[committed,PR,tree-optimization/109392] Handle failure from maybe_push_res_to_seq
Commit Message
I was scanning through the P3s today and saw this patch from Jakub to
fix 109392. It looked quite reasonable, so I went ahead and did the
bootstrap & regression test. Everything looked good and I've applied it
to the trunk.
Jeff
commit 101380a8394c22a7a2ea70de2060ee93716156e2
Author: Jakub Jelinek <jakub@redhat.com>
Date: Sat Apr 8 12:21:42 2023 -0600
PR tree-optimization/109392
If we have an object with SSA_NAME_OCCURS_IN_ABNORMAL_PHI, then
maybe_push_res_to_seq may fail. Directly build the extraction
for that case.
PR tree-optimization/109392
gcc/
* tree-vect-generic.cc (tree_vec_extract): Handle failure
of maybe_push_res_to_seq better.
gcc/testsuite/
* gcc.dg/pr109392.c: New test.
new file mode 100644
@@ -0,0 +1,15 @@
+/* PR tree-optimization/109392 */
+/* { dg-do compile } */
+/* { dg-options "-O2 -Wno-psabi" } */
+
+typedef short __attribute__ ((__vector_size__ (64))) V;
+V v, w;
+void bar (void) __attribute__((returns_twice));
+
+V
+foo (V a, V b)
+{
+ bar ();
+ b &= v < b;
+ return (V) { foo (b, w)[3], (V) {}[3] };
+}
@@ -174,7 +174,16 @@ tree_vec_extract (gimple_stmt_iterator *gsi, tree type,
opr.resimplify (NULL, follow_all_ssa_edges);
gimple_seq stmts = NULL;
tree res = maybe_push_res_to_seq (&opr, &stmts);
- gcc_assert (res);
+ if (!res)
+ {
+ /* This can happen if SSA_NAME_OCCURS_IN_ABNORMAL_PHI are
+ used. Build BIT_FIELD_REF manually otherwise. */
+ t = build3 (BIT_FIELD_REF, type, t, bitsize, bitpos);
+ res = make_ssa_name (type);
+ gimple *g = gimple_build_assign (res, t);
+ gsi_insert_before (gsi, g, GSI_SAME_STMT);
+ return res;
+ }
gsi_insert_seq_before (gsi, stmts, GSI_SAME_STMT);
return res;
}