Re: [PATCH 1/2] tcg/optimize: Fix fold_multiply2 vs 1

[Jim MacArthur <jim.macarthur@linaro.org>]

On Mon, Sep 14, 2026 at 03:41:14PM -1000, Richard Henderson wrote:
> For muls2, we returned from fold_multiply2 without finish_folding
> or any other mechanism to update value tracking for the rh output.
> We might as well do this via fold_shift, which will also mark rh
> as having all copies of the sign bit.
> 
> For the movi paths of multiply vs 0 or mulu2, we're better off tail
> calling to tcg_opt_gen_movi rather than explicitly returning true.
> 
> Cc: qemu-stable@nongnu.org
> Fixes: c9349965ce4 ("tcg: Optimize INDEX_op_mul[us]2 for 0 and 1")
> Resolves: https://gitlab.com/qemu-project/qemu/-/work_items/4448
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
>  tcg/optimize.c | 17 +++++++----------
>  1 file changed, 7 insertions(+), 10 deletions(-)
> 

Reviewed-by: Jim MacArthur <jim.macarthur@linaro.org>