Nadym Salem wrote:
> On Mon, July 30, 2007 13:09, Jan Kiszka wrote:
>> Nadym Salem wrote:
>>> sorry, reply and reply-all is to close together ;)
>>>
>>> On Mon, July 30, 2007 12:35, Jan Kiszka wrote:
>>>> As you want to connect two applications locally, there is no Ethernet
>>>> involved, of course. Only rt_loopback. And as you are using the _raw_
>>>> access, you have to do _raw_ addressing as well, which includes to
>>>> specify the outgoing device. In the local case, this is no longer some
>>>> rtethX, but rtlo. Thus, some awareness is required in your application,
>>>> but that is likely already needed for the addressing via MAC, isn't it?
>>> That's what I originally intented to do. Changing the addressing and the
>>> interfaces inside my applications is not a problem, just parameters.
>>> But then my communication between App1 (tdma master) and App2 over rtlo
>>> is
>>> separated from the communication from app1 (tdma master) and the other
>>> slaves on rteth0.
>>> That would either mean that the other nodes couldn't communicate with
>>> App2
>>> or that App2 would "steal" app1 its timeslot. That's what I meant by
>>> "two
>>> rings" ;)
>> Sending over loopback doesn't have any impact on the slot usage or
>> scheduling over Ethernet. The sender just pushes its frame directly into
>> the local receiver's input queue (which may be the same queue/socket
>> that is also listening on Ethernet channels) and wakes it up. That's all.
>>
>> If you want to synchronise the sender on the TDMA cycle even when doing
>> local communication, you can additionally use the TDMA API to wait on
>> the cycle tick of the TDMA master.
> 
> Ah ok, that helps for the synchronisation problem. But my nodes connected
> via ethernet cannot communicate with my App2, or do I see something wrong
> ?
> A frame from a node with destination address of my master machine will
> reach only my App1, since I don't have something like ports to address the
> application on the machine.

If you are looking for multicast/broadcast delivery to more than one
local receiver socket - no, sorry, this doesn't work with RTnet yet. We
are currently lacking a scheme for RT-safe buffer replication, thus only
one socket can listen on the same IP port or the same Ethernet frame type.

Jan

Attachment: signature.asc
Description: OpenPGP digital signature

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
RTnet-users mailing list
RTnet-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rtnet-users

Reply via email to