detect auto baud rate on can bus in linux driver.

Post by Ratheendran
Hi All,
Need you help in understanding detecting auto baud rate of CAN bus.
please share your valuable experience.
I am using imx6 processor which has flexcan hardware driver
implemented as socketcan stack.

I've used auto baud rate detection with the Microchip MCP2515 CAN Controller
connected to a PIC controller, but not with Linux, as yet. I'm working on my
own CAN Bus network so have control of all the nodes. There is a "Ping"
Protocol built into the nodes so that if there is no network traffic in a
certain period of time a random node will fire a "Ping" frame onto the CAN
Bus. That way I'm guaranteed that there will be network traffic to listen to.

In addition to network traffic my network uses a fixed number of baud rates:

10K, 20K, 50K, 125K, 250K, 500, 800K, and 1M

Now given all that on the network I simply set the MCP2515 into listen only
mode at a baud rate of 10K. If the Ping message is set for 10 seconds and I
listen for 30 Seconds I should at least get 3 CAN Frames. If I've got the
wrong baud rate I'll get MERRE errors in listen mode. My driver simply, when
in listen mode, counts received good frames and errors. If after 30 seconds
I've got good frames and no errors then I've found the baud rate being used by
the network and I can drop out of listen mode. If there are any errors then
simply move up to the next baud rate and start listening again.

This all works for me as I'm assured of traffic and have a limited number of
baud rates to search through.

I keep meaning to look at this on Linux but I've been distracted with other
work. I'll get back to it. Not sure how it would fit into the Linux can
architecture. The Driver code would have to count errors and userspace would
have to look at the results and change baud rate, or connect.

Hope that's of some help.

Post by Ratheendran
With Regards,
Ratheendran
--
To unsubscribe from this list: send the line "unsubscribe linux-can" in
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 ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Kurt Van Dijck

2014-04-27 18:01:35 UTC

Hey,

Post by Ratheendran
Hi All,
Need you help in understanding detecting auto baud rate of CAN bus.
please share your valuable experience.

The baudrate detection principle is: configure your CAN chip, and if you receive
something, then the baudrate was good. This can only work when there is traffic.
The CAN chip should be set in a mode that does not disturb the bus.
This mode is called 'listen-only' for CAN.

Be sure to read
http://www.can-cia.org/fileadmin/cia/files/icc/9/koppe.pdf
I did only read it quickly, and it appears most items have been explained
a bit.

John Whitmore explained his scenario very well.
I just want to add some comments with special regard to linux.

Post by Ratheendran
I am using imx6 processor which has flexcan hardware driver
implemented as socketcan stack.

Theoretically/ideally, all linux-can drivers behave the same.
Practically, a required feature for baudrate detection on CAN is
the listen-only mode. linux-can provides the API, but not all
drivers/devices support it.
Look into the code for CAN_CTRLMODE_LISTENONLY. My 3.12 tree flexcan driver
seems to support it, but I clearly did not test with real hardware.

Post by John Whitmore
I've used auto baud rate detection with the Microchip MCP2515 CAN Controller
connected to a PIC controller, but not with Linux, as yet.
I'm working on my
own CAN Bus network so have control of all the nodes. There is a "Ping"
Protocol built into the nodes so that if there is no network traffic in a
certain period of time a random node will fire a "Ping" frame onto the CAN
Bus. That way I'm guaranteed that there will be network traffic to listen to.
10K, 20K, 50K, 125K, 250K, 500, 800K, and 1M
Now given all that on the network I simply set the MCP2515 into listen only
mode at a baud rate of 10K. If the Ping message is set for 10 seconds and I
listen for 30 Seconds I should at least get 3 CAN Frames. If I've got the
wrong baud rate I'll get MERRE errors in listen mode. My driver simply, when
in listen mode, counts received good frames and errors. If after 30 seconds
I've got good frames and no errors then I've found the baud rate being used by
the network and I can drop out of listen mode. If there are any errors then
simply move up to the next baud rate and start listening again.
This all works for me as I'm assured of traffic and have a limited number of
baud rates to search through.
I keep meaning to look at this on Linux but I've been distracted with other
work. I'll get back to it. Not sure how it would fit into the Linux can
architecture.

Autobaud is, IMO, a userspace thing.
The principle has been explained, and with iproute2 & candump piping into
grep or wc, you could even make a shell script for that.

Post by John Whitmore
The Driver code would have to count errors and userspace would
have to look at the results and change baud rate, or connect.
Hope that's of some help.

Post by Ratheendran
With Regards,
Ratheendran

Kurt

--
To unsubscribe from this list: send the line "unsubscribe linux-can" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Ratheendran S. Rajeshwaran

2014-04-29 08:33:59 UTC

Thanks John , Kurt for your valuable support.

