Unaligned block moves and MEM_ALIGN re-broken

[David Edelsohn]

>>>>> Richard Kenner writes:

> I am a little concerned what other problems are going to show up with
> this patch which now is going to make a lot of objects appear much
> less aligned.

Richard> I'm not.  On a non-STRICT_ALIGNMENT machine, alignment problems affect only
Richard> efficiency, not correctness.

	We are trying to improve performance in GCC 3.1, so I do not want
to introduce any performance regressions due to this patch.

Richard> This patch needs to be thoroughly regression tested.

> I was hoping you'd do that on an rs6000 target.

	I have bootstrapped sucessfully on powerpc-ibm-aix4.3.3.0 both
with and without your patch.  There were no testsuite regressions for the
languages that are built on AIX (i.e., everything except Java).

Thanks, David