Hi!
On Mon, Jan 08, 2018 at 08:41:55PM +0100, Uros Bizjak wrote:
> Attached patch corrects wrong mode argument in the call to
> force_to_mode call for ASHIFT operator. The patch uses "mode" mode,
> the same as for all binop and unop operators in the force_int_to_mode
> function.
>
> Also, the unpatched function would force operand to op_mode and later
> truncate to op_mode again, so it all looks like a typo to me.
Indeed.
> 2018-01-08 Uros Bizjak <ubiz...@gmail.com>
>
> PR target/83628
> * combine.c (force_int_to_mode) <case ASHIFT>: Use mode instead of
> op_mode in the force_to_mode call.
>
> Together with a follow-up target patch, the patch fixes
> gcc.target/alpha/pr83628-2.c scan-asm failures on alpha.
>
> 2018-01-08 Uros Bizjak <ubiz...@gmail.com>
>
> PR target/83628
> * combine.c (force_int_to_mode) <case ASHIFT>: Use mode instead of
> op_mode in the force_to_mode call.
>
> Bootstrapped and regression tested on x86_64-linux-gnu {,-m32}.
>
> OK for mainline and branches?
OK for trunk; OK for the branches after a suitable burn-in period.
Thanks,
Segher
> diff --git a/gcc/combine.c b/gcc/combine.c
> index 3a42de53455c..6adc0a7d6f85 100644
> --- a/gcc/combine.c
> +++ b/gcc/combine.c
> @@ -8908,7 +8908,7 @@ force_int_to_mode (rtx x, scalar_int_mode mode,
> scalar_int_mode xmode,
> mask = fuller_mask;
>
> op0 = gen_lowpart_or_truncate (op_mode,
> - force_to_mode (XEXP (x, 0), op_mode,
> + force_to_mode (XEXP (x, 0), mode,
> mask, next_select));
>
> if (op_mode != xmode || op0 != XEXP (x, 0))
