diff options
Diffstat (limited to 'target/arm/tcg/vec_helper.c')
| -rw-r--r-- | target/arm/tcg/vec_helper.c | 15 |
1 files changed, 12 insertions, 3 deletions
diff --git a/target/arm/tcg/vec_helper.c b/target/arm/tcg/vec_helper.c index 42bb43ac..aefcd07 100644 --- a/target/arm/tcg/vec_helper.c +++ b/target/arm/tcg/vec_helper.c @@ -2178,19 +2178,28 @@ void HELPER(sve2_fmlal_zzzw_s)(void *vd, void *vn, void *vm, void *va, CPUARMState *env, uint32_t desc) { intptr_t i, oprsz = simd_oprsz(desc); - uint16_t negn = extract32(desc, SIMD_DATA_SHIFT, 1) << 15; + bool is_s = extract32(desc, SIMD_DATA_SHIFT, 1); intptr_t sel = extract32(desc, SIMD_DATA_SHIFT + 1, 1) * sizeof(float16); float_status *status = &env->vfp.fp_status_a64; bool fz16 = get_flush_inputs_to_zero(&env->vfp.fp_status_f16_a64); + int negx = 0, negf = 0; + + if (is_s) { + if (env->vfp.fpcr & FPCR_AH) { + negf = float_muladd_negate_product; + } else { + negx = 0x8000; + } + } for (i = 0; i < oprsz; i += sizeof(float32)) { - float16 nn_16 = *(float16 *)(vn + H1_2(i + sel)) ^ negn; + float16 nn_16 = *(float16 *)(vn + H1_2(i + sel)) ^ negx; float16 mm_16 = *(float16 *)(vm + H1_2(i + sel)); float32 nn = float16_to_float32_by_bits(nn_16, fz16); float32 mm = float16_to_float32_by_bits(mm_16, fz16); float32 aa = *(float32 *)(va + H1_4(i)); - *(float32 *)(vd + H1_4(i)) = float32_muladd(nn, mm, aa, 0, status); + *(float32 *)(vd + H1_4(i)) = float32_muladd(nn, mm, aa, negf, status); } } |