The previous changes to irange::constant_p return TRUE for
VARYING, since VARYING has numerical end points like any other
constant range. The problem is that some users of constant_p
depended on constant_p excluding the full domain. The
range handler for __builtin_clz, that is shared between ranger
and vr_values, is one such user.
This patch excludes varying_p(), to match the original behavior
for clz.
Tested on x86-64 Linux.
gcc/ChangeLog:
PR c/100521
* gimple-range.cc (range_of_builtin_call): Skip out on
processing __builtin_clz when varying.
---
gcc/gimple-range.cc | 2 +-
gcc/testsuite/gcc.dg/pr100521.c | 8 ++++++++
2 files changed, 9 insertions(+), 1 deletion(-)
create mode 100644 gcc/testsuite/gcc.dg/pr100521.c
diff --git a/gcc/gimple-range.cc b/gcc/gimple-range.cc
index e94bb355de3..5b288d8e6a7 100644
--- a/gcc/gimple-range.cc
+++ b/gcc/gimple-range.cc
@@ -737,7 +737,7 @@ range_of_builtin_call (range_query &query, irange &r, gcall
*call)
query.range_of_expr (r, arg, call);
// From clz of minimum we can compute result maximum.
- if (r.constant_p ())
+ if (r.constant_p () && !r.varying_p ())
{
int newmaxi = prec - 1 - wi::floor_log2 (r.lower_bound ());
// Argument is unsigned, so do nothing if it is [0, ...] range.
diff --git a/gcc/testsuite/gcc.dg/pr100521.c b/gcc/testsuite/gcc.dg/pr100521.c
new file mode 100644
index 00000000000..fd9f0db5412
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/pr100521.c
@@ -0,0 +1,8 @@
+/* { dg-do compile } */
+/* { dg-options "-O2" } */
+
+int
+__builtin_clz (int a)
+{
+ return __builtin_clz(a);
+}
--
2.31.1
