Syncronization problem in rtmain.c and scicos test

Roberto Bucher roberto.bucher at supsi.ch
Wed Aug 9 11:08:36 CEST 2006


Thanks Gianluca

You are not the first who ahs this (strange) behaviour with the 
synchronisation between the RT task and the rtMain task. Perhaps it can be 
connected with the send/receive mechanism, which gives some troubles in the 
new version of RTAI with the provided msg.c from Paolo Mantegazza. We are 
still investigating why this module dosn't work correctly with rtmain in both 
Matlab and Scilab environments. I think that your solution can be introduced 
to solve this problem.

The "test.cos" file is not compatible with the RTAI-3.3 version, I'm sorry for 
this. I posted a corrected version on my homepage. 

It should be OK for the next RTAI-3.4 release. In this new version I canged a 
lot in the Scicos part, but I lived the old modules for compatibility. I now 
substituted all the RTAI/Scicos bock with function type "4" instead of the 
function type "2004" used in the other RTAI-3.x versions. Writing new scicos 
blocks should be now easy, because Scicos interface and corrsponding C 
function are now better separate.

I'll check your new rtmain.c files ASAP.

Thanks

Roberto


On Wednesday 09 August 2006 10:44, Gianluca Palli wrote:
> Hi Roberto
>
> i found a problem in the realtime applications generated both from
> matlab7/simulink/rtw and scilab-4.0/scicos with rtai-3.3
>
> During the first run of the application after the rtai modules loading, all
> works fine, but at any subsequent run the application seems to be blocked
> before the resuming of the rt_BaseRate task. I think the problem is in
> which task is running after the rt_send/rt_receive calls (may be a bug?).
>
> I solved the problem replacing  the syncronization between the rt_Main task
> and rt_BaseRate made by means of rt_send and rt_receive with a semaphore
> based syncronization.
>
> I send you the modified rtmain.c
>
> Another problem is in the test.cos example: to compile this example 
> correctly i replaced the rtai blocks with the corresponding ones from the
> RTAI-Lib palette. It seems the blocks in the example come from an older
> version of the RTAI-Lib.
>
> Best Regards
>
> Gianluca

-- 
University of Applied Sciences of Southern Switzerland
Dept. Innovative Technologies
CH-6928 Lugano-Manno
http://www.dti.supsi.ch/~bucher




More information about the Rtai mailing list