On 01-03-16 10:45, Richard Biener wrote:
On Tue, 1 Mar 2016, Tom de Vries wrote:
Hi,
this patch fixes graphite PR68659, which is a 6 regression.
The patch fixes two things:
- by handling new_expr == NULL_TREE in collect_all_ssa_names, it handles a
COMPONENT_REF where operand 2 is NULL_TREE.
- it handles an ADDR_EXPR in get_new_name.
Together, they allow the currently failing test-case
gcc.dg/graphite/id-pr45230-1.c to pass.
Bootstrapped and reg-tested on x86_64.
OK for stage4 trunk?
@@ -1804,6 +1806,9 @@ get_new_name (basic_block new_bb, tree op,
if (is_constant (op))
return op;
+ if (TREE_CODE (op) == ADDR_EXPR)
+ return op;
+
return get_rename (new_bb, op, old_bb, phi_kind);
}
please instead change the is_constant (op) condition to
TREE_CODE (op) != SSA_NAME.
Done.
Bootstrapped and reg-tested on x86_64.
OK for stage4 trunk?
Thanks,
- Tom
Handle addr_expr and component_ref in graphite-ast-to-ast
2016-03-01 Tom de Vries <t...@codesourcery.com>
PR tree-optimization/68659
* graphite-isl-ast-to-gimple.c (collect_all_ssa_names): Handle
new_expr == NULL_TREE.
(get_new_name): Handle ADDR_EXPR.
---
gcc/graphite-isl-ast-to-gimple.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/gcc/graphite-isl-ast-to-gimple.c b/gcc/graphite-isl-ast-to-gimple.c
index d3614e4..89a4118 100644
--- a/gcc/graphite-isl-ast-to-gimple.c
+++ b/gcc/graphite-isl-ast-to-gimple.c
@@ -1403,6 +1403,8 @@ gsi_insert_earliest (gimple_seq seq)
void translate_isl_ast_to_gimple::
collect_all_ssa_names (tree new_expr, vec<tree> *vec_ssa)
{
+ if (new_expr == NULL_TREE)
+ return;
/* Rename all uses in new_expr. */
if (TREE_CODE (new_expr) == SSA_NAME)
@@ -1801,7 +1803,7 @@ get_new_name (basic_block new_bb, tree op,
basic_block old_bb, phi_node_kind phi_kind) const
{
/* For constants the names are the same. */
- if (is_constant (op))
+ if (TREE_CODE (op) != SSA_NAME)
return op;
return get_rename (new_bb, op, old_bb, phi_kind);
