Hi, Peter,
On Wed, Jul 28, 2021 at 9:21 PM Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
>
> On Wed, Jul 28, 2021 at 03:16:54PM +0200, Peter Zijlstra wrote:
> > On Wed, Jul 28, 2021 at 02:58:35PM +0200, Peter Zijlstra wrote:
> > > The below isn't quite right, because it'll use try_cmpxchg() for
> > > atomic_andnot_or(), which by being a void atomic should be _relaxed. I'm
> > > not entirely sure how to make that happen in a hurry.
> > >
> > > ---
> >
> > This seems to do the trick.
> >
>
> Mark suggested this, which is probably nicer still.
Wow, Amazing! so the architecture dependent can be implemented one by one.
Regards
Rui
>
> ---
> diff --git a/scripts/atomic/atomics.tbl b/scripts/atomic/atomics.tbl
> index fbee2f6190d9..3aaa0caa6b2d 100755
> --- a/scripts/atomic/atomics.tbl
> +++ b/scripts/atomic/atomics.tbl
> @@ -39,3 +39,4 @@ inc_not_zero b v
> inc_unless_negative b v
> dec_unless_positive b v
> dec_if_positive i v
> +andnot_or vF v i:m i:o
> diff --git a/scripts/atomic/fallbacks/andnot_or b/scripts/atomic/fallbacks/andnot_or
> new file mode 100644
> index 000000000000..0fb3a728c0ff
> --- /dev/null
> +++ b/scripts/atomic/fallbacks/andnot_or
> @@ -0,0 +1,24 @@
> +local try_order=${order}
> +
> +#
> +# non-value returning atomics are implicity relaxed
> +#
> +if [ -z "${retstmt}" ]; then
> + try_order="_relaxed"
> +fi
> +
> +cat <<EOF
> +static __always_inline ${ret}
> +arch_${atomic}_${pfx}andnot_or${sfx}${order}(${atomic}_t *v, ${int} m, ${int} o)
> +{
> + ${retstmt}({
> + ${int} N, O = atomic_read(v);
> + do {
> + N = O;
> + N &= ~m;
> + N |= o;
> + } while (!arch_${atomic}_try_cmpxchg${try_order}(v, &O, N));
> + O;
> + });
> +}
> +EOF
> diff --git a/scripts/atomic/fallbacks/fetch_andnot_or b/scripts/atomic/fallbacks/fetch_andnot_or
> deleted file mode 100644
> index e69de29bb2d1..000000000000
