accel/tcg: Handle MO_ATOM_WITHIN16 in do_st16_leN

Otherwise we hit the default assert not reached.
Handle it as MO_ATOM_NONE, because of size and misalignment.
We already handle this correctly in do_ld16_beN.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c
index 5e2ca472438b6bfeaa86348d1e97404bd00dc282..14ce97c33be904cff7caaee4fa6b57e968b5b8b9 100644 (file)
--- a/accel/tcg/cputlb.c
+++ b/accel/tcg/cputlb.c
@@ -2727,6 +2727,7 @@ static uint64_t do_st16_leN(CPUArchState *env, MMULookupPageData *p,
          * and so neither is atomic.
          */
     case MO_ATOM_IFALIGN:
+    case MO_ATOM_WITHIN16:
     case MO_ATOM_NONE:
         stq_le_p(p->haddr, int128_getlo(val_le));
         return store_bytes_leN(p->haddr + 8, p->size - 8,