On Sun, May 10, 2026 at 1:37 PM Richard Henderson
<[email protected]> wrote:
>
> On 5/9/26 21:13, James Hilliard wrote:
> > SAA is an Octeon+ instruction that atomically adds rt to the naturally
> > aligned 32-bit word at base and discards the old memory value.
> >
> > Add an Octeon+ feature bit, enable it for Octeon68XX, and implement the
> > common SAA/SAAD translator with TCG atomic_fetch_add_i64.  The MemOp
> > selects the word or doubleword transaction size.
> >
> > Signed-off-by: James Hilliard <[email protected]>
> > ---
> > Changes v2 -> v3:
> >    - Split SAA out of the combined Octeon arithmetic and memory
> >      instruction patch.  (requested by Richard Henderson)
> >
> > Changes v3 -> v4:
> >    - Gate SAA/SAAD behind an Octeon+ feature bit.  (reported by Richard
> >      Henderson)
> >    - Use the i64 TCG atomic add path for both word and doubleword sizes.
> >      (suggested by Richard Henderson)
> > ---
> >   target/mips/cpu-defs.c.inc         |  2 +-
> >   target/mips/mips-defs.h            |  1 +
> >   target/mips/tcg/octeon.decode      |  4 ++++
> >   target/mips/tcg/octeon_translate.c | 19 +++++++++++++++++++
> >   4 files changed, 25 insertions(+), 1 deletion(-)
>
> I beg your pardon here.  Given that we implement only Octeon68XX, which is 
> Octeon3,
> there's not really a need for new feature bits.  For clarity, we could rename 
> INSN_OCTEON
> to INSN_OCTEON3, but that's not really necessary either.

You're right that a separate INSN_OCTEONP bit does not buy us much while
QEMU only has one Octeon CPU model, so I'll drop that in v5.

One small correction: Octeon68XX corresponds to CN68XX / Octeon II, not
Octeon III. My understanding is roughly:

  CN3XXX -> original Octeon
  CN58XX -> Octeon / Octeon+
  CN68XX -> Octeon II
  CN7XXX -> Octeon III

Some later patches model Octeon III-style state/forms because existing
binaries/toolchains use those encodings and QEMU has no separate Octeon
III CPU model today. But for SAA/SAAD, I agree the extra feature bit is
not worth carrying in this series.

>
> Anyway,
> Reviewed-by: Richard Henderson <[email protected]>
>
>
> r~
>
>

Reply via email to