I am going ahead with the shell script & iproute2 supported tool testing a set of fixed baud rate from table.

My observations with the existing imx6 flexcan hardware running linux 3.0.x.

Two imx6 board ,both are enabled for FLEXCAN communication.
CAN dev bring up command with baud rate.
command -> ip link set can0 up type can bitrate 125000
one board run cansend while other end candump--e command

Case 1 both imx6 operate on same CAN baudrate
cansend can0 -i0x100 11 22 33 44
on candump terminal I receive <0x100> [4] 0b 16 21 2c

I do ip -details -statistics link show can0
2: can0: <NOARP,UP,LOWER_UP,ECHO> mtu 16 qdisc pfifo_fast state UNKNOWN mode DEFAULT qlen 10
link/can
can state ERROR-ACTIVE (berr-counter tx 0 rx 0) restart-ms 0
bitrate 250000 sample-point 0.866
tq 266 prop-seg 6 phase-seg1 6 phase-seg2 2 sjw 1
flexcan: tseg1 4..16 tseg2 2..8 sjw 1..4 brp 1..256 brp-inc 1
clock 30000000
re-started bus-errors arbit-lost error-warn error-pass bus-off
0 0 0 5 0 0
RX: bytes packets errors dropped overrun mcast
40 5 0 0 0 0
TX: bytes packets errors dropped carrier collsns
32 8 0 2 0 0

Case 2: imx6 CAN on different baud rate.
received packets( I am not able to interpret the below packet)
<0x004> [8] 00 04 00 00 00 00 00 00
ip -details -statistics link show can0
2: can0: <NOARP,UP,LOWER_UP,ECHO> mtu 16 qdisc pfifo_fast state UNKNOWN mode DEFAULT qlen 10
link/can
can state ERROR-WARNING (berr-counter tx 0 rx 128) restart-ms 0
bitrate 250000 sample-point 0.866
tq 266 prop-seg 6 phase-seg1 6 phase-seg2 2 sjw 1
flexcan: tseg1 4..16 tseg2 2..8 sjw 1..4 brp 1..256 brp-inc 1
clock 30000000
re-started bus-errors arbit-lost error-warn error-pass bus-off
0 0 0 14 1 0
RX: bytes packets errors dropped overrun mcast
144 22 0 0 0 0
TX: bytes packets errors dropped carrier collsns
0 0 0 0 0 0

I am wondering a shell script setting the different baud rate and then checking the can status to decide the correct baud rate.
our custom application need two can bus,one raw and other j1939.in case of j1939 I will use iproute2-j1939 and can-utility for j1939.

is there any better approach to do it,I am open for better Ideas.

Regards,
Ratheendran

________________________________________
From: Kurt Van Dijck <***@vandijck-laurijssen.be>
Sent: 27 April 2014 23:31
To: Ratheendran S. Rajeshwaran
Cc: John Whitmore; linux-***@vger.kernel.org
Subject: Re: detect auto baud rate on can bus in linux driver.

Hey,

Post by Ratheendran
Hi All,
Need you help in understanding detecting auto baud rate of CAN bus.
please share your valuable experience.

Post by Ratheendran
I am using imx6 processor which has flexcan hardware driver
implemented as socketcan stack.

Autobaud is, IMO, a userspace thing.
The principle has been explained, and with iproute2 & candump piping into
grep or wc, you could even make a shell script for that.

Post by John Whitmore
The Driver code would have to count errors and userspace would
have to look at the results and change baud rate, or connect.
Hope that's of some help.

Post by Ratheendran
With Regards,
Ratheendran

Kurt

--
To unsubscribe from this list: send the line "unsubscribe linux-can" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Ratheendran S. Rajeshwaran

2014-04-29 08:35:15 UTC

Post by Ratheendran
Hi All,
Need you help in understanding detecting auto baud rate of CAN bus.
please share your valuable experience.

Post by Ratheendran
I am using imx6 processor which has flexcan hardware driver
implemented as socketcan stack.

Autobaud is, IMO, a userspace thing.
The principle has been explained, and with iproute2 & candump piping into
grep or wc, you could even make a shell script for that.

Post by John Whitmore
The Driver code would have to count errors and userspace would
have to look at the results and change baud rate, or connect.
Hope that's of some help.

Post by Ratheendran
With Regards,
Ratheendran

Kurt

--
To unsubscribe from this list: send the line "unsubscribe linux-can" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Robert Schwebel

2014-04-29 19:58:41 UTC

Post by Ratheendran S. Rajeshwaran
My observations with the existing imx6 flexcan hardware running linux 3.0.x.

