+ if (a->rd && a->rs1 && a->rs2) {
+ TCGv_i32 shift = tcg_temp_new_i32();
+ tcg_gen_extrl_i64_i32(shift, cpu_gpr[a->rs2]);
+ tcg_gen_andi_i32(shift, shift, mask);
+ f64(vece, offsetof(CPURISCVState, gpr[a->rd]),
+ offsetof(CPURISCVState, gpr[a->rs1]),
+ shift, 8, 8);
+ tcg_temp_free_i32(shift);
+ return true;
+ }
+#endif
+ return rvp_shift_ool(ctx, a, fn, mask);
+}
+
+#define GEN_RVP_SHIFT(NAME, GVEC, VECE) \
+static bool trans_##NAME(DisasContext *s, arg_r *a) \
+{ \
+ return rvp_shift(s, a, VECE, GVEC, gen_helper_##NAME, \
+ (8 << VECE) - 1); \
+}
+
+GEN_RVP_SHIFT(sra16, tcg_gen_gvec_sars, 1);
+GEN_RVP_SHIFT(srl16, tcg_gen_gvec_shrs, 1);
+GEN_RVP_SHIFT(sll16, tcg_gen_gvec_shls, 1);
+GEN_RVP_R_OOL(sra16_u);
+GEN_RVP_R_OOL(srl16_u);
+GEN_RVP_R_OOL(ksll16);
+GEN_RVP_R_OOL(kslra16);
+GEN_RVP_R_OOL(kslra16_u);
+
+static bool rvp_shifti_ool(DisasContext *ctx, arg_shift *a,
+ gen_helper_rvp_r *fn)
+{
+ TCGv src1, dst, shift;
+
+ src1 = tcg_temp_new();
+ dst = tcg_temp_new();
+
+ gen_get_gpr(src1, a->rs1);
+ shift = tcg_const_tl(a->shamt);
+ fn(dst, cpu_env, src1, shift);
+ gen_set_gpr(a->rd, dst);
+
+ tcg_temp_free(src1);
+ tcg_temp_free(dst);
+ tcg_temp_free(shift);
+ return true;
+}
+
+static inline bool
+rvp_shifti(DisasContext *ctx, arg_shift *a,
+ void (* f64)(TCGv_i64, TCGv_i64, int64_t),
+ gen_helper_rvp_r *fn)
+{
+ if (!has_ext(ctx, RVP)) {
+ return false;
+ }
+
+#ifdef TARGET_RISCV64
+ if (a->rd && a->rs1 && f64) {
+ f64(cpu_gpr[a->rd], cpu_gpr[a->rs1], a->shamt);
+ return true;
+ }
+#endif
+ return rvp_shifti_ool(ctx, a, fn);
+}
+
+#define GEN_RVP_SHIFTI(NAME, OP, GVEC) \
+static bool trans_##NAME(DisasContext *s, arg_shift *a) \
+{ \
+ return rvp_shifti(s, a, GVEC, gen_helper_##OP); \
+}
+
+GEN_RVP_SHIFTI(srai16, sra16, tcg_gen_vec_sar16i_i64);
+GEN_RVP_SHIFTI(srli16, srl16, tcg_gen_vec_shr16i_i64);
+GEN_RVP_SHIFTI(slli16, sll16, tcg_gen_vec_shl16i_i64);
+GEN_RVP_SHIFTI(srai16_u, sra16_u, NULL);
+GEN_RVP_SHIFTI(srli16_u, srl16_u, NULL);
+GEN_RVP_SHIFTI(kslli16, ksll16, NULL);
diff --git a/target/riscv/packed_helper.c b/target/riscv/packed_helper.c
index 62db072204..7e31c2fe46 100644
--- a/target/riscv/packed_helper.c
+++ b/target/riscv/packed_helper.c
@@ -425,3 +425,107 @@ static inline void do_uksub8(CPURISCVState *env, void
*vd, void *va,
}
RVPR(uksub8, 1, 1);
+
+/* 16-bit Shift Instructions */
+static inline void do_sra16(CPURISCVState *env, void *vd, void *va,
+ void *vb, uint8_t i)
+{
+ int16_t *d = vd, *a = va;
+ uint8_t shift = *(uint8_t *)vb & 0xf;
+ d[i] = a[i] >> shift;
+}
+
+RVPR(sra16, 1, 2);
+
+static inline void do_srl16(CPURISCVState *env, void *vd, void *va,
+ void *vb, uint8_t i)
+{
+ uint16_t *d = vd, *a = va;
+ uint8_t shift = *(uint8_t *)vb & 0xf;
+ d[i] = a[i] >> shift;
+}
+
+RVPR(srl16, 1, 2);
+
+static inline void do_sll16(CPURISCVState *env, void *vd, void *va,
+ void *vb, uint8_t i)
+{
+ uint16_t *d = vd, *a = va;
+ uint8_t shift = *(uint8_t *)vb & 0xf;
+ d[i] = a[i] << shift;
+}
+
+RVPR(sll16, 1, 2);
+
+static inline void do_sra16_u(CPURISCVState *env, void *vd, void *va,
+ void *vb, uint8_t i)
+{
+ int16_t *d = vd, *a = va;
+ uint8_t shift = *(uint8_t *)vb & 0xf;
+
+ d[i] = vssra16(env, 0, a[i], shift);
+}
+
+RVPR(sra16_u, 1, 2);
+
+static inline void do_srl16_u(CPURISCVState *env, void *vd, void *va,
+ void *vb, uint8_t i)
+{
+ uint16_t *d = vd, *a = va;
+ uint8_t shift = *(uint8_t *)vb & 0xf;
+
+ d[i] = vssrl16(env, 0, a[i], shift);
+}
+
+RVPR(srl16_u, 1, 2);
+
+static inline void do_ksll16(CPURISCVState *env, void *vd, void *va,
+ void *vb, uint8_t i)
+{
+ int16_t *d = vd, *a = va, result;
+ uint8_t shift = *(uint8_t *)vb & 0xf;
+
+ result = a[i] << shift;
+ if (shift > (clrsb32(a[i]) - 16)) {
+ env->vxsat = 0x1;
+ d[i] = (a[i] & INT16_MIN) ? INT16_MIN : INT16_MAX;
+ } else {
+ d[i] = result;
+ }
+}
+
+RVPR(ksll16, 1, 2);
+
+static inline void do_kslra16(CPURISCVState *env, void *vd, void *va,
+ void *vb, uint8_t i)
+{
+ int16_t *d = vd, *a = va;
+ int32_t shift = sextract32((*(target_ulong *)vb), 0, 5);
+
+ if (shift >= 0) {
+ do_ksll16(env, vd, va, vb, i);
+ } else {
+ shift = -shift;
+ shift = (shift == 16) ? 15 : shift;
+ d[i] = a[i] >> shift;
+ }
+}
+
+RVPR(kslra16, 1, 2);
+
+static inline void do_kslra16_u(CPURISCVState *env, void *vd, void *va,
+ void *vb, uint8_t i)
+{
+ int16_t *d = vd, *a = va;
+ int32_t shift = sextract32((*(uint32_t *)vb), 0, 5);
+
+ if (shift >= 0) {
+ do_ksll16(env, vd, va, vb, i);
+ } else {
+ shift = -shift;
+ shift = (shift == 16) ? 15 : shift;
+ d[i] = vssra16(env, 0, a[i], shift);
+ }
+}
+
+RVPR(kslra16_u, 1, 2);
--
2.17.1