RTEMS build problems

[Ralf Corsepius]

On Mon, 2003-12-22 at 19:17, Paul Cadaret wrote:
> Ralph/Joel,
> Thank you for responding.  
> You are right.  My old 300MHz Ultra-10 is not fast... but it is very 
> reliable.
Ah, the same kind of machine I had sitting around in my office for quite
a while ;)

> You mentioned that Solaris 2.7 is likely to be broken.  
> I may be able to setup a build environment on a solaris-9-sparc machine.
> What are my chances getting the host-based version to build on a 
> solaris-9-sparc machine?
No, in contrary. 

Solaris2.7 unix/posix once definitely had worked, Solaris2.7 is also is
the Solaris release I still have access to (so actually could test) and
for which I have i386-linux->sparc-sun-solaris2.7 gcc cross toolchains.

The actual problem is the Unix/Posix  BSP having had far less attention
paid to it than the embedded cross-targets, so it is suffering from code
aging/rott. 

At present time, the unix/posix BSP RTEMS-CVS-trunk does not build for
me for any "unixy" target :-(

> BTW:  Just a note regarding my interest in RTEMS.  I am familiar with 
> PSOS as I used it for a few
> years many years ago.   I remember when Motorola put out the RTEID spec 
> many years ago and I remembered
> that RTEMS was based on that spec.  I always thought that PSOS was a 
> very reasonable RTOS model and
> I thought that I would see if  I could get RTEMS running on an old 
> MVME162 I have.  Since that time I
> have worked on alot of other rtos based products using other rtos 
> environments.
> 
>  As a 1st step I thought that I would just try to build the host-based 
> solaris version simply to provide a simple
> compile-link-test environment to familiarize myself with the 
> environment.

IMO, the easiest way to getting you started probably is to use an
embedded target on a simulator.

> If any solaris host based
> environment is likely to turn into a big science-project, let me know 
> and I'll stop and just move on to
> the m68k/mvme162 bsp environment.

Ralf