Problems building RTEMS-4.0.0 for MC68332-BSP

[Quality Quorum]

On Fri, 14 Jan 2000, Tobias Kolb wrote:

> Hi,
> 
> I´m new to RTEMS and would appreciate any help on my problem.
> 
> I´m trying to build RTEMS-4.0.0 for a BSP derived from the efi332-BSP
> which worked fine under RTEMS-3.6 and IMB-AIX (I didn´t do the changes
> to make this work and the collegue who did the adaptions is far ...) 
> 
> I´m trying to build RTEMS-4.0.0 for the MC68332 under Linux.
> I´m using the sources from "ftp.OARcorp.com" installing them as
> described in "Getting Started with RTEMS for C/C++ Users".
> I can build RTEMS using the "bit_rtems" script and the
> "CPU_CFLAGS=-m68020" in /rtems-4.0.0/make/custom/efi332.cfg as told for
> the efi332-BSP, convert the "hello.exe" sample application to a
> srecord-file and load it up on the target. Running the sample
> application on the target gives me the following message:
> 
> Spurious interrupt: line 1111 Emulation
> Vector: 0x0000000b  sp: 0x00f42bc4
> 
> Is it right that this means that there is floating-point-opcode which
> the MC68332
> doesn´t understand? 
> 
> I tried various compiler switches in /rtems-4.0.0/make/custom/BSP.cfg
> like:
> -m68020                                ; error message as cited
> -m68020 -msoft-float                   ; compiles, program on target
> doesn´t start
> -m68000 -msoft-float -Wa, -mcpu32      ; compiles, program on target
> doesn´t start
> ...
> 
> Which compiler switches do I have to use to create appropriate code for
> my target?
> Or is it a totally different problem? I´m really stuck at this point and
> would greatly appreciate any help.
> 
> Tank you in advance, greetings from Stuttgart

It may be fixed already, however, last time I was doing it I found that
it was required to manually set multilib options in gcc/config/m68k to
support CPU-32 properly (gcc won't bark if specified -mXXX is not
copiled into libs).

> 
> Tobi
> 

Thanks,

Aleksey