BSP choices

[Joel Sherrill]

lange92 at 2067.resnet.uni.edu wrote:
> 
> On Sat, 28 Jul 2001, Thomas Bocek wrote:
> 
> > On Friday 27 July 2001 17:25, you wrote:
> > > Hello rtems-users,
> > >   I have been working with the GNAT/RTEMS 4.5.0 release packages...
> 
> > > When I try to compile either of the first 2 with --enable-posix and
> > > defaults for the rest, it aborts with a nice error during the `make all`
> > > step: ../../../../../i386ex/lib/linkcmds:115 parse error.
> 
> > I had the same problem, but I dont know how to fix it.
> 
> After a quick analysis, I saw linkcmds:115 to contain
>    _rodata_start 5 . ;
> I changed it to:
>    _rodata_start = .;
> and this BSP compiled as far as the other ones. I considered the
> possibility of
>    _rodata_start % .;
> but haven't tested it, and haven't decided whether it would be parsable
> (instructing a linker is hardly my forte.)

Exactly on the "=".  THis slipped through from some corruption caused
by our CVS server having a HW problem.

> Now my problem is that all the BSPs fail out with the following error:
> /opt/gnatrtems/i386-rtems/bin/ld: cannot open crti.o: No such file or
> directory
> 
> Best I can figure is that something failed when I built the cross
> compiler. Any ideas?

Seem my longer post but the root cause is that gcc 2.8.1 and hence
can't treat i386-rtems as COFF while newer RTEMS and gcc 2.9+ use
ELF for i386-rtems.  I think you can delete the references from
bsp_specs and move forward.  The linkcmds script is elf-ish so
might need a hack or two.

> DanL

-- 
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