On Wed, Nov 2, 2022 at 10:48 PM apinski--- via Gcc-patches
<gcc-patches@gcc.gnu.org> wrote:
>
> From: Andrew Pinski <apin...@marvell.com>
>
> Right now anyone could call tree_nonzero_bits with
> either complex or vector types and this will return
> the wrong thing. So just assert that nobody calls
> it with this.
>
> OK? Bootstrapped and tested with no regressions on x86_64-linux-gnu.
The function comment needs adjustment. To me, nonzero bits of
a v4si would simply include all bits of v4si - isn't that what happens?
I see the SSA get_nonzero_bits uses element_precision but
tree_nonzero_bits uses TYPE_PRECISON as fallback, so I guess
the assert is correct.
Still the function comment should be amended, also ...
> gcc/ChangeLog:
>
> * fold-const.cc (tree_nonzero_bits): Add
> assert.
> ---
> gcc/fold-const.cc | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/gcc/fold-const.cc b/gcc/fold-const.cc
> index 7e1ea58518b..3ccac9b28df 100644
> --- a/gcc/fold-const.cc
> +++ b/gcc/fold-const.cc
> @@ -16567,6 +16567,9 @@ c_getstr (tree str)
> wide_int
> tree_nonzero_bits (const_tree t)
> {
> + gcc_assert (TREE_CODE (TREE_TYPE (t)) != VECTOR_TYPE
> + && TREE_CODE (TREE_TYPE (t)) != COMPLEX_TYPE);
... please perform a "positive" test. Like INTEGRAL_TYPE_P (TREE_TYPE
(t)) || POINTER_TYPE_P (TREE_TYPE (t))
Richard.
> +
> switch (TREE_CODE (t))
> {
> case INTEGER_CST:
> --
> 2.17.1
>
