On Mon, Mar 30, 2020 at 06:24:32PM +0800, Kito Cheng wrote:
> Hi Jakub:
>
> I saw the omp and oacc related passes are in the head of all_passes,
> so I plan added after pass_omp_target_link, does it late enough?
Yes, that is ok.
> diff --git a/gcc/passes.def b/gcc/passes.def
> index
Hi Andrew:
> > + FOR_EACH_FUNCTION_WITH_GIMPLE_BODY (node)
> > +{
> > + function *fun = node->get_fun ();
> > + FOR_EACH_LOCAL_DECL (fun, i, var)
> > + {
> > + align = LOCAL_DECL_ALIGNMENT (var);
> > +
> > + SET_DECL_ALIGN (var, align);
>
>
> I think this is
Hi Jakub:
I saw the omp and oacc related passes are in the head of all_passes,
so I plan added after pass_omp_target_link, does it late enough?
diff --git a/gcc/passes.def b/gcc/passes.def
index 2bf2cb78fc5..92cbe587a8a 100644
--- a/gcc/passes.def
+++ b/gcc/passes.def
@@ -183,6 +183,7 @@ along
> But I wonder if we can instead fix the memcpy inlining issue by
> making the predicates involved honor LOCAL_ALIGNMENT
> instead of relying on DECL_ALIGN?
The memcpy inlining issue is not the only one affected by alignment
issue, I guess?
So I think it would be better to fix DECL_ALIGN?
On
On Mon, Mar 30, 2020 at 11:09:40AM +0200, Richard Biener wrote:
> On Fri, Mar 27, 2020 at 7:27 PM Jakub Jelinek wrote:
> >
> > On Sat, Mar 28, 2020 at 02:06:56AM +0800, Kito Cheng wrote:
> > > PR target/90811
> > > * ipa-increase-alignment.cc (increase_alignment_local_var): New.
> > >
On Fri, Mar 27, 2020 at 7:27 PM Jakub Jelinek wrote:
>
> On Sat, Mar 28, 2020 at 02:06:56AM +0800, Kito Cheng wrote:
> > PR target/90811
> > * ipa-increase-alignment.cc (increase_alignment_local_var): New.
> > (increase_alignment_global_var): New.
> >
On Sat, Mar 28, 2020 at 02:06:56AM +0800, Kito Cheng wrote:
> PR target/90811
> * ipa-increase-alignment.cc (increase_alignment_local_var): New.
> (increase_alignment_global_var): New.
> (pass_ipa_increase_alignment::gate): Remove.
I'm afraid this is too early,
On Fri, Mar 27, 2020 at 11:07 AM Kito Cheng wrote:
>
> - The alignment for local variable was adjust during
> estimate_stack_frame_size,
>however it seems wrong spot to adjust that, expand phase will adjust that
>but it little too late to some gimple optimization, which rely on certain
- The alignment for local variable was adjust during estimate_stack_frame_size,
however it seems wrong spot to adjust that, expand phase will adjust that
but it little too late to some gimple optimization, which rely on certain
target hooks need to check alignment, forwprop is an example
