Re: candump error flags

[Oliver Hartkopp <socketcan@hartkopp.net>]

On 11/08/2014 04:34 PM, Paul Lever wrote:

>> In this case the -d option should work.
>> But just for the records: The -d option detects drops from the network layer
>> to the application which opened the socket (here: candump).
>>
>> So it basically detects problems when your user space application is not fast
>> enough to read and handle the CAN frames. It does NOT detect whether the CAN
>> driver dropped any frames.
> 
> 
> Thanks for the info. Is there an easy way to detect driver level
> packet drops or overruns?The CANBUS is at 250,000 bitrate.

When the CAN driver supports it, there's a bit in the error message

	CAN_ERR_CRTL_RX_OVERFLOW

for that.

See:
http://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/tree/include/uapi/linux/can/error.h?h=linux-3.17.y#n63

Looking for it in the tree:
http://lxr.free-electrons.com/ident?i=CAN_ERR_CRTL_RX_OVERFLOW


It is used in c_can_handle_lost_msg_obj()

http://lxr.free-electrons.com/source/drivers/net/can/c_can/c_can.c#L348

So it should also increase rx_errors and rx_over_errors.

I'm not familiar with c_can - but can you see these statistics increasing the
values in your setup?

ip -stats link show can0

should display them.

Regards,
Oliver