Hi all,
First of all, thanks for your insightful comments and support to our
project of building a simulator for RTL/RTAI.
My fellow developers and I finally come up with the following
reasoning:
As you may probably have noticed, developing a RTOS personality
on top of CarbonKernel is mainly a matter of implementing a set of
simulation objects that will stand for their "real world" counterparts
(e.g. threads/tasks, mutexes, semaphores and so on), then writing the proper
interface to interact with them. In other words, it's easy for us to dissociate
the behaviour of these objects from the mean we allow the applications to
control this behaviour.
Aside from the specific interfaces (RTL/RTLAI) and provided the
simulation development process won't have to stick to the actual
kernel implementations but rather follow its own way to do the same
things under the hood, what seems really important is the general
behaviour exhibited by the RTOS objects.
We have a reasonably common background between RT-Linux and RTAI at
the behaviour level at least. Albeit some slight differences exist,
common objects behave the same way in both worlds.
I propose we start implementing the common simulation core while
prototyping the RTL interface to test it, since it's simpler than RTAI/LXRT in
a first attempt. Right after this, we'll be able to complete the simulation
system implementing the LXRT interface, having a useful experience and a
working code base.
Comments are as usual welcome.
Cheers,
--Philippe
--
Philippe Gerum http://www.realiant.com/ [EMAIL PROTECTED]
19-21 rue du 8 Mai 1945 http://www.carbonkernel.org/ [EMAIL PROTECTED]
94110 Arcueil, France Tel: 33 1 41 24 31 10
-- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
---
For more information on Real-Time Linux see:
http://www.rtlinux.org/rtlinux/
