Re: [Socketcan-users] real-time transmission?

Re: [Socketcan-users] real-time transmission?

[Marc Kleine-Budde]

[-- Attachment #1: Type: text/plain, Size: 1128 bytes --]

Hello,

please come to the linux-can mailinglist.

On 09/03/2013 04:31 PM, Aiken, Scott wrote:
> Hi all,
> 
>  
> 
> I’m porting an old implementation of CAL/CiA to Linux.  The device
> connection protocol has some pretty tight timing.  I’ve observed that
> some CAN packets are buffered for 500 ms between the socket write() call
> and the actual transmission of the packet.  This is enough to sabotage
> the connection process, and possibly take out other devices as well. 
> Are there any socket options to disable buffering of write packets, and
> send them when write() is called? 
> 
>  
> 
> Thanks,
> 
> Scott
> 
>  
> 
> 
> 
> _______________________________________________
> Socketcan-users mailing list
> Socketcan-users@lists.berlios.de
> https://lists.berlios.de/mailman/listinfo/socketcan-users
> 


-- 
Pengutronix e.K.                  | Marc Kleine-Budde           |
Industrial Linux Solutions        | Phone: +49-231-2826-924     |
Vertretung West/Dortmund          | Fax:   +49-5121-206917-5555 |
Amtsgericht Hildesheim, HRA 2686  | http://www.pengutronix.de   |


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 259 bytes --]

Re: real-time transmission?

[Marc Kleine-Budde]

[-- Attachment #1: Type: text/plain, Size: 1224 bytes --]

On 09/03/2013 04:56 PM, Marc Kleine-Budde wrote:
>> I’m porting an old implementation of CAL/CiA to Linux.  The device
>> connection protocol has some pretty tight timing.  I’ve observed that
>> some CAN packets are buffered for 500 ms between the socket write() call
>> and the actual transmission of the packet.  This is enough to sabotage
>> the connection process, and possibly take out other devices as well. 
>> Are there any socket options to disable buffering of write packets, and
>> send them when write() is called? 

Depends who introduces that big delay. If you're sending too much CAN
traffic on your system you can use traffic shaping to prioritize your
communication. It there are other interrupts in the system you may want
to use PREEMPT_RT and play with the interrupt and net-rx/tx softirq
priorities.

I think the first step should be measurements to figure out who
introduces the delay.

Regards,
Marc

-- 
Pengutronix e.K.                  | Marc Kleine-Budde           |
Industrial Linux Solutions        | Phone: +49-231-2826-924     |
Vertretung West/Dortmund          | Fax:   +49-5121-206917-5555 |
Amtsgericht Hildesheim, HRA 2686  | http://www.pengutronix.de   |


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 259 bytes --]