Re: [PATCH v8 40/62] target/riscv: vector floating-point compare instructions

[Richard Henderson]

On 5/21/20 2:43 AM, LIU Zhiwei wrote:
> +static uint8_t float16_eq_quiet(uint16_t a, uint16_t b, float_status *s)

Return bool, better than uint8_t.

> +{
> +    int compare = float16_compare_quiet(a, b, s);

New since your last revision is that compare should be type FloatRelation.

And similarly for the other 13 comparison helpers.

> +        vext_set_elem_mask(vd, mlen, i, DO_OP(s2, (ETYPE)s1,        \
> +                           &env->fp_status));                       \

Indentation is off.  env->fp_status should align with DO_OP.

With that,

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>

r~