[PATCH v2 3/6] target/riscv: Add support for Zvfbfmin extension

[Weiwei Li]

Add trans_* and helper function for Zvfbfmin 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>
---
 target/riscv/helper.h                      |  3 +
 target/riscv/insn32.decode                 |  4 ++
 target/riscv/insn_trans/trans_rvbf16.c.inc | 64 ++++++++++++++++++++++
 target/riscv/vector_helper.c               |  6 ++
 4 files changed, 77 insertions(+)

diff --git a/target/riscv/helper.h b/target/riscv/helper.h
index ef8487f1ee..fc48853e07 100644
--- a/target/riscv/helper.h
+++ b/target/riscv/helper.h
@@ -1157,3 +1157,6 @@ DEF_HELPER_FLAGS_2(cm_jalt, TCG_CALL_NO_WG, tl, env, i32)
 /* BF16 functions */
 DEF_HELPER_FLAGS_2(fcvt_bf16_s, TCG_CALL_NO_RWG, i64, env, i64)
 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)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 45fdcad185..10d001f14d 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -912,3 +912,7 @@ czero_nez   0000111  ..... ..... 111 ..... 0110011 @r
 # *** Zfbfmin Standard Extension ***
 fcvt_bf16_s       0100010  01000 ..... ... ..... 1010011 @r2_rm
 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
diff --git a/target/riscv/insn_trans/trans_rvbf16.c.inc 
b/target/riscv/insn_trans/trans_rvbf16.c.inc
index 8cafde505f..f794a3f745 100644
--- a/target/riscv/insn_trans/trans_rvbf16.c.inc
+++ b/target/riscv/insn_trans/trans_rvbf16.c.inc
@@ -22,6 +22,12 @@
     } \
 } while (0)
 
+#define REQUIRE_ZVFBFMIN(ctx) do { \
+    if (!ctx->cfg_ptr->ext_zvfbfmin) { \
+        return false; \
+    } \
+} while (0)
+
 static bool trans_fcvt_bf16_s(DisasContext *ctx, arg_fcvt_bf16_s *a)
 {
     REQUIRE_FPU;
@@ -51,3 +57,61 @@ static bool trans_fcvt_s_bf16(DisasContext *ctx, 
arg_fcvt_s_bf16 *a)
     mark_fs_dirty(ctx);
     return true;
 }
+
+static bool trans_vfncvtbf16_f_f_w(DisasContext *ctx, arg_vfncvtbf16_f_f_w *a)
+{
+    REQUIRE_FPU;
+    REQUIRE_ZVFBFMIN(ctx);
+
+    if (opfv_narrow_check(ctx, a) && (ctx->sew == MO_16)) {
+        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_3_ptr(vreg_ofs(ctx, a->rd), vreg_ofs(ctx, 0),
+                           vreg_ofs(ctx, a->rs2), cpu_env,
+                           ctx->cfg_ptr->vlen / 8,
+                           ctx->cfg_ptr->vlen / 8, data,
+                           gen_helper_vfncvtbf16_f_f_w);
+        mark_vs_dirty(ctx);
+        gen_set_label(over);
+        return true;
+    }
+    return false;
+}
+
+static bool trans_vfwcvtbf16_f_f_v(DisasContext *ctx, arg_vfwcvtbf16_f_f_v *a)
+{
+    REQUIRE_FPU;
+    REQUIRE_ZVFBFMIN(ctx);
+
+    if (opfv_widen_check(ctx, a) && (ctx->sew == MO_16)) {
+        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_3_ptr(vreg_ofs(ctx, a->rd), vreg_ofs(ctx, 0),
+                           vreg_ofs(ctx, a->rs2), cpu_env,
+                           ctx->cfg_ptr->vlen / 8,
+                           ctx->cfg_ptr->vlen / 8, data,
+                           gen_helper_vfwcvtbf16_f_f_v);
+        mark_vs_dirty(ctx);
+        gen_set_label(over);
+        return true;
+    }
+    return false;
+}
diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c
index 1e06e7447c..4d2bd42155 100644
--- a/target/riscv/vector_helper.c
+++ b/target/riscv/vector_helper.c
@@ -4535,6 +4535,9 @@ RVVCALL(OPFVV1, vfwcvt_f_f_v_w, WOP_UU_W, H8, H4, 
float32_to_float64)
 GEN_VEXT_V_ENV(vfwcvt_f_f_v_h, 4)
 GEN_VEXT_V_ENV(vfwcvt_f_f_v_w, 8)
 
+RVVCALL(OPFVV1, vfwcvtbf16_f_f_v, WOP_UU_H, H4, H2, bfloat16_to_float32)
+GEN_VEXT_V_ENV(vfwcvtbf16_f_f_v, 4)
+
 /* Narrowing Floating-Point/Integer Type-Convert Instructions */
 /* (TD, T2, TX2) */
 #define NOP_UU_B uint8_t,  uint16_t, uint32_t
@@ -4581,6 +4584,9 @@ RVVCALL(OPFVV1, vfncvt_f_f_w_w, NOP_UU_W, H4, H8, 
float64_to_float32)
 GEN_VEXT_V_ENV(vfncvt_f_f_w_h, 2)
 GEN_VEXT_V_ENV(vfncvt_f_f_w_w, 4)
 
+RVVCALL(OPFVV1, vfncvtbf16_f_f_w, NOP_UU_H, H2, H4, float32_to_bfloat16)
+GEN_VEXT_V_ENV(vfncvtbf16_f_f_w, 2)
+
 /*
  * Vector Reduction Operations
  */
-- 
2.25.1