On Thu, Jan 4, 2018 at 11:06 AM, Richard Sandiford
<richard.sandif...@linaro.org> wrote:
> The first BIT_FIELD_REF folding pattern assumed without checking that
> operands satisfy tree_fits_uhwi_p. The second pattern does check this:
>
> /* On constants we can use native encode/interpret to constant
> fold (nearly) all BIT_FIELD_REFs. */
> if (CONSTANT_CLASS_P (arg0)
> && can_native_interpret_type_p (type)
> && BITS_PER_UNIT == 8
> && tree_fits_uhwi_p (op1)
> && tree_fits_uhwi_p (op2))
>
> so this patch adds the checks to the first pattern too. This is needed
> for POLY_INT_CST bit positions.
>
> Tested on aarch64-linux-gnu, x86_64-linux-gnu and powerpc64le-linux-gnu.
> Also tested by comparing the before and after assembly output for at
> least one target per CPU directory. OK to install?
Ok. It's of course because verify_expr verified that.
Richard.
> Richard
>
>
> 2018-01-04 Richard Sandiford <richard.sandif...@linaro.org>
>
> gcc/
> * fold-const.c (fold_ternary_loc): Check tree_fits_uhwi_p before
> using tree_to_uhwi.
>
> Index: gcc/fold-const.c
> ===================================================================
> --- gcc/fold-const.c 2018-01-03 21:42:34.349039784 +0000
> +++ gcc/fold-const.c 2018-01-04 10:06:02.165809202 +0000
> @@ -11643,7 +11643,9 @@ fold_ternary_loc (location_t loc, enum t
> if (TREE_CODE (arg0) == VECTOR_CST
> && (type == TREE_TYPE (TREE_TYPE (arg0))
> || (TREE_CODE (type) == VECTOR_TYPE
> - && TREE_TYPE (type) == TREE_TYPE (TREE_TYPE (arg0)))))
> + && TREE_TYPE (type) == TREE_TYPE (TREE_TYPE (arg0))))
> + && tree_fits_uhwi_p (op1)
> + && tree_fits_uhwi_p (op2))
> {
> tree eltype = TREE_TYPE (TREE_TYPE (arg0));
> unsigned HOST_WIDE_INT width = tree_to_uhwi (TYPE_SIZE (eltype));
