As the following testcase shows, BIT_FIELD_REF result doesn't have to have
just integral type, it can also have vector type. And in that case
cxx_eval_bit_field_ref just ICEs on it because it is unprepared for that
case, creates the initial value with build_int_cst (sure, that one could be
easily replaced with build_zero_cst) and then expects it can through shifts,
ands and ors come up with the final value, but that doesn't work for
vectors.
We already call fold_ternary if whole is a VECTOR_CST, this patch does the
same if the result doesn't have integral type. And, there is no guarantee
fold_ternary will succeed and the callers certainly don't expect NULL
being returned, so it also diagnoses those as non-constant and returns
original t in that case.
2022-06-09 Jakub Jelinek <jakub@redhat.com>
PR c++/105871
* constexpr.cc (cxx_eval_bit_field_ref): For BIT_FIELD_REF with
non-integral result type use fold_ternary too like for BIT_FIELD_REFs
from VECTOR_CST. If fold_ternary returns NULL, diagnose non-constant
expression, set *non_constant_p and return t, instead of returning
NULL.
* g++.dg/pr105871.C: New test.
(cherry picked from commit
4c334e0e4ff05d3a7ca9ebb832428c446cd0ae8d)
if (*non_constant_p)
return t;
- if (TREE_CODE (whole) == VECTOR_CST)
- return fold_ternary (BIT_FIELD_REF, TREE_TYPE (t), whole,
- TREE_OPERAND (t, 1), TREE_OPERAND (t, 2));
+ if (TREE_CODE (whole) == VECTOR_CST || !INTEGRAL_TYPE_P (TREE_TYPE (t)))
+ {
+ if (tree r = fold_ternary (BIT_FIELD_REF, TREE_TYPE (t), whole,
+ TREE_OPERAND (t, 1), TREE_OPERAND (t, 2)))
+ return r;
+ if (!ctx->quiet)
+ error ("%qE is not a constant expression", orig_whole);
+ *non_constant_p = true;
+ return t;
+ }
start = TREE_OPERAND (t, 2);
istart = tree_to_shwi (start);
--- /dev/null
+// PR c++/105871
+// { dg-do compile }
+// { dg-options "-Wno-psabi" }
+
+typedef __attribute__((__vector_size__ ( 1))) unsigned char U;
+typedef __attribute__((__vector_size__ (16))) unsigned char V;
+
+U
+foo (void)
+{
+ return __builtin_shufflevector ((U){}, (V){}, 0);
+}