Re: [Qemu-devel] [PATCH 15/26] target/s390x: fix COMPARE LOGICAL LONG EXTENDED

[Richard Henderson]

On 05/25/2017 02:04 PM, Aurelien Jarno wrote:
> There are multiple issues with the COMPARE LOGICAL LONG EXTENDED
> instruction:
> - The test between the two operands is inverted, leading to an inversion
>    of the cc values 1 and 2.
> - The address and length of an operand continue to be decreased after
>    reaching the end of this operand. These values are then wrong write
>    back to the registers.
> - We should limit the amount of bytes to process, so that interrupts can
>    be served correctly.
>
> Signed-off-by: Aurelien Jarno <address@hidden>
> ---
>   target/s390x/mem_helper.c | 36 ++++++++++++++++++++++++++++--------
>   1 file changed, 28 insertions(+), 8 deletions(-)
>
> diff --git a/target/s390x/mem_helper.c b/target/s390x/mem_helper.c
> index 1dc71fe5f0..bd3bce3623 100644
> --- a/target/s390x/mem_helper.c
> +++ b/target/s390x/mem_helper.c
> @@ -716,28 +716,48 @@ uint32_t HELPER(clcle)(CPUS390XState *env, uint32_t r1, 
> uint64_t a2,
>       uint64_t srclen = get_length(env, r3 + 1);
>       uint64_t src = get_address(env, r3);
>       uint8_t pad = a2 & 0xff;
> +    uint64_t len = MAX(srclen, destlen);
>       uint32_t cc = 0;
>   
>       if (!(destlen || srclen)) {

Might as well replace this with len == 0 now.

>           if (v1 != v2) {
> -            cc = (v1 < v2) ? 1 : 2;
> +            cc = (v1 > v2) ? 1 : 2;

This would have been less confusing, perhaps, if we'd not used "src" and "dest" 
-- and doubly so since they are two sources -- but rather "src1" and "src3", 
which corresponds better to the PoO text.

That said,

Reviewed-by: Richard Henderson <address@hidden>


r~