[okl4-developer] Fast and Optimized System StartUp / OKL4 in an automotive environment

[Oliver Mayer-Buschmann]

Hi,

I still need help in using the iguana-api.
I've tried a lot to setup a loaded iguana application
but I failed in starting the loaded code.

That's what I've accomplished:
In my loader-server I have access to a shared memsection, containing an
elf-file, loaded from flash. This elf-file corresponds to a simple
iguana application (not yet a server).
The elf-file is linked to a predetermined virtual address and I'm able
to create a fixed memsection for the needed address range.
Then I use the nicta elf-loader to copy the code and data sections to
this memsection.
So far so good.

I've played a litte bit in creating threads, e.g. using
thread_create_simple(), etc.
In case I pass a local function of my loader-server to
thread_create_simple(), the new thread is created an the passed function
get's correctly executed. But this does not work, if I pass loaded
functions contained in the fixed memsection. 

I'm still wondering what a basic scenario for creating memsections a pd,
etc. for a dynamically created iguana application would be.

Shall I first create the fixed memsection,
then a pd and attach the memsection to the pd?

What else is needed? Caps, function calls like
memsection_register_server(), etc. ?

How can I create a thread in context of the target pd?

So basically what is the standard scenario and about what do I have take
care?

Thx,

Oliver

> > There is an Iguana reference manual on the OKL4 wiki.  Is there
> > something specific that you are after?
> 
> I'm after a sample Implementation how additional server's are
> registered
> to iguana. Some explanation for which API calls have to be performed
> in
> which order. That's what I mean when I talk about semantic.
> > 
> >       -gl