Though Victor Yodaikens RT-Linux is great for developing
hard real-time applications using Linux, it does not
allow real-time tasks to use any of Linux's features (like
networking, etc...) To write a real-time application
that uses Linux's features, you need to split it into
two parts. A part that does not need such features (the
real-time part) and a part that needs to use these features
(the non-real-time part). These two parts can communicate
by using FIFOs (i think). Note that the non real-time part is not given any real-time guarantees. If the real-time application cannot be split into two parts, then you cannot use RT-Linux.
See http://hegel.ittc.ukans.edu/projects/kurt for further
details.
balaji
|