[riscv64] Implement simd128 in swap

enable simd on riscv64

Change-Id: I446d6b14e4f89164b49a66367340d904ba104911
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3347493Reviewed-by: ji qiu <qiuji@iscas.ac.cn>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Yahan Lu <yahan@iscas.ac.cn>
Cr-Commit-Position: refs/heads/main@{#78553}

src/compiler/backend/riscv64/code-generator-riscv64.cc

@@ -1723,7 +1723,13 @@ CodeGenerator::CodeGenResult CodeGenerator::AssembleArchInstruction(
     case kRiscvStoreToStackSlot: {
       if (instr->InputAt(0)->IsFPRegister()) {
         if (instr->InputAt(0)->IsSimd128Register()) {
-          UNREACHABLE();
+          Register dst = sp;
+          if (i.InputInt32(1) != 0) {
+            dst = kScratchReg2;
+            __ Add64(kScratchReg2, sp, Operand(i.InputInt32(1)));
+          }
+          __ VU.set(kScratchReg, E8, m1);
+          __ vs(i.InputSimd128Register(0), dst, 0, E8);
         } else {
           __ StoreDouble(i.InputDoubleRegister(0),
                          MemOperand(sp, i.InputInt32(1)));
@@ -4192,17 +4198,18 @@ void CodeGenerator::AssembleMove(InstructionOperand* source,
       VRegister src = g.ToSimd128Register(source);
       if (destination->IsSimd128Register()) {
         VRegister dst = g.ToSimd128Register(destination);
+        __ VU.set(kScratchReg, E8, m1);
         __ vmv_vv(dst, src);
       } else {
         DCHECK(destination->IsSimd128StackSlot());
-        Register dst = g.ToMemOperand(destination).offset() == 0
-                           ? g.ToMemOperand(destination).rm()
-                           : kScratchReg;
-        if (g.ToMemOperand(destination).offset() != 0) {
-          __ Add64(dst, g.ToMemOperand(destination).rm(),
-                   g.ToMemOperand(destination).offset());
+        __ VU.set(kScratchReg, E8, m1);
+        MemOperand dst = g.ToMemOperand(destination);
+        Register dst_r = dst.rm();
+        if (dst.offset() != 0) {
+          dst_r = kScratchReg;
+          __ Add64(dst_r, dst.rm(), dst.offset());
         }
-        __ vs(src, dst, 0, E8);
+        __ vs(src, dst_r, 0, E8);
       }
     } else {
       FPURegister src = g.ToDoubleRegister(source);
@@ -4224,24 +4231,25 @@ void CodeGenerator::AssembleMove(InstructionOperand* source,
     MemOperand src = g.ToMemOperand(source);
     MachineRepresentation rep = LocationOperand::cast(source)->representation();
     if (rep == MachineRepresentation::kSimd128) {
-      Register src_reg = src.offset() == 0 ? src.rm() : kScratchReg;
+      __ VU.set(kScratchReg, E8, m1);
+      Register src_r = src.rm();
       if (src.offset() != 0) {
-        __ Add64(src_reg, src.rm(), src.offset());
+        src_r = kScratchReg;
+        __ Add64(src_r, src.rm(), src.offset());
       }
       if (destination->IsSimd128Register()) {
-        __ vl(g.ToSimd128Register(destination), src_reg, 0, E8);
+        __ vl(g.ToSimd128Register(destination), src_r, 0, E8);
       } else {
         DCHECK(destination->IsSimd128StackSlot());
         VRegister temp = kSimd128ScratchReg;
-        Register dst = g.ToMemOperand(destination).offset() == 0
-                           ? g.ToMemOperand(destination).rm()
-                           : kScratchReg;
-        if (g.ToMemOperand(destination).offset() != 0) {
-          __ Add64(dst, g.ToMemOperand(destination).rm(),
-                   g.ToMemOperand(destination).offset());
+        MemOperand dst = g.ToMemOperand(destination);
+        Register dst_r = dst.rm();
+        if (dst.offset() != 0) {
+          dst_r = kScratchReg2;
+          __ Add64(dst_r, dst.rm(), dst.offset());
         }
-        __ vl(temp, src_reg, 0, E8);
