAN device. You need to be connected to a receiving CAN device so that ACK (acknowledge) packets get sent and the CAN sender can mark it as sent (i.e. stop buffering it), otherwise the buffer gets full after a short period of time and all future writes to the CAN Bus will fail.

We’ve pretty much duplicated the exact steps you’ve done with the exception of connecting the beaglebone + serial cape to an external CAN device. We’ve also tested kernels 3.8.x, 3.14.x, and 4.1.x( most recently ). They all work. Most curious though . .

I am working with a BeagleBone Black rev C for a project using CAN. I am using the CBB-Serial-r02 cape because we need the UART features too

We’re using . . .

debian@192.168.7.2s password:
Last login: Mon May 11 20:23:49 2015 from 192.168.7.1
debian@beaglebone:~$ cat /sys/devices/platform/bone_capemgr/slots
0: 54:P—L cape-CBB-Serial,r01,Logic Supply,cape-CBB-Serial
1: 55:PF—
2: 56:PF—
3: 57:PF—

Curious because you say you’re using an r2, where we’re using and r1. Is this a Logic Supply board ?

4

Well, testing it with the oscilloscope, following the Robert Nelson guide, it didn’t worked. It is very strange that nothing is outputted while I am probing the CANH/CANL/GND channels, as I did with the DSP (with the DSP it worked). I am going to try to find some CAN device here to test it, but I still think that is strange.

Meanwhile, as your setup is almost the same of mine, could you run the cangen and test the output (canh + gnd or canl + gnd) with an oscilloscope to see if it transmits something without any CAN device?

PS: the buffer didn’t got full when without CAN devices, as the manual said that would happen. Although, in earlier tests, when I was without the cape (I tested right off the DCAN1 pins) this problem happened.

5

My “dmesg | grep cape” seems right

[ 3.373227] bone_capemgr bone_capemgr: Baseboard: ‘A335BNLT,000C,4014BBBK3429’
[ 3.373249] bone_capemgr bone_capemgr: compatible-baseboard=ti,beaglebone-black
[ 3.433139] bone_capemgr bone_capemgr: slot #0: No cape found
[ 3.493138] bone_capemgr bone_capemgr: slot #1: No cape found
[ 3.553156] bone_capemgr bone_capemgr: slot #2: No cape found
[ 3.583214] bone_capemgr bone_capemgr: slot #3: ‘cape-CBB-Serial,r01,Logic Supply,cape-CBB-Serial’
[ 3.583478] bone_capemgr bone_capemgr: initialized OK.
[ 3.602192] bone_capemgr bone_capemgr: slot #3: dtbo ‘cape-CBB-Serial-r01.dtbo’ loaded; overlay id #0

and my “lsmod” too

can_raw 5383 1
can 26347 1 can_raw
omap_sham 18661 0
omap_aes 12035 0
usb_f_acm 6692 1
u_serial 9442 3 usb_f_acm
usb_f_ecm 8898 1
g_multi 5722 0
usb_f_mass_storage 40017 2 g_multi
usb_f_rndis 21091 2 g_multi
u_ether 11080 3 usb_f_ecm,usb_f_rndis,g_multi
libcomposite 43102 5 usb_f_acm,usb_f_ecm,usb_f_rndis,g_multi,usb_f_mass_storage
omap_rng 4188 0
c_can_platform 6169 0
rng_core 6965 1 omap_rng
c_can 9105 1 c_can_platform
can_dev 11344 1 c_can
uio_pdrv_genirq 3169 0
uio 8006 1 uio_pdrv_genirq

My “/sys/kernel/debug/pinctrl/44e10800.pinmux/pingroups”

registered pin groups:
group: pinmux_clkout2_pin
pin 109 (44e109b4.0)

group: pinmux_uart0_pins
pin 92 (44e10970.0)
pin 93 (44e10974.0)

group: nxp_hdmi_bonelt_pins
pin 108 (44e109b0.0)
pin 40 (44e108a0.0)
pin 41 (44e108a4.0)
pin 42 (44e108a8.0)
pin 43 (44e108ac.0)
pin 44 (44e108b0.0)
pin 45 (44e108b4.0)
pin 46 (44e108b8.0)
pin 47 (44e108bc.0)
pin 48 (44e108c0.0)
pin 49 (44e108c4.0)
pin 50 (44e108c8.0)
pin 51 (44e108cc.0)
pin 52 (44e108d0.0)
pin 53 (44e108d4.0)
pin 54 (44e108d8.0)
pin 55 (44e108dc.0)
pin 56 (44e108e0.0)
pin 57 (44e108e4.0)
pin 58 (44e108e8.0)
pin 59 (44e108ec.0)

