Increasing MAX_SLOTS on RTAI 3.0

henrique at jpl.nasa.gov henrique at jpl.nasa.gov
Wed Jun 22 19:02:12 CEST 2005


I looked at nam2num() and it seems that it will generate 378 unique names 
(i.e. 6*(26*2+10+1) objects) and therefore the nam2num() could at most be 
used to register that many objects.  Therefore, I would need to increase 
the nam2num() to accept at least 159 char names to support my 10K registry.

         - David

At 05:32 AM 6/22/2005, you wrote:
>Peter Soetens wrote:
>>On Wednesday 22 June 2005 13:45, Paolo Mantegazza wrote:
>>
>>>Ciao,
>>>
>>>Malte experience is OK. The problems reported by Peter, in a following
>>>message, are likey due to registering names not univocally assigned by
>>>the user, but relying on something generated by RTAI when using a zero id.
>>
>>Interesting. This is what we do and I thought this was the safest way. 
>>How should an application then find new names when creating say, 5000+ 
>>semaphores ? Generating the names yourselves in a 'reserved' namespace ? 
>>I believe Takis would be interested in this topic too, since he's the one 
>>who actually stress-tested RTAI within this regard (or did I miss an 
>>earlier thread on this list ? )
>
>It is safe if you use just that, none can garantte there is no name 
>conflict if both users and internal namings are used.
>
>The reservation of different reserved named spaces within an application 
>is not a complex problem, xrtailab has one. The real problem is the 
>constraint to use just 6 characters that limis the ease of the scheme we 
>are using. So it will be solved, as I'm being bitten by it too, along the 
>line of my email.
>
>Paolo.
>






More information about the Rtai mailing list