-        __ vs(temp, dst, 0, E8);
+        __ vl(temp, src_r, 0, E8);
+        __ vs(temp, dst_r, 0, E8);
       }
     } else {
       if (destination->IsFPRegister()) {
@@ -4272,91 +4280,106 @@ void CodeGenerator::AssembleMove(InstructionOperand* source,
 void CodeGenerator::AssembleSwap(InstructionOperand* source,
                                  InstructionOperand* destination) {
   RiscvOperandConverter g(this, nullptr);
-  // Dispatch on the source and destination operand kinds.  Not all
-  // combinations are possible.
-  if (source->IsRegister()) {
-    // Register-register.
-    Register temp = kScratchReg;
-    Register src = g.ToRegister(source);
-    if (destination->IsRegister()) {
-      Register dst = g.ToRegister(destination);
-      __ Move(temp, src);
-      __ Move(src, dst);
-      __ Move(dst, temp);
-    } else {
-      DCHECK(destination->IsStackSlot());
-      MemOperand dst = g.ToMemOperand(destination);
-      __ Move(temp, src);
-      __ Ld(src, dst);
-      __ Sd(temp, dst);
-    }
-  } else if (source->IsStackSlot()) {
-    DCHECK(destination->IsStackSlot());
-    Register temp_0 = kScratchReg;
-    Register temp_1 = kScratchReg2;
-    MemOperand src = g.ToMemOperand(source);
-    MemOperand dst = g.ToMemOperand(destination);
-    __ Ld(temp_0, src);
-    __ Ld(temp_1, dst);
-    __ Sd(temp_0, dst);
-    __ Sd(temp_1, src);
-  } else if (source->IsFPRegister()) {
-    MachineRepresentation rep = LocationOperand::cast(source)->representation();
-    if (rep == MachineRepresentation::kSimd128) {
-      UNIMPLEMENTED();
-    } else {
-      FPURegister temp = kScratchDoubleReg;
-      FPURegister src = g.ToDoubleRegister(source);
-      if (destination->IsFPRegister()) {
-        FPURegister dst = g.ToDoubleRegister(destination);
+  switch (MoveType::InferSwap(source, destination)) {
+    case MoveType::kRegisterToRegister:
+      if (source->IsRegister()) {
+        Register temp = kScratchReg;
+        Register src = g.ToRegister(source);
+        Register dst = g.ToRegister(destination);
         __ Move(temp, src);
         __ Move(src, dst);
         __ Move(dst, temp);
       } else {
-        DCHECK(destination->IsFPStackSlot());
+        if (source->IsFloatRegister() || source->IsDoubleRegister()) {
+          FPURegister temp = kScratchDoubleReg;
+          FPURegister src = g.ToDoubleRegister(source);
+          FPURegister dst = g.ToDoubleRegister(destination);
+          __ Move(temp, src);
+          __ Move(src, dst);
+          __ Move(dst, temp);
+        } else {
+          DCHECK(source->IsSimd128Register());
+          VRegister src = g.ToDoubleRegister(source).toV();
+          VRegister dst = g.ToDoubleRegister(destination).toV();
+          VRegister temp = kSimd128ScratchReg;
+          __ VU.set(kScratchReg, E8, m1);
+          __ vmv_vv(temp, src);
+          __ vmv_vv(src, dst);
+          __ vmv_vv(dst, temp);
+        }
+      }
+      return;
+    case MoveType::kRegisterToStack: {
+      MemOperand dst = g.ToMemOperand(destination);
+      if (source->IsRegister()) {
+        Register temp = kScratchReg;
+        Register src = g.ToRegister(source);
+        __ mv(temp, src);
+        __ Ld(src, dst);
+        __ Sd(temp, dst);
+      } else {
         MemOperand dst = g.ToMemOperand(destination);
-        if (rep == MachineRepresentation::kFloat32) {
-          __ MoveFloat(temp, src);
+        if (source->IsFloatRegister()) {
+          DoubleRegister src = g.ToDoubleRegister(source);
+          DoubleRegister temp = kScratchDoubleReg;
+          __ fmv_s(temp, src);
           __ LoadFloat(src, dst);
           __ StoreFloat(temp, dst);
