Some Haifa scheduler bugs

[Jeffrey A Law]

  In message < 199709041633.MAA20790@melange.gnu.ai.mit.edu >you write:
  > Good summary.
It's not 100% accurate, but it's close enough for this discussion :-)

  > The way I look at it is, all things being equal, I'd prefer an OS
  > default to faulting on dereferecing NULL when running generic C code
  > simply because C code shouldn't be dereferencing NULL.
Agreed.

HPUX does it backwards -- it defaults to allowing the dereference.
However the GCC port passes the magic flags to the linker to turn on
faulting by default :-)

It's a choice that can be made on a file by file basis -- and with
a little work one can even change the magic bit in the executable
header to change the behavior of a particular file.

I'm less familiar with how this works under aix or with ultrasparcs/solaris,
but I assume they both do something relatively reasonable.

  > not scheduling NULL dereferences.  Generally it is better to
  > default to not doing potentially hurtful optimizations, since users
  > who know they want optimizations know enough to specify them;
That's my general opinion too.

  > Note that the documentation should be *very* clear that, even if
  > this option is enabled, it does *not* guarantee that C (or Fortran
  > etc.) code can then be written that explicitly dereferences NULL.
Absolutely.

Jeff