The following avoids a mistake with r17-3692 by restoring the
way vectype_in was computed, pushing down the vec_in_num compute
before the input array is possibly mangled.
Bootstrapped and tested on aarch64-linux-gnu, pushed.
PR tree-optimization/127152
* tree-vect-loop.cc (vect_transform_reduction): Restore
vectype_in computation, compute vec_in_num from unaltered
vec_oprnds.
* gcc.dg/vect/vect-pr127152.c: New testcase.
---
gcc/testsuite/gcc.dg/vect/vect-pr127152.c | 29 +++++++++++++++++++++++
gcc/tree-vect-loop.cc | 7 +++---
2 files changed, 33 insertions(+), 3 deletions(-)
create mode 100644 gcc/testsuite/gcc.dg/vect/vect-pr127152.c
diff --git a/gcc/testsuite/gcc.dg/vect/vect-pr127152.c
b/gcc/testsuite/gcc.dg/vect/vect-pr127152.c
new file mode 100644
index 00000000000..4b9a599ff0d
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/vect/vect-pr127152.c
@@ -0,0 +1,29 @@
+/* { dg-additional-options "--param vect-partial-vector-usage=0" } */
+/* { dg-additional-options "-march=armv8-a+sve" { target aarch64-*-* } } */
+
+#include "tree-vect.h"
+
+__attribute__((noipa)) float
+f (float *a, int n)
+{
+ float s = 0.0f;
+ for (int i = 0; i < n; i++)
+ if (a[i] > 0.0f)
+ s += a[i];
+ return s;
+}
+
+int
+main (void)
+{
+ float a[16];
+
+ check_vect ();
+
+#pragma GCC novector
+ for (int i = 0; i < 16; i++)
+ a[i] = (i & 1) ? 1.0f : -1.0f;
+ if (f (a, 16) != 8.0f)
+ __builtin_abort ();
+ return 0;
+}
diff --git a/gcc/tree-vect-loop.cc b/gcc/tree-vect-loop.cc
index e2636c5b767..78f84c6269a 100644
--- a/gcc/tree-vect-loop.cc
+++ b/gcc/tree-vect-loop.cc
@@ -7992,9 +7992,9 @@ vect_transform_reduction (loop_vec_info loop_vinfo,
assumption is not true: we use reduc_index to record the index of the
reduction variable. */
int reduc_index = SLP_TREE_REDUC_IDX (slp_node);
- slp_tree node_in = SLP_TREE_CHILDREN (slp_node)[reduc_index == 0 ? 1 : 0];
- tree vectype_in = SLP_TREE_VECTYPE (node_in);
- unsigned vec_in_num = SLP_TREE_VEC_DEFS (node_in).length ();
+ tree vectype_in = SLP_TREE_VECTYPE (slp_node);
+ if (lane_reducing_op_p (op.code))
+ vectype_in = SLP_TREE_VECTYPE (SLP_TREE_CHILDREN (slp_node)[0]);
code_helper code = canonicalize_code (op.code, op.type);
internal_fn cond_fn
@@ -8085,6 +8085,7 @@ vect_transform_reduction (loop_vec_info loop_vinfo,
/* For single def-use cycles get one copy of the vectorized reduction
definition. */
+ unsigned vec_in_num = vec_oprnds[reduc_index == 0 ? 1 : 0].length ();
if (single_defuse_cycle)
{
vect_get_vec_defs (loop_vinfo, slp_node,
--
2.51.0
