On sparc we end up choosing vector(8) <signed-boolean:1> for the
condition but vector(2) int for the value of a COND_EXPR but we
fail to verify their shapes match and thus things go downhill.
This is a missed-optimization on the pattern recognition side
as well as unhandled vector decomposition in vectorizable_condition.
The following plugs just the observed ICE for now.
Bootstrap and regtest ongoing on x86_64-unknown-linux-gnu.
PR tree-optimization/118529
* tree-vect-stmts.cc (vectorizable_condition): Check the
shape of the vector and condition vector type are compatible.
* gcc.target/sparc/pr118529.c: New testcase.
---
gcc/testsuite/gcc.target/sparc/pr118529.c | 17 +++++++++++++++++
gcc/tree-vect-stmts.cc | 5 +++--
2 files changed, 20 insertions(+), 2 deletions(-)
create mode 100644 gcc/testsuite/gcc.target/sparc/pr118529.c
diff --git a/gcc/testsuite/gcc.target/sparc/pr118529.c
b/gcc/testsuite/gcc.target/sparc/pr118529.c
new file mode 100644
index 00000000000..1393763e9db
--- /dev/null
+++ b/gcc/testsuite/gcc.target/sparc/pr118529.c
@@ -0,0 +1,17 @@
+/* { dg-do compile } */
+/* { dg-options "-O3 -mvis3" } */
+
+long c;
+int d[10];
+int e;
+void g() {
+ int b = 1 & e;
+ int *f = d;
+ b = -b;
+ c = 0;
+ for (; c < 10; c++) {
+ int h = f[c] ^ c;
+ h &= b;
+ f[c] ^= h;
+ }
+}
diff --git a/gcc/tree-vect-stmts.cc b/gcc/tree-vect-stmts.cc
index b5dd1a2e40f..833029fcb00 100644
--- a/gcc/tree-vect-stmts.cc
+++ b/gcc/tree-vect-stmts.cc
@@ -12676,8 +12676,9 @@ vectorizable_condition (vec_info *vinfo,
masked = !COMPARISON_CLASS_P (cond_expr);
vec_cmp_type = truth_type_for (comp_vectype);
-
- if (vec_cmp_type == NULL_TREE)
+ if (vec_cmp_type == NULL_TREE
+ || maybe_ne (TYPE_VECTOR_SUBPARTS (vectype),
+ TYPE_VECTOR_SUBPARTS (vec_cmp_type)))
return false;
cond_code = TREE_CODE (cond_expr);
--
2.43.0
