Re: [PATCH 1/2] target/s390x: Fix determination of overflow condition code after addition

[Thomas Huth]

On 30/03/2022 16.03, Richard Henderson wrote:
> On 3/30/22 02:52, David Hildenbrand wrote:
> > >   static uint32_t cc_calc_add_64(int64_t a1, int64_t a2, int64_t ar)
> > >   {
> > > -    if ((a1 > 0 && a2 > 0 && ar < 0) || (a1 < 0 && a2 < 0 && ar > 0)) {
> > > +    if ((a1 > 0 && a2 > 0 && ar < 0) || (a1 < 0 && a2 < 0 && ar >= 0)) {
> >
> >
> > Intuitively, I'd have checked for any overflow/underflow by comparing
> > with one of the input variables:
> >
> > a) Both numbers are positive
> >
> > Adding to positive numbers has to result in something that's bigger than
> > the input parameters.
> >
> > "a1 > 0 && a2 > 0 && ar < a1"
> >
> > b) Both numbers are negative
> >
> > Adding to negative numbers has to result in something that's smaller
> > than the input parameters.
> >
> > "a1 < 0 && a2 < 0 && ar > a1"
>
> If we're not going to just fix the >= typo,
> I'd suggest using the much more compact bitwise method:
>
>      ((ar ^ a1) & ~(a1 ^ a2)) < 0
>
> See sadd64_overflow in qemu/host-utils.h.

Thanks, sounds like a good idea. Anyway, I'd like to go with Bruno's patch 
for 7.0 and do the optimization in the 7.1 cycle instead.

 Thomas