On 23/10/15 16:27, Thomas Schwinge wrote:
Hi Jakub and Tom!
On Thu, 1 Oct 2015 15:37:26 +0200, Tom de Vries <tom_devr...@mentor.com> wrote:
On 01/10/15 14:49, Jakub Jelinek wrote:
On Thu, Oct 01, 2015 at 02:46:01PM +0200, Tom de Vries wrote:
this patch adds initialization in zero_iter_bb of counters introduced in
expand_omp_for_init_counts.
This removes the need to set TREE_NO_WARNING on those counters.
Why do you think it is a good idea?
[...]
Furthermore, I'd prefer not to change this area of code before
gomp-4_1-branch is merged, as it will be a nightmare for the merge
otherwise.
Committing to gomp-4_0-branch for now would work for me.
Well, the "nightmare" to merge this thus fell onto me... In particular,
as part of my gomp-4_0-branch r229255,
<http://news.gmane.org/find-root.php?message_id=%3C877fmd7lig.fsf%40schwinge.name%3E>,
I "butchered" your (Tom's) code (gomp-4_0-branch r228595) to work in
context of Jakub's changes -- would you please check that out (current
gomp-4_0-branch)? Even though there are no testsuite regressions, given
my lack of detailed understanding of this code, I'm not too sure about my
changes.
Here's, briefly, what I did: in gcc/omp-low.c:expand_omp_for_init_counts
extend created_zero_iter_bb handling for fd->ordered; at the end of that
function, individually decide (lame...) whether to use zero_iter1_bb or
zero_iter2_bb; in gcc/omp-low.c:expand_omp_for_generic remove
TREE_NO_WARNING code for the new zero_iter2_bb case, too.
Hi Thomas,
the merge looks good to me.
Thanks,
- Tom
