[09/11] sim: riscv: Add double precision floating-point MAC instructions

Message ID 20240226142845.1629113-1-bhushan.attarde@imgtec.com
State New
Headers
Series sim: riscv: simulation of single and double precision floating point instructions |

Checks

Context Check Description
linaro-tcwg-bot/tcwg_gdb_build--master-aarch64 success Testing passed
linaro-tcwg-bot/tcwg_gdb_build--master-arm success Testing passed
linaro-tcwg-bot/tcwg_gdb_check--master-arm success Testing passed
linaro-tcwg-bot/tcwg_gdb_check--master-aarch64 success Testing passed

Commit Message

Bhushan Attarde Feb. 26, 2024, 2:28 p.m. UTC
  From: Bhushan Attarde <bhushan.attarde@imgtec.com>

Added simulation of following single precision floating-point instructions
fmadd.d, fnmadd.d, fmsub.d and fnmsub.d.

Added test file sim/testsuite/riscv/d-basic-arith.s to test these instructions.
---
 sim/riscv/sim-main.c                | 117 ++++++++++++++++++++++++++--
 sim/testsuite/riscv/d-basic-arith.s |  78 +++++++++++++++++++
 2 files changed, 190 insertions(+), 5 deletions(-)
 create mode 100644 sim/testsuite/riscv/d-basic-arith.s
  

Patch

diff --git a/sim/riscv/sim-main.c b/sim/riscv/sim-main.c
index 4f347fbfc5e..4a102df74e0 100644
--- a/sim/riscv/sim-main.c
+++ b/sim/riscv/sim-main.c
@@ -1361,20 +1361,36 @@  float64_compare (SIM_CPU *cpu, int rd, int rs1, int rs2, int flags)
 
 /* Handle double precision floating point math instructions.  */
 static void
-float64_math (SIM_CPU *cpu, int rd, int rs1, int rs2, int flags)
+float64_math (SIM_CPU *cpu, int rd, int rs1, int rs2, int rs3, int rm,
+	      int flags)
 {
   struct riscv_sim_cpu *riscv_cpu = RISCV_SIM_CPU (cpu);
-  double a, b, result = 0;
-  uint64_t rs1_bits, rs2_bits, rd_bits;
+  double a, b, c, result = 0;
+  int old_rm, old_except, new_except;
+  uint64_t rs1_bits, rs2_bits, rs3_bits, rd_bits;
   const char *frd_name = riscv_fpr_names_abi[rd];
   const char *frs1_name = riscv_fpr_names_abi[rs1];
   const char *frs2_name = riscv_fpr_names_abi[rs2];
+  const char *frs3_name = riscv_fpr_names_abi[rs3];
+
+  if (rm == DYN)
+    rm = riscv_cpu->csr.frm;
+
+  old_rm = set_riscv_rounding_mode (rm);
+  old_except = fetestexcept (FE_ALL_EXCEPT);
 
   rs1_bits = (uint64_t) riscv_cpu->fpregs[rs1];
   memcpy (&a, &rs1_bits, sizeof (a));
   rs2_bits = (uint64_t) riscv_cpu->fpregs[rs2];
   memcpy (&b, &rs2_bits, sizeof (b));
 
+  if (flags == FMADD || flags == FNMADD
+      || flags == FMSUB || flags == FNMSUB)
+    {
+      rs3_bits = (uint64_t) riscv_cpu->fpregs[rs3];
+      memcpy (&c, &rs3_bits, sizeof (c));
+    }
+
   switch (flags)
     {
     case FMAX:
@@ -1385,12 +1401,85 @@  float64_math (SIM_CPU *cpu, int rd, int rs1, int rs2, int flags)
       TRACE_INSN (cpu, "fmin.d %s, %s, %s;", frd_name, frs1_name, frs2_name);
       result = fmin (a, b);
       break;
+    case FMADD:
+      TRACE_INSN (cpu, "fmadd.d %s, %s, %s, %s, rm=%d;",
+		  frd_name, frs1_name, frs2_name, frs3_name, rm);
+      result = (a * b) + c;
+      break;
+    case FNMADD:
+      TRACE_INSN (cpu, "fnmadd.d %s, %s, %s, %s, rm=%d;",
+		  frd_name, frs1_name, frs2_name, frs3_name, rm);
+      result = -((a * b) - c);
+      break;
+    case FMSUB:
+      TRACE_INSN (cpu, "fmsub.d %s, %s, %s, %s, rm=%d;",
+		  frd_name, frs1_name, frs2_name, frs3_name, rm);
+      result = (a * b) - c;
+      break;
+    case FNMSUB:
+      TRACE_INSN (cpu, "fnmsub.d %s, %s, %s, %s, rm=%d;",
+		  frd_name, frs1_name, frs2_name, frs3_name, rm);
+      result = -((a * b) + c);
+      break;
+    }
+
+  if (rm == RMM)
+    {
+      if (is_float_halfway (result))
+	{
+	  if (result > 0)
+	    result = nextafterf (result, INFINITY);
+	  else
+	    result = nextafterf (result, -INFINITY);
+	}
     }
 
   /* Store result.  */
   memcpy (&rd_bits, &result, sizeof (result));
   store_fp (cpu, rd, rd_bits);
 
+  /* Restore rounding mode.  */
+  fesetround (old_rm);
+
+  /* Set exception.  */
+  new_except = fetestexcept (FE_ALL_EXCEPT);
+
+  if (old_except != new_except)
+    {
+      if (new_except & FE_OVERFLOW)
+	{
+	  riscv_cpu->csr.fcsr |= FCSR_OF;
+	  riscv_cpu->csr.fflags |= FCSR_OF;
+	  TRACE_REGISTER (cpu, "wrote CSR fcsr |= OF");
+	}
+      else if (new_except & FE_UNDERFLOW)
+	{
+	  riscv_cpu->csr.fcsr |= FCSR_UF;
+	  riscv_cpu->csr.fflags |= FCSR_UF;
+	  TRACE_REGISTER (cpu, "wrote CSR fcsr |= UF");
+	}
+      else if (new_except & FE_INEXACT)
+	{
+	  riscv_cpu->csr.fcsr |= FCSR_NX;
+	  riscv_cpu->csr.fflags |= FCSR_NX;
+	  TRACE_REGISTER (cpu, "wrote CSR fcsr |= NX");
+	}
+      else if (new_except & FE_DIVBYZERO)
+	{
+	  riscv_cpu->csr.fcsr |= FCSR_DZ;
+	  riscv_cpu->csr.fflags |= FCSR_DZ;
+	  TRACE_REGISTER (cpu, "wrote CSR fcsr |= DZ");
+	}
+      else if (new_except & FE_INVALID)
+	{
+	  riscv_cpu->csr.fcsr |= FCSR_NV;
+	  riscv_cpu->csr.fflags |= FCSR_NV;
+	  TRACE_REGISTER (cpu, "wrote CSR fcsr |= NV");
+	}
+
+      feclearexcept (FE_ALL_EXCEPT);
+      feraiseexcept (old_except);
+    }
 }
 
