[4/11][POC][Clang][RISCV] Define tuple type variant of vsseg2e32

[eopXD]

Summary

For the cover letter of this patch-set, please checkout D146872.

Depends on D147731.

This is the 4th patch of the patch-set.

This patch is a proof-of-concept and will be extended to full coverage
in the future. Currently, the old non-tuple unit-stride segment store is
not removed, and only signed integer unit-strided segment store of NF=2,
EEW=32 is defined here.

[eopXD]

Minor code change: Separate defm of segment load and store under riscv_vector.td.

[eopXD]

Rebase upon update of parent patch.

[craig.topper]

clang/include/clang/Basic/riscv_vector.td

1571

Offset variable

[eopXD]

ddress comment from Craig.

[craig.topper]

clang/include/clang/Basic/riscv_vector.td

1571

Similar comment to patch 3. We can index the Ops array directly in the push_back calls instead of having temporary variables.

[eopXD]

Address comments from Craig.

[eopXD]

Removed another variable that is too verbose.

[craig.topper]

LGTM

[eopXD]

Rebase.

clang/include/clang/Basic/riscv_vector.td

foreach type = ["i"] in {
     }
     }] in {
           defvar T = "(Tuple:" # nf # ")";
           def : RVVBuiltin<"v", T # "vPCe", type>;
       }
     }
   }
 }
+
+multiclass RVVUnitStridedSegStoreTuple<string op> {
+  foreach type = ["i"] in {
+    defvar eew = !cond(!eq(type, "i") : "32");
+    foreach nf = [2] in {
+      let Name = op # nf # "e" # eew # "_v_tuple",
+          OverloadedName = op # nf # "e" # eew # "_tuple",
+          IRName = op # nf,
+          MaskedIRName = op # nf # "_mask",
+          NF = nf,
+          HasMaskedOffOperand = false,
+          ManualCodegen = [{
+    {
+      // Masked
+      // Builtin: (mask, ptr, v_tuple, vl)
+      // Intrinsic: (val0, val1, ..., ptr, mask, vl)
+      // Unmasked
+      // Builtin: (ptr, v_tuple, vl)
+      // Intrinsic: (val0, val1, ..., ptr, vl)
+      unsigned Offset = IsMasked ? 1 : 0;
+      llvm::Value *VTupleOperand = Ops[Offset + 1];
+
+      SmallVector<llvm::Value*, 12> Operands;
+      for (unsigned I = 0; I < NF; ++I) {
+        llvm::Value *V = Builder.CreateExtractValue(VTupleOperand, {I});
+        Operands.push_back(V);
+      }
+      Operands.push_back(Ops[Offset]); // Ptr
+      if (IsMasked)
+        Operands.push_back(Ops[0]);
+      Operands.push_back(Ops[Offset + 2]); // VL
+
+      IntrinsicTypes = {Operