RISC-V: Support RVV VFMUL rounding mode intrinsic API

Update in v2:

* Sync with upstream for the vfmul duplicated declaration.

Original log:

This patch would like to support the rounding mode API for the VFMUL
for the below samples.

* __riscv_vfmul_vv_f32m1_rm
* __riscv_vfmul_vv_f32m1_rm_m
* __riscv_vfmul_vf_f32m1_rm
* __riscv_vfmul_vf_f32m1_rm_m

Signed-off-by: Pan Li <pan2.li@intel.com>

gcc/ChangeLog:

	* config/riscv/riscv-vector-builtins-bases.cc
	(vfmul_frm_obj): New declaration.
	(Base): Likewise.
	* config/riscv/riscv-vector-builtins-bases.h: Likewise.
	* config/riscv/riscv-vector-builtins-functions.def
	(vfmul_frm): New function definition.
	* config/riscv/vector.md: Add vfmul to frm_mode.

gcc/testsuite/ChangeLog:

	* gcc.target/riscv/rvv/base/float-point-single-mul.c: New test.

gcc/config/riscv/riscv-vector-builtins-bases.cc

@@ -277,6 +277,7 @@ public:
 
 /* Implements below instructions for now.
    - vfadd
+   - vfmul
 */
 template<rtx_code CODE>
 class binop_frm : public function_base
@@ -2103,6 +2104,7 @@ static CONSTEXPR const widen_binop_frm<PLUS> vfwadd_frm_obj;
 static CONSTEXPR const widen_binop<MINUS> vfwsub_obj;
 static CONSTEXPR const widen_binop_frm<MINUS> vfwsub_frm_obj;
 static CONSTEXPR const binop<MULT> vfmul_obj;
+static CONSTEXPR const binop_frm<MULT> vfmul_frm_obj;
 static CONSTEXPR const binop<DIV> vfdiv_obj;
 static CONSTEXPR const reverse_binop<DIV> vfrdiv_obj;
 static CONSTEXPR const widen_binop<MULT> vfwmul_obj;
@@ -2334,6 +2336,7 @@ BASE (vfwadd_frm)
 BASE (vfwsub)
 BASE (vfwsub_frm)
 BASE (vfmul)
+BASE (vfmul_frm)
 BASE (vfdiv)
 BASE (vfrdiv)
 BASE (vfwmul)

gcc/config/riscv/riscv-vector-builtins-bases.h

@@ -152,6 +152,7 @@ extern const function_base *const vfwadd_frm;
 extern const function_base *const vfwsub;
 extern const function_base *const vfwsub_frm;
 extern const function_base *const vfmul;
+extern const function_base *const vfmul_frm;
 extern const function_base *const vfdiv;
 extern const function_base *const vfrdiv;
 extern const function_base *const vfwmul;

gcc/config/riscv/riscv-vector-builtins-functions.def

@@ -319,6 +319,8 @@ DEF_RVV_FUNCTION (vfmul, alu, full_preds, f_vvf_ops)
 DEF_RVV_FUNCTION (vfdiv, alu, full_preds, f_vvv_ops)
 DEF_RVV_FUNCTION (vfdiv, alu, full_preds, f_vvf_ops)
 DEF_RVV_FUNCTION (vfrdiv, alu, full_preds, f_vvf_ops)
+DEF_RVV_FUNCTION (vfmul_frm, alu_frm, full_preds, f_vvv_ops)
+DEF_RVV_FUNCTION (vfmul_frm, alu_frm, full_preds, f_vvf_ops)
 
 // 13.5. Vector Widening Floating-Point Multiply
 DEF_RVV_FUNCTION (vfwmul, alu, full_preds, f_wvv_ops)

gcc/config/riscv/vector.md

@@ -866,7 +866,7 @@
 
 ;; Defines rounding mode of an floating-point operation.
 (define_attr "frm_mode" "rne,rtz,rdn,rup,rmm,dyn,dyn_exit,dyn_call,none"
-  (cond [(eq_attr "type" "vfalu,vfwalu")
+  (cond [(eq_attr "type" "vfalu,vfwalu,vfmul")
          (cond
 	   [(match_test "INTVAL (operands[9]) == riscv_vector::FRM_RNE")
 	    (const_string "rne")

gcc/testsuite/gcc.target/riscv/rvv/base/float-point-single-mul.c

@@ -0,0 +1,44 @@
+/* { dg-do compile } */
+/* { dg-options "-march=rv64gcv -mabi=lp64 -O3 -Wno-psabi" } */
+
+#include "riscv_vector.h"
+
+typedef float float32_t;
+
+vfloat32m1_t
+test_riscv_vfmul_vv_f32m1_rm (vfloat32m1_t op1, vfloat32m1_t op2, size_t vl) {
+  return __riscv_vfmul_vv_f32m1_rm (op1, op2, 0, vl);
+}
+
+vfloat32m1_t
+test_vfmul_vv_f32m1_rm_m (vbool32_t mask, vfloat32m1_t op1, vfloat32m1_t op2,
+			  size_t vl) {
+  return __riscv_vfmul_vv_f32m1_rm_m (mask, op1, op2, 1, vl);
+}
+
+vfloat32m1_t
+test_vfmul_vf_f32m1_rm (vfloat32m1_t op1, float32_t op2, size_t vl) {
+  return __riscv_vfmul_vf_f32m1_rm (op1, op2, 2, vl);
+}
+
+vfloat32m1_t
+test_vfmul_vf_f32m1_rm_m (vbool32_t mask, vfloat32m1_t op1, float32_t op2,
+			  size_t vl) {
+  return __riscv_vfmul_vf_f32m1_rm_m (mask, op1, op2, 3, vl);
+}
+
+vfloat32m1_t
+test_riscv_vfmul_vv_f32m1 (vfloat32m1_t op1, vfloat32m1_t op2, size_t vl) {
+  return __riscv_vfmul_vv_f32m1 (op1, op2, vl);
+}
+
+vfloat32m1_t
+test_vfmul_vv_f32m1_m (vbool32_t mask, vfloat32m1_t op1, vfloat32m1_t op2,
+		       size_t vl) {
+  return __riscv_vfmul_vv_f32m1_m (mask, op1, op2, vl);
+}
+
+/* { dg-final { scan-assembler-times {vfmul\.v[vf]\s+v[0-9]+,\s*v[0-9]+,\s*[fav]+[0-9]+} 6 } } */
+/* { dg-final { scan-assembler-times {frrm\s+[axs][0-9]+} 4 } } */
+/* { dg-final { scan-assembler-times {fsrm\s+[axs][0-9]+} 4 } } */
+/* { dg-final { scan-assembler-times {fsrmi\s+[01234]} 4 } } */