 /* Simulate single precision floating point instructions.  */
@@ -1644,6 +1733,8 @@  execute_d (SIM_CPU *cpu, unsigned_word iw, const struct riscv_opcode *op)
   int rd = (iw >> OP_SH_RD) & OP_MASK_RD;
   int rs1 = (iw >> OP_SH_RS1) & OP_MASK_RS1;
   int rs2 = (iw >> OP_SH_RS2) & OP_MASK_RS2;
+  int rs3 = (iw >> OP_SH_RS3) & OP_MASK_RS3;
+  int rm = (iw >> OP_SH_RM) & OP_MASK_RM;
   const char *frd_name = riscv_fpr_names_abi[rd];
   const char *rd_name = riscv_gpr_names_abi[rd];
   const char *frs1_name = riscv_fpr_names_abi[rs1];
@@ -1728,10 +1819,26 @@  execute_d (SIM_CPU *cpu, unsigned_word iw, const struct riscv_opcode *op)
 	break;
       }
     case MATCH_FMIN_D:
-      float64_math (cpu, rd, rs1, rs2, FMIN);
+      float64_math (cpu, rd, rs1, rs2, 0, -1, FMIN);
       break;
     case MATCH_FMAX_D:
-      float64_math (cpu, rd, rs1, rs2, FMAX);
+      float64_math (cpu, rd, rs1, rs2, 0, -1, FMAX);
+      break;
+    case MATCH_FMADD_D:
+    case MATCH_FMADD_D | MASK_RM:
+      float64_math (cpu, rd, rs1, rs2, rs3, rm, FMADD);
+      break;
+    case MATCH_FNMADD_D:
+    case MATCH_FNMADD_D | MASK_RM:
+      float64_math (cpu, rd, rs1, rs2, rs3, rm, FNMADD);
+      break;
+    case MATCH_FMSUB_D:
+    case MATCH_FMSUB_D | MASK_RM:
+      float64_math (cpu, rd, rs1, rs2, rs3, rm, FMSUB);
+      break;
+    case MATCH_FNMSUB_D:
+    case MATCH_FNMSUB_D | MASK_RM:
+      float64_math (cpu, rd, rs1, rs2, rs3, rm, FNMSUB);
       break;
     default:
       TRACE_INSN (cpu, "UNHANDLED INSN: %s", op->name);
diff --git a/sim/testsuite/riscv/d-basic-arith.s b/sim/testsuite/riscv/d-basic-arith.s
new file mode 100644
index 00000000000..996f603e91d
--- /dev/null
+++ b/sim/testsuite/riscv/d-basic-arith.s
@@ -0,0 +1,78 @@ 
+# Double precision basic arithmetic tests.
+# mach: riscv64
+# sim(riscv64): --model RV64ID
+# ld(riscv64): -m elf64lriscv
+# as(riscv64): -march=rv64id
+
+.include "testutils.inc"
+
+	.section	.data
+	.align 3
+
+_arg1:
+	.double -12.5
+
+_arg2:
+	.double 2.5
+
+_arg3:
+	.double 7.45
+
+_result:
+	.double -23.799999
+	.double 38.7000008
+	.double -38.7000008
+	.double 23.7999992
+
+	start
+	.option push
+	.option norelax
+	la	a0,_arg1
+	la	a1,_arg2
+	la	a2,_arg3
+	la	a3,_result
+	li	a4,1
+	.option pop
+
+	# Test fmadd instruction.
+	fld	fa0,0(a0)
+	fld	fa1,0(a1)
+	fld	fa2,0(a2)
+	fld	fa3,0(a3)
+	fmadd.d	fa4,fa0,fa1,fa0,rne
+	feq.d	a5,fa4,fa4
+	bne	a5,a4,test_fail
+
+	# Test fnmadd instruction.
+	fld	fa0,0(a0)
+	fld	fa1,0(a1)
+	fld	fa2,0(a2)
+	fld	fa3,8(a3)
+	fnmadd.d	fa4,fa0,fa1,fa0,rne
+	feq.d	a5,fa4,fa4
+	bne	a5,a4,test_fail
+
+	# Test fmsub instruction.
+	fld	fa0,0(a0)
+	fld	fa1,0(a1)
+	fld	fa2,0(a2)
+	fld	fa3,16(a3)
+	fmsub.d	fa4,fa0,fa1,fa0,rne
+	feq.d	a5,fa4,fa4
+	bne	a5,a4,test_fail
+
+	# Test fnmsub instruction.
+	fld	fa0,0(a0)
+	fld	fa1,0(a1)
+	fld	fa2,0(a2)
+	fld	fa3,24(a3)
+	fmsub.d	fa4,fa0,fa1,fa0,rne
+	feq.d	a5,fa4,fa4
+	bne	a5,a4,test_fail
+
+
+test_pass:
+	pass
+
+test_fail:
+	fail