Could you retry with 3.14? There has been a lot of activity recently.

rsc

--
Pengutronix e.K. | |
Industrial Linux Solutions | http://www.pengutronix.de/ |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
--
To unsubscribe from this list: send the line "unsubscribe linux-can" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Kurt Van Dijck

2014-04-30 14:41:24 UTC

Hey,

Thanks again, I Appreciate your support, this has helped my understanding
,build confidence on socketcan.

[...]

Hi kurt,
I am not able to set the timeout scenario for candump command, do i need to
send a killall signal to candump after a wait period.

I believe this approach is a good one, but the timeout thing is the tricky part.
I'll take a look next few days.

Kurt

With regards,
Ratheendran

--
To unsubscribe from this list: send the line "unsubscribe linux-can" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Oliver Hartkopp

2014-04-30 16:35:12 UTC

Post by Kurt Van Dijck
Hey,

Thanks again, I Appreciate your support, this has helped my understanding
,build confidence on socketcan.

[...]

Hi kurt,
I am not able to set the timeout scenario for candump command, do i need to
send a killall signal to candump after a wait period.

I believe this approach is a good one, but the timeout thing is the tricky part.
I'll take a look next few days.

There is already an option '-T <msecs>' for this in the latest candump on the
latest gitorious can-utils git:

Usage: candump [options] <CAN interface>+
(use CTRL-C to terminate candump)

Options: -t <type> (timestamp: (a)bsolute/(d)elta/(z)ero/(A)bsolute w date)
-c (increment color mode level)
-i (binary output - may exceed 80 chars/line)
-a (enable additional ASCII output)
-S (swap byte order in printed CAN data[] - marked with '`' )
-s <level> (silent mode - 0: off (default) 1: animation 2: silent)
-b <can> (bridge mode - send received frames to <can>)
-B <can> (bridge mode - like '-b' with disabled loopback)
-u <usecs> (delay bridge forwarding by <usecs> microseconds)
-l (log CAN-frames into file. Sets '-s 2' by default)
-L (use log file format on stdout)
-n <count> (terminate after receiption of <count> CAN frames)
-r <size> (set socket receive buffer to <size>)
-d (monitor dropped CAN frames)
-e (dump CAN error frames in human-readable format)
-x (print extra message infos, rx/tx brs esi)
-T <msecs> (terminate after <msecs> without any reception)

So

candump -T 1000 -n 4 can0 | wc -l

would produce values from 1 to 4 when there was some traffic within one
second of reception or zero when not traffic was detected.

if [ `candump -T 1000 -n 4 can0 | wc -l` -ne 0 ]; then

echo "traffic";
else
echo "nothing";
fi

So updating your can-utils helps :-)

Regards,
Oliver

--
To unsubscribe from this list: send the line "unsubscribe linux-can" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Kurt Van Dijck

2014-05-01 18:49:57 UTC

Hey Oliver,

Post by Oliver Hartkopp

Hey,

[...]

Post by Oliver Hartkopp
So
candump -T 1000 -n 4 can0 | wc -l
would produce values from 1 to 4 when there was some traffic within one
second of reception or zero when not traffic was detected.
if [ `candump -T 1000 -n 4 can0 | wc -l` -ne 0 ]; then
echo "traffic";
else
echo "nothing";
fi

4 means here then "return success after 4 frames, and don't keep waiting 1sec"?
This 'if' line is the right one to use!

Thanks,
Kurt
--
To unsubscribe from this list: send the line "unsubscribe linux-can" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Kurt Van Dijck

2014-05-02 11:00:07 UTC

Hey,

Hey=C2=A0 Kurt,
=20
I just looked into linux-can/can-j1939-utils which I am plan to use o=

n my final

target, for this source candump -T < > option is missing.

That is very likely, as the can-j1939-utils forked some time ago,
and the -T option is recent.
You could merge both using git, or, since can-j1939-utils does not
really touch candump, you could just copy the newer candump.c into
your can-j1399-utils directory, overwriting the old candump.c.

Kind regards,
Kurt
--
To unsubscribe from this list: send the line "unsubscribe linux-can" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Heinz-Jürgen Oertel

2014-04-29 12:47:19 UTC

Post by Kurt Van Dijck
Autobaud is, IMO, a userspace thing.
The principle has been explained, and with iproute2 & candump piping into
grep or wc, you could even make a shell script for that.