group: nxp_hdmi_bonelt_off_pins
pin 108 (44e109b0.0)

group: pinmux_mmc1_pins
pin 88 (44e10960.0)

group: pinmux_emmc_pins
pin 32 (44e10880.0)
pin 33 (44e10884.0)
pin 0 (44e10800.0)
pin 1 (44e10804.0)
pin 2 (44e10808.0)
pin 3 (44e1080c.0)
pin 4 (44e10810.0)
pin 5 (44e10814.0)
pin 6 (44e10818.0)
pin 7 (44e1081c.0)

group: user_leds_s0
pin 21 (44e10854.0)
pin 22 (44e10858.0)
pin 23 (44e1085c.0)
pin 24 (44e10860.0)

group: pinmux_i2c0_pins
pin 98 (44e10988.0)
pin 99 (44e1098c.0)

group: pinmux_i2c2_pins
pin 94 (44e10978.0)
pin 95 (44e1097c.0)

group: nxp_hdmi_bonelt_pins
pin 108 (44e109b0.0)
pin 40 (44e108a0.0)
pin 41 (44e108a4.0)
pin 42 (44e108a8.0)
pin 43 (44e108ac.0)
pin 44 (44e108b0.0)
pin 45 (44e108b4.0)
pin 46 (44e108b8.0)
pin 47 (44e108bc.0)
pin 48 (44e108c0.0)
pin 49 (44e108c4.0)
pin 50 (44e108c8.0)
pin 51 (44e108cc.0)
pin 52 (44e108d0.0)
pin 53 (44e108d4.0)
pin 54 (44e108d8.0)
pin 55 (44e108dc.0)
pin 56 (44e108e0.0)
pin 57 (44e108e4.0)
pin 58 (44e108e8.0)
pin 59 (44e108ec.0)

group: nxp_hdmi_bonelt_off_pins
pin 108 (44e109b0.0)

group: cpsw_default
pin 66 (44e10908.0)
pin 67 (44e1090c.0)
pin 68 (44e10910.0)
pin 69 (44e10914.0)
pin 70 (44e10918.0)
pin 71 (44e1091c.0)
pin 72 (44e10920.0)
pin 73 (44e10924.0)
pin 74 (44e10928.0)
pin 75 (44e1092c.0)
pin 76 (44e10930.0)
pin 77 (44e10934.0)
pin 78 (44e10938.0)
pin 79 (44e1093c.0)
pin 80 (44e10940.0)

group: cpsw_sleep
pin 66 (44e10908.0)
pin 67 (44e1090c.0)
pin 68 (44e10910.0)
pin 69 (44e10914.0)
pin 70 (44e10918.0)
pin 71 (44e1091c.0)
pin 72 (44e10920.0)
pin 73 (44e10924.0)
pin 74 (44e10928.0)
pin 75 (44e1092c.0)
pin 76 (44e10930.0)
pin 77 (44e10934.0)
pin 78 (44e10938.0)
pin 79 (44e1093c.0)
pin 80 (44e10940.0)

group: davinci_mdio_default
pin 82 (44e10948.0)
pin 83 (44e1094c.0)

group: davinci_mdio_sleep
pin 82 (44e10948.0)
pin 83 (44e1094c.0)

group: pinmux_bb_uart2_pins
pin 85 (44e10954.0)
pin 84 (44e10950.0)

group: pinmux_bb_uart4_pins
pin 28 (44e10870.0)
pin 29 (44e10874.0)
pin 16 (44e10840.0)
pin 34 (44e10888.0)

group: pinmux_dcan1_pins
pin 97 (44e10984.0)
pin 96 (44e10980.0)

6

Meanwhile, as your setup is almost the same of mine, could you run the cangen and test the output (canh + gnd or canl + gnd) with an oscilloscope to see if it transmits something without any CAN device?

No can do. The board is hooked up to an external CAN device. Logging data.

7

Not much good to play with can with no other can bus things on the bus.
If your writing
software use Vcan and simulate your can device so you can get a jump
programming your application
until you have a device on your bus.

