This is a partial roadmap for the project, composed of the currently
undergoing efforts, and other developments which are already planned
to start before Q2 2006. This roadmap will likely be complemented by
other tasks over time, this is just a somewhat sketchy vision of the
next major steps for which we do have an immediate plan and the
resources to achieve them.
Issues related to the usual debugging and extension of existing
features or skins are not covered here, Dmitry, Gilles, Jan and the
Uni-Hannover crew are usually taking care of these depending on the
code in question, with the help from other contributors; let's just
consider those issues as implicit, usual business for now. In any
case, the Xenomai-core list is open for discussing the matter and
filling the gaps in the roadmap.
o Web site.
- Bruno is working on this. His basic idea of the contents is
about being clear, simple and informational. Crafting a
useful and lively site is something of a daunting and tireless
task, so if you feel helping, just drop him a mail.
ETA: October 20 (initial version).
o Xenomai 2.0 release
ETA: October 22.
It's a bit early to define a timeframe for 2.1, we first need
to wait for the feedback we get with 2.0. Between both
releases, updates (2.0.1 and so on) will be made on a regular
basis.
o Automated benchmarking.
- We are still considering the best way to do that; actually,
my take is that we would just need to bootstrap the thing and
flesh it out over time, writing one or two significant
benchmark tests to start with, choosing a tool to plot the
collected data and push the results to some web page for
public consumption on a regular basis, but so far, we did not
manage to spark this. It's still in the short-term plan,
though, because we currently have neither metrics nor data to
check for basics, and we deeply need both of them now.
ETA: Q4 2005.
o Build system revamping.
- In order to allow binding the Xenomai core statically
to the Linux kernel while keeping the ability to have it as
loadable modules, we would need to refactor a number of
things in the existing build system. We are going to do
exactely that, which should make the use of Xenomai in
embedded setups more straightforward and efficient.
ETA: Q4 2005.
- Heikki has a plan to merge the ppc32 and ppc64 trees so that
we would track the same refactoring effort than the PPC kernel
folks are undertaking.
ETA: unspecified.
o Architecture ports.
- Analog Devices (http://www.analog.com/) have just offered
the project two Blackfin boards (bf533 and bf537) running
uClinux, so that we can first port Adeos over this
architecture, then the Xenomai core of course.
ETA: Adeos port, Q4 2005. Xenomai port, Q1 2006.
- An ARM port is finally underway (yes, really, for sure, no
kidding, I swear it!). For now, what we have is an almost
working Adeos/I-pipe patch, on an Integrator CP board running
an ARM1136 core. Stelian Pop will tell you more as the work
progresses.
ETA: Q1 2006.
o Kernel ports.
- We are going to backport Xenomai over 2.4, initially
targeting the PPC architecture. I do believe that Xenomai
will progress faster by confronting itself to low-end
hardware, which implies that we should also support the kernel
architecture which is running most of such hardware, and will
likely keep on doing so for a long time. For this task, we
will make good use of the boards the Denx's people will give
us access to. This task depends on the build system revamping
to be achieved.
ETA: Q1 2006.
o Scalability.
- Gilles is going to work on improving the scalability of the
timer management code, so that a large number of outstanding
timers would be more efficiently supported. This is
particularly important when it comes to port telecom-oriented
applications from traditional RTOS to Xenomai: those
applications could just create an insane number of concurrent
timers the way they are usually implemented.
ETA: Q1 2006.
--
Philippe.
