[Bdi4emc-help] Writing a stepper driver and have a couple questions

[Paul]

Hi Matt

On Wednesday 08 February 2006 14:12, Matt Timmermans wrote:
> I'm writing a new stepper driver on BDI 4.38, so I can run my x10
> microstepping Gecko drives on the P266 laptop that I have allocated to the
> purpose.

With a P266, you might want to consider running xfce4, evilwm, or one of the 
other lightweight window managers - I fear KDE will suck up too much in the 
way of resources.. The downside is no pretty desktop icons to launch EMC 
with.

> The effort has taught me what a pain it is to get a patched kernel and EMC
> running from available source and information.  Now that I've figured out
> that I don't have to patch the kernel or adeos, I can use BDI, and in
> thanks for all the hard work that must have gone into it, I'm willing to
> release the driver under GPL when it's done.

BDI-4 is an ongoing project, and with a little help, we aim to release 
BDI-4.40 early in the summer. Things have settled down with Debian declared 
Sarge as stable, and the RTAI team are moving ever closer to a final release 
of rtai-3.3. EMC it's self is undergoing quite a bit of work in an attempt to 
resolve a number of outstanding issues, although it is possible some of the 
more radical changes won't be ready in time..

> Are you guys still accepting code submissions for the BDI branch of EMC?
>  If so, what's the best way to go about that?

Currently, patches (preferably with -uw flags), complete files, or source 
tarballs are acceptable - Contact me off list and we can get you up to speed 
with some of the changes in the pipeline that might affect the implementation 
of a new driver.

> Also a little technical question: freqmod has an empty function called
> freqsig().  What's that for?

The RTAI scheduler will call freqsig() when freqfunc() becomes the active 
task. Within the context of freqmod (and EMC generally), this feature is not 
used, hence a dummy function. In more complex multithreaded RT applications, 
this could be used to service interrupts, initialise variables, or do 
nothing.


Regards, Paul.


-- 
From the Klingon book of C:
Klingon function calls do not have 'parameters' - they have 'arguments' - and
they ALWAYS WIN THEM.