what is fusion?

Paolo Mantegazza mantegazza at aero.polimi.it
Tue Oct 5 11:30:29 CEST 2004

Gavin Li wrote:
> Hi,
> I browsed all the mail list archives, I couldn't find a description what is
> fusion? 
> another question: The ADEOS is not patented and suppose to replace the
> rthal, but it still needs the support of rthal which is the patented. why?
> at least when I use rtai3.1, I have to insmod rthal first.

Philippe has answered already but I do want to be picky on a point again.

Even with the old RTHAL way and RTAI-(NEW)LXRT, the only thing that 
should have counted for years, RTAI was not the dual kernel approach.

On the other hand the patent could have covered the RTAI legacy stuff 
related to kernel only applications, but for that there is the well 
known Moglen's clarification.

The fact that there are RTAI users that keep working in kernel only with 
sched_up/mup/smp and fifos, while everybody should have transitioned to 
LXRT (without_using_RTAI_PROPER_TASKS) for kernel only applications 
also, is just a masochistic constraint they want to put on themselves. 
They should thank Philippe for having cared of migrating such a support 
to 3.x also. On my side, after splitting the APIs from the schedulers 
proper, I never wrote a single line for kernel space only support in 3.x 
and never will again.

Finally I want to stress what I said many many times already. ADEOS has 
not been chosen for RTAI because of the patent, but because it grants a 
freedom community effort and because it is a technically superior 
support for _any_ Real Time Application Interface on a general purpose 
OS. When I say _any_ I mean not just RTAI, present and fusion, as with 
ADEOS anybody can easily do what (s)he wants.


More information about the Rtai mailing list