In can4linux it can be done in user space, like

int do_autobaud(can_fd)
{
int bitrate[] = {10, 20, 50, 125, 250, 500, 800, 1000 , 0};
int i;
int ret;
can_statuspar_t status;

printf("\tswitch into listen only mode\n");
set_lomode(can_fd, 1);
/* for all possible bit rates */
for (i = 0; ; i++) {
if (bitrate[i] == 0) {
ret = 0;
break;
}
printf("\tTry %d Kbit/s\n", bitrate[i]);
set_bitrate(can_fd, bitrate[i]);
/* wait one second */
sleep(1);
ioctl(can_fd, CAN_IOCTL_STATUS, &status);
/* printf("Rx buffer: %d\n", status.rx_buffer_used); */
if (status.rx_buffer_used > 1) {
printf("\t\tCould receive frames. Switch back to
active mode\n");
ret = bitrate[i];
break;
}
}
if (ret > 0) {
/* bit rate detected */
set_lomode(can_fd, 0);
}
return ret;
}

Of course the CAN controller should have a 'listen only' mode.

On a running EnergyBus it looks like:

using CAN device /dev/can1
doing auto bit rate detection now
take care that traffic is on the bus
switch into listen only mode
Try 10 Kbit/s
Try 20 Kbit/s
Try 50 Kbit/s
Try 125 Kbit/s
Try 250 Kbit/s
Could receive frames. Switch back to active mode

--
Mit freundlichen Grüßen

Heinz-Jürgen Oertel

-------------------------------------------------------------------------

emtas - your embedded solution partner

Heinz-Jürgen Oertel

2014-04-29 14:12:51 UTC

this mode, transmission is disabled, all error counters are frozen and
the module operates
in a CAN Error Passive mode. Only messages *acknowledged by another CAN
station* will be received.

In this case always two nodes already communicating are needed in order for
the node doing the auto detection.

If FLEXCAN detects a message that has not been acknowledged, it will
flag a BIT0 error (without changing the REC), as if it was trying to
acknowledge the message."

Is this BIT0 somehow exported to user space by SocketCAN to get the
information
"received valid (+ CRC) message but without seeing an ACK" ?

--
Mit freundlichen Grüßen

Heinz-Jürgen Oertel

-------------------------------------------------------------------------

emtas - your embedded solution partner

Kurt Van Dijck

2014-04-29 13:33:55 UTC

Thanks John , Kurt Van Dijck for your valuable support.
=20
I am going ahead with the shell script & iproute2 supported tool test=

ing a set

of fixed baud rate from table.
=20
My observations with the existing imx6 flexcan hardware running linux=

3.0.x.

=20
Two imx6 board ,both are enabled for FLEXCAN=C2=A0 communication.
CAN dev bring up command with baud rate.
command -> ip link set can0 up type can bitrate 125000
one board run cansend while other end candump--e command

The 'autobaud' node must do
$ ip link set canX type can listen-only on

=20
=20
Case 1 both imx6 operate on same CAN baudrate
=C2=A0 cansend can0 -i0x100 11 22 33 44
=C2=A0 on candump terminal I receive=C2=A0 <0x100> [4] 0b 16 21 2c

Do you use the latest tools from gitorious?
The syntax of cansend and the output of candump look different.

[...]

=C2=A0 Case 2: imx6 CAN on different baud rate.=C2=A0
=C2=A0received packets( I am not able to interpret the below packet)
=C2=A0<0x004> [8] 00 04 00 00 00 00 00 00=C2=A0

Well, this is not ok.
I'd expect an error or status frame, but I don't see it in your candump=
output.
I do notice that the received frame has 8 bytes, and only 1 bit set.
So, if I suppose that the CAN_ERR_FLAG has been set,
I see (with include/linux/can/error.h aside)
<0x004> Controller problem, info in data[1]

data[1] 04: Reached warning level for RX errors.

I'd expect something like that.
Wolfgang Grandegger may explain more, as he did author these.

Next step is to make a script that does all that, and that makes candum=
p
stop after a limited time (few seconds).
I would suggest to skip a baudrate if you receive any error frame, so t=
he
interpretation may be not necessary.

Kurt
--
To unsubscribe from this list: send the line "unsubscribe linux-can" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Wolfgang Grandegger

2014-04-29 14:12:02 UTC

On Tue, 29 Apr 2014 15:33:55 +0200, Kurt Van Dijck

Thanks John , Kurt Van Dijck for your valuable support.
=20
I am going ahead with the shell script & iproute2 supported tool

testing

a set
of fixed baud rate from table.
=20
My observations with the existing imx6 flexcan hardware running linu=

3.0.x.
=20
Two imx6 board ,both are enabled for FLEXCAN=C2=A0 communication.
CAN dev bring up command with baud rate.
command -> ip link set can0 up type can bitrate 125000

=46or bus error reporting the option "berr-reporting on" needs to
be added. Otherwise only CAN error state changes will be send.