On Fri, Aug 1, 2025 at 10:38 AM <pan2...@intel.com> wrote:
>
> From: Pan Li <pan2...@intel.com>
>
> This patch would like to try to match the the unsigned
> SAT_MUL form 2, aka below:
>
> #define DEF_SAT_U_MUL_FMT_2(T) \
> T __attribute__((noinline)) \
> sat_u_mul_##T##_fmt_2 (T a, T b) \
> { \
> T result; \
> if (__builtin_mul_overflow(a, b, &result)) \
> return -1; \
> else \
> return result; \
> }
>
> While T is uint8_t, uint16_t, uint32_t and uint64_t.
OK.
Thanks,
Richard.
> gcc/ChangeLog:
>
> * match.pd: Add form 2 for unsigned SAT_MUL.
>
> Signed-off-by: Pan Li <pan2...@intel.com>
> ---
> gcc/match.pd | 12 ++++++++++++
> 1 file changed, 12 insertions(+)
>
> diff --git a/gcc/match.pd b/gcc/match.pd
> index 82e6e291ae1..a94730680ad 100644
> --- a/gcc/match.pd
> +++ b/gcc/match.pd
> @@ -3623,6 +3623,18 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
> bool c2_is_max_p = wi::eq_p (c2, max);
> }
> (if (widen_prec > prec && c2_is_max_p)))))
> + (match (unsigned_integer_sat_mul @0 @1)
> + /* SAT_U_MUL (X, Y) = {
> + T result;
> + if (__builtin_mul_overflow (X, Y, &result))
> + return -1;
> + else
> + return -(T)overflow_p | result;
> + }
> + while T can be uint8_t, uint16_t, uint32_t and uint64_t. */
> + (cond^ (ne (imagpart (IFN_MUL_OVERFLOW:c@2 @0 @1)) integer_zerop)
> + integer_minus_onep (realpart @2))
> + (if (types_match (type, @0, @1))))
> )
>
> /* The boundary condition for case 10: IMM = 1:
> --
> 2.43.0
>
