Re: [commit] Fix testcase for SPU (Re: [RFA/testsuite] stack check)

["Ulrich Weigand" <uweigand@de.ibm.com>]

Jeroms Guitton wrote:

> Ulrich Weigand (uweigand@de.ibm.com):
> 
> > Jerome Guitton wrote:
> > 
> > > void big_frame ()
> > > {
> > >   char S [524188];
> > >   small_frame ();
> > > }
> > 
> > This breaks on the SPU where we only have 256 KB local store.
> 
> 
> Sorry about that. I guess that this may break other configurations;
> maybe we should enable this test on a limited set of targets?

Hmm, the GCC testsuite supports a gcc,stack_size target_info property
and modified the behaviour of some test cases depending on that info.
Maybe we want to use something like that in GDB as well ...

But for now, I'd be fine with either leaving this particular as is
and see if other platforms complain, or else restrict it only those
platforms it is supposed to be testing (the particular levels of
frame size where prologue code generation changes is platform-specific
anyway).

Bye,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com