-        } else {
-          DCHECK_EQ(rep, MachineRepresentation::kFloat64);
-          __ MoveDouble(temp, src);
+        } else if (source->IsDoubleRegister()) {
+          DoubleRegister src = g.ToDoubleRegister(source);
+          DoubleRegister temp = kScratchDoubleReg;
+          __ fmv_d(temp, src);
           __ LoadDouble(src, dst);
           __ StoreDouble(temp, dst);
+        } else {
+          DCHECK(source->IsSimd128Register());
+          VRegister src = g.ToDoubleRegister(source).toV();
+          VRegister temp = kSimd128ScratchReg;
+          __ VU.set(kScratchReg, E8, m1);
+          __ vmv_vv(temp, src);
+          Register dst_v = dst.rm();
+          if (dst.offset() != 0) {
+            dst_v = kScratchReg2;
+            __ Add64(dst_v, dst.rm(), Operand(dst.offset()));
+          }
+          __ vl(src, dst_v, 0, E8);
+          __ vs(temp, dst_v, 0, E8);
         }
       }
+      return;
     }
-  } else if (source->IsFPStackSlot()) {
-    DCHECK(destination->IsFPStackSlot());
-    Register temp_0 = kScratchReg;
-    MemOperand src0 = g.ToMemOperand(source);
-    MemOperand src1(src0.rm(), src0.offset() + kIntSize);
-    MemOperand dst0 = g.ToMemOperand(destination);
-    MemOperand dst1(dst0.rm(), dst0.offset() + kIntSize);
-    MachineRepresentation rep = LocationOperand::cast(source)->representation();
-    if (rep == MachineRepresentation::kSimd128) {
-      UNIMPLEMENTED();
-    } else {
-      FPURegister temp_1 = kScratchDoubleReg;
-      if (rep == MachineRepresentation::kFloat32) {
-        __ LoadFloat(temp_1, dst0);  // Save destination in temp_1.
-        __ Lw(temp_0, src0);  // Then use temp_0 to copy source to destination.
-        __ Sw(temp_0, dst0);
-        __ StoreFloat(temp_1, src0);
+    case MoveType::kStackToStack: {
+      MemOperand src = g.ToMemOperand(source);
+      MemOperand dst = g.ToMemOperand(destination);
+      if (source->IsSimd128StackSlot()) {
+        __ VU.set(kScratchReg, E8, m1);
+        Register src_v = src.rm();
+        Register dst_v = dst.rm();
+        if (src.offset() != 0) {
+          src_v = kScratchReg;
+          __ Add64(src_v, src.rm(), Operand(src.offset()));
+        }
+        if (dst.offset() != 0) {
+          dst_v = kScratchReg2;
+          __ Add64(dst_v, dst.rm(), Operand(dst.offset()));
+        }
+        __ vl(kSimd128ScratchReg, src_v, 0, E8);
+        __ vl(kSimd128ScratchReg2, dst_v, 0, E8);
+        __ vs(kSimd128ScratchReg, dst_v, 0, E8);
+        __ vs(kSimd128ScratchReg2, src_v, 0, E8);
       } else {
-        DCHECK_EQ(rep, MachineRepresentation::kFloat64);
-        __ LoadDouble(temp_1, dst0);  // Save destination in temp_1.
-        __ Lw(temp_0, src0);  // Then use temp_0 to copy source to destination.
-        __ Sw(temp_0, dst0);
-        __ Lw(temp_0, src1);
-        __ Sw(temp_0, dst1);
-        __ StoreDouble(temp_1, src0);
+        UseScratchRegisterScope scope(tasm());
+        Register temp_0 = kScratchReg;
+        Register temp_1 = kScratchReg2;
+        __ Ld(temp_0, src);
+        __ Ld(temp_1, dst);
+        __ Sd(temp_0, dst);
+        __ Sd(temp_1, src);
       }
+      return;
     }
-  } else {
-    // No other combinations are possible.
-    UNREACHABLE();
+    default:
+      UNREACHABLE();
   }
 }
 

tools/testrunner/base_runner.py

@@ -694,10 +694,6 @@ class BaseTestRunner(object):
        utils.GuessPowerProcessorVersion() < 9:
        no_simd_hardware = True
 
-    # riscv64 do not support Simd instructions
-    if self.build_config.arch == 'riscv64':
-       no_simd_hardware = True
-
     return {
       "arch": self.build_config.arch,
       "asan": self.build_config.asan,