target/riscv: Add support for Zvfbfwma extension

Add trans_* and helper function for Zvfbfwma instructions.

Signed-off-by: Weiwei Li <liweiwei@iscas.ac.cn>
Signed-off-by: Junqiang Wang <wangjunqiang@iscas.ac.cn>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230615063302.102409-5-liweiwei@iscas.ac.cn>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>

target/riscv/helper.h

@@ -1160,3 +1160,6 @@ DEF_HELPER_FLAGS_2(fcvt_s_bf16, TCG_CALL_NO_RWG, i64, env, i64)
 
 DEF_HELPER_5(vfncvtbf16_f_f_w, void, ptr, ptr, ptr, env, i32)
 DEF_HELPER_5(vfwcvtbf16_f_f_v, void, ptr, ptr, ptr, env, i32)
+
+DEF_HELPER_6(vfwmaccbf16_vv, void, ptr, ptr, ptr, ptr, env, i32)
+DEF_HELPER_6(vfwmaccbf16_vf, void, ptr, ptr, i64, ptr, env, i32)

target/riscv/insn32.decode

@@ -916,3 +916,7 @@ fcvt_s_bf16       0100000  00110 ..... ... ..... 1010011 @r2_rm
 # *** Zvfbfmin Standard Extension ***
 vfncvtbf16_f_f_w  010010 . ..... 11101 001 ..... 1010111 @r2_vm
 vfwcvtbf16_f_f_v  010010 . ..... 01101 001 ..... 1010111 @r2_vm
+
+# *** Zvfbfwma Standard Extension ***
+vfwmaccbf16_vv    111011 . ..... ..... 001 ..... 1010111 @r_vm
+vfwmaccbf16_vf    111011 . ..... ..... 101 ..... 1010111 @r_vm

target/riscv/insn_trans/trans_rvbf16.c.inc

@@ -28,6 +28,12 @@
     } \
 } while (0)
 
+#define REQUIRE_ZVFBFWMA(ctx) do { \
+    if (!ctx->cfg_ptr->ext_zvfbfwma) { \
+        return false; \
+    } \
+} while (0)
+
 static bool trans_fcvt_bf16_s(DisasContext *ctx, arg_fcvt_bf16_s *a)
 {
     REQUIRE_FPU;
@@ -115,3 +121,55 @@ static bool trans_vfwcvtbf16_f_f_v(DisasContext *ctx, arg_vfwcvtbf16_f_f_v *a)
     }
     return false;
 }
+
+static bool trans_vfwmaccbf16_vv(DisasContext *ctx, arg_vfwmaccbf16_vv *a)
+{
+    REQUIRE_FPU;
+    REQUIRE_ZVFBFWMA(ctx);
+
+    if (require_rvv(ctx) && vext_check_isa_ill(ctx) && (ctx->sew == MO_16) &&
+        vext_check_dss(ctx, a->rd, a->rs1, a->rs2, a->vm)) {
+        uint32_t data = 0;
+        TCGLabel *over = gen_new_label();
+
+        gen_set_rm_chkfrm(ctx, RISCV_FRM_DYN);
+        tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_vl, 0, over);
+        tcg_gen_brcond_tl(TCG_COND_GEU, cpu_vstart, cpu_vl, over);
+
+        data = FIELD_DP32(data, VDATA, VM, a->vm);
+        data = FIELD_DP32(data, VDATA, LMUL, ctx->lmul);
+        data = FIELD_DP32(data, VDATA, VTA, ctx->vta);
+        data = FIELD_DP32(data, VDATA, VMA, ctx->vma);
+        tcg_gen_gvec_4_ptr(vreg_ofs(ctx, a->rd), vreg_ofs(ctx, 0),
+                           vreg_ofs(ctx, a->rs1),
+                           vreg_ofs(ctx, a->rs2), cpu_env,
+                           ctx->cfg_ptr->vlen / 8,
+                           ctx->cfg_ptr->vlen / 8, data,
+                           gen_helper_vfwmaccbf16_vv);
+        mark_vs_dirty(ctx);
+        gen_set_label(over);
+        return true;
+    }
+    return false;
+}
+
+static bool trans_vfwmaccbf16_vf(DisasContext *ctx, arg_vfwmaccbf16_vf *a)
+{
+    REQUIRE_FPU;
+    REQUIRE_ZVFBFWMA(ctx);
+
+    if (require_rvv(ctx) && (ctx->sew == MO_16) && vext_check_isa_ill(ctx) &&
+        vext_check_ds(ctx, a->rd, a->rs2, a->vm)) {
+        uint32_t data = 0;
+
+        gen_set_rm(ctx, RISCV_FRM_DYN);
+        data = FIELD_DP32(data, VDATA, VM, a->vm);
+        data = FIELD_DP32(data, VDATA, LMUL, ctx->lmul);
+        data = FIELD_DP32(data, VDATA, VTA, ctx->vta);
+        data = FIELD_DP32(data, VDATA, VMA, ctx->vma);
+        return opfvf_trans(a->rd, a->rs1, a->rs2, data,
+                           gen_helper_vfwmaccbf16_vf, ctx);
+    }
+
+    return false;
+}

target/riscv/vector_helper.c

@@ -3554,6 +3554,17 @@ RVVCALL(OPFVF3, vfwmacc_vf_w, WOP_UUU_W, H8, H4, fwmacc32)
 GEN_VEXT_VF(vfwmacc_vf_h, 4)
 GEN_VEXT_VF(vfwmacc_vf_w, 8)
 
+static uint32_t fwmaccbf16(uint16_t a, uint16_t b, uint32_t d, float_status *s)
+{
+    return float32_muladd(bfloat16_to_float32(a, s),
+                          bfloat16_to_float32(b, s), d, 0, s);
+}
+
+RVVCALL(OPFVV3, vfwmaccbf16_vv, WOP_UUU_H, H4, H2, H2, fwmaccbf16)
+GEN_VEXT_VV_ENV(vfwmaccbf16_vv, 4)
+RVVCALL(OPFVF3, vfwmaccbf16_vf, WOP_UUU_H, H4, H2, fwmacc16)
+GEN_VEXT_VF(vfwmaccbf16_vf, 4)
+
 static uint32_t fwnmacc16(uint16_t a, uint16_t b, uint32_t d, float_status *s)
 {
     return float32_muladd(float16_to_float32(a, true, s),