Re: BCM loopback - Information request

[Oliver Hartkopp <socketcan@hartkopp.net>]

On 09/20/2016 05:19 PM, Angeloni Andrea wrote:
> Ok clear.
>
> With local host you mean same Linux system

Yes.

"MSG_DONTROUTE: set when the received frame was created on the local host."


or same CAN interface? On the uPC I'm currently using there are two CAN 
interfaces, and it could happen that one sends data to be processed to 
the other.

When the CAN frame is physically received by the CAN controller it comes 
'from outer space' and not from the local host.

When you sent a frame with CAN ID 0x123 with the CAN_BCM on can0 and you 
have can0 and can1 connected together the received frame on can0 will 
have MSG_DONTROUTE set and the frame received from can1 will have 
MSG_DONTROUTE unset.

Regards,
Oliver

>
> Regards,
> Andrea
>
> -----Original Message-----
> From: Oliver Hartkopp [mailto:socketcan@hartkopp.net]
> Sent: martedì 20 settembre 2016 17:07
> To: Angeloni Andrea <andrea.angeloni@alfamationglobal.com>; Alexander Stein <alexander.stein@systec-electronic.com>
> Cc: linux-can@vger.kernel.org
> Subject: Re: BCM loopback - Information request
>
> Hi Andrea,
>
> the CAN_BCM sets the loopback by default:
>
> http://lxr.linux.no/#linux+v4.7.4/net/can/bcm.c#L1229
>
> So it is not configurable by now.
>
> An possibility could be to implement socket options for the CAN_BCM so that a setsockopt() with a new CAN_BCM_LOOPBACK becomes available.
>
> But this is not in mainline Linux now.
>
> Alternatively you can check for the fames you received on a CAN_RAW socket whether these frames have been created on the local host:
>
> http://lxr.linux.no/#linux+v4.7.4/Documentation/networking/can.txt#L632
>
> Which means that you have to use recvmsg() to read from CAN_RAW.
>
> 'candump' does it here:
> https://github.com/linux-can/can-utils/blob/master/candump.c#L793
>
> Regards,
> Oliver
>
> On 09/20/2016 12:19 PM, Angeloni Andrea wrote:
>> Ciao Alexander,
>>
>> thanks for your answer.
>>
>> Unfortunately, CAN_RAW_LOOPBACK is not working. Option is indeed referring to socket and not device, while BCM messages are coming from a different socket.
>>
>> Andrea
>>
>> -----Original Message-----
>> From: Alexander Stein [mailto:alexander.stein@systec-electronic.com]
>> Sent: martedì 20 settembre 2016 11:54
>> To: Angeloni Andrea <andrea.angeloni@alfamationglobal.com>
>> Cc: linux-can@vger.kernel.org
>> Subject: Re: BCM loopback - Information request
>>
>> Hi Andrea,
>>
>> On Tuesday 20 September 2016 09:42:53, Angeloni Andrea wrote:
>>> In my application there is a thread for each initialized CAN
>>> interface that sends over TCP all the read data to a Windows Client.
>>> Data are read using Socket RAW. Application is also capable to send
>>> cyclic messages using Broadcast manager.
>>>
>>> Here is my question: how can I configure my application so that
>>> reading thread does not get messages sent by BCM on the same interface?
>>
>> I think the socket option CAN_RAW_LOOPBACK is what you need. Please
>> refer to
>> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/D
>> ocumentation/networking/can.txt#n536
>> You will not receive CAN frames sent from interface (BCM and other CAN frames), thus only messages from other CAN nodes.
>>
>> HTH
>> Alexander
>>
>>
>> --
>> 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
>>
>
>