8

I ran “candump can0” in one BeagleBone and, at the other BBB, “cangen can0 -D 11223344DEADBEEF -L 8”. Both BBB had the same disk image (I just changed the device IPs) and use the same cbb cape model. Each cape is connected with a cutted ethernet cable, so the only thing that is not in the bus is the 120 ohms parallel resistor. Well, after the commands it, didn’t worked (no logs or some output on osciloscope).

ifconfig log:

############### candump device:
can0 Link encap:UNSPEC HWaddr 00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00
UP RUNNING NOARP MTU:16 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:10
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
Interrupt:182

############### cangen device:
can0 Link encap:UNSPEC HWaddr 00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00
UP NOARP MTU:16 Metric:1
RX packets:3 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:10
RX bytes:24 (24.0 B) TX bytes:0 (0.0 B)
Interrupt:182

The strangest thing was the “cat /proc/net/can/stats”. When I was running the “cangen device” it increases TXF, but the “candump device” change nothing. I can’t figure out what could be happening. The stats logs:

############### cangen device:
5363 transmitted frames (TXF) # This increases
3 received frames (RXF)
0 matched frames (RXMF)

0 % total match ratio (RXMR)
5 frames/s total tx rate (TXR)
0 frames/s total rx rate (RXR)

0 % current match ratio (CRXMR)
5 frames/s current tx rate (CTXR)
0 frames/s current rx rate (CRXR)

0 % max match ratio (MRXMR)
5 frames/s max tx rate (MTXR)
3 frames/s max rx rate (MRXR)

0 current receive list entries (CRCV)
0 maximum receive list entries (MRCV)

############### candump device:
0 transmitted frames (TXF)
0 received frames (RXF)
0 matched frames (RXMF)

0 % total match ratio (RXMR)
0 frames/s total tx rate (TXR)
0 frames/s total rx rate (RXR)

0 % current match ratio (CRXMR)
0 frames/s current tx rate (CTXR)
0 frames/s current rx rate (CRXR)

0 % max match ratio (MRXMR)
0 frames/s max tx rate (MTXR)
0 frames/s max rx rate (MRXR)

1 current receive list entries (CRCV)
1 maximum receive list entries (MRCV)

9

I checked that when I run any “cansend” or “cangen” the interface go to BUS_OFF state (with ip -details link show can0). I tried to install libsocketcan, but it didn’t worked too. How I check if my kernel is configured rightly for SocketCAN?

10

I put BBB to send data via DCAN_TX but the problem is that the CBB-Cape don’t output nothing (and I have other BBB with the same cape, running candump). To make the BBB send something I used the RobertCNelson guide (above) and created this script:

#!/bin/bash

wget http://www.pengutronix.de/software/libsocketcan/download/libsocketcan-0.0.10.tar.bz2
tar xvjf libsocketcan-0.0.10.tar.bz2
cd libsocketcan-0.0.10
./autogen.sh
./configure --prefix=/usr
make
make install

wget http://pengutronix.de/software/socket-can/download/canutils/v4.0/canutils-4.0.6.tar.bz2
tar xvjf canutils-4.0.6.tar.bz2
cd canutils-4.0.6
./autogen.sh
./configure --prefix=/usr
make
make install

It installed the most “recent” version of both. The canconfig was:
canconfig can0 bitrate 1000000 ctrlmode triple-sampling on loopback off listen-only off restart-ms 50

Without the restart-ms the BBB DCAN always go to BUS-OFF state. In loopback tests it didn’t happen.

11

After the BBB CANTX worked (as I reported above), I tried some other things to discover why the transceiver was not generating any output. The main issue which was causing trouble is that the CBB needs an 5V power source connected to the BeagleBone. Now all is working fine and I did some scripts and a small guide for other devs, available at https://github.com/brnluiz/bbb-can-installer.

Strange that 5V power source requirement was not alerted at the CBB-Serial cape manual. I discovered that checking the electrical diagram, which took time and could be alerted by the manufacturer.

Other thing is that I tested the device without any CAN Device or 120 ohms resistor at the CANH and CANL and it worked (I used an oscilloscope to check it). The cape already provides a 120 resistor (R10). After that I tried with some CAN devices here and it worked too.

12

This had me for the whole morning. I solved it eventually by reading the schematic - thanks for making that available.
I’ve posted a message to the maintainer of the google doc with a suggestion on how to make the setup clearer with regard to the necessity of either 5v barrel jack or J4.