Can we run RTEMS under AT91SAM7S256?

[Joel Sherrill <joel@OARcorp.com>]

Jay Monkman wrote:
> On Thu, Oct 21, 2004 at 05:28:49AM +0200, Ralf Corsepius wrote:
> 
>>>I'm sure those could be reduced some with effort.
>>
>>Definitely. 
>>
>>Did you try to compile RTEMS w/ -Os?
> 
> 
> Doesn't make a difference. I'm using -O3 already. In the ARM stuff, I
> know the processor stacks could be reduced, and there's a 4K buffer
> used for an incompletely implemented data abort handler and it should
> be removed altogether.

As Ralf and I also spoke about in Munich, there is a lot in RTEMS
to be "complete" which could conditionally be turned off or
reduced:

   + 256 priorities -> a smaller power of two >= 4
   + disble POSIX signals
   + Classic notepads
   + disable Classic signals
   + build without POSIX enabled (already supported)
   + reuse stack from init
   + disable priority ceiling and inheritance
   + 16 bit object Ids
   + a bit more MP code can be disable when single CPU
   + when not using a manager drop init out
   + disable all exit support
   + disable C library reentrancy support

It is ALWAYS possible that a BSP is linking in things that
are not required.  Some may call printf which we know is
20K of code most of the time.  Reducing depencies
in the BSP is always a good thing to do.

THe goal of a minimum RTEMS application right now is not
to be the smallest, least functional code base.  It is to
be the baseline for a real full-featured application.



-- 
Joel Sherrill, Ph.D.             Director of Research & Development
joel at OARcorp.com                 On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
    Support Available             (256) 722-9985