On 07/26/2016 09:53 AM, Dominik Vogt wrote:
Finally a patch that works and is simple. Bootstrapped and
regression tested on s390, s390x biarch and x86_64. The new patch
exploits the known alignment of (stack pointer +
STACK_DYNAMIC_OFFSET) as described earlier (see below). I think
that is the right way to get rid of the extra allocation. It
took a long time to understand the problem.
As the patch triggers a bug in the fortran compiler, the
der_type.f90 test case may fail on some targets if this patch is
used without the fortran fix that I've posted in another thread.
(The patch also contains a fix for a typo in a comment in the
patched function.)
See ChangeLog for a full description of the new patch.
Since the patch is all new, we're not going to commit it without a
new OK.
I like this one much better :-)
OK.
Thanks for your patience,
JEff
