problems compiling latest snapshoot

[Joel Sherrill]

Chris Johns wrote:
> 
> Vladimir Nesic wrote:
> >
> > m68k-rtems-gcc --pipe -B../../../../lib/ -B../../../../gen68360/lib/
> > -specs bsp_specs -qrtems         -qrtems_debug -Wno-unused  -g -Wall
> > -mcpu32  -o o-debug/capture-cli.o -c
> > ../../../../../../rtems-ss-20021007/c/src/libmisc/capture/capture-cli.c
> >
> > ../../../../../../rtems-ss-20021007/c/src/libmisc/capture/capture-cli.c:
> > In function `rtems_capture_task_time':
> > ../../../../../../rtems-ss-20021007/c/src/libmisc/capture/capture-cli.c:1404:
> > Internal compiler error in instantiate_virtual_regs_1, at
> > function.c:3972
> 
> I get the error when doing the debug build on the cpu32 as you do. I did
> not see the error on the o-optimize or o-debug build for the Coldfire
> (-m5200).
> 
> After a little more testing the problem shows on:
> 
>   68040, 68030, 68020, cpu32
> 
> and passes on:
> 
>   68000, 68060, 5200
> 
> The -g option does not seem to make any difference so something else in
> the code seems to trip the error.
> 
> I am not sure what the solution to this problem is.

This sounds like a bug in gcc that needs to be reported to the gcc
maintainers.
Worse, it sounds like a regression from 2.95 which makes it important.

The standard procedure for filing gcc bugs is as follows:

  + Cut the failure case down to the smallest test case possible.
  + File a GCC GNATS PR and list me as an interested party and the CPU
    maintainers.

Please do this as we are coming down to a release point and it might
get fixed.

> --
>  Chris Johns, cjohns at cybertec.com.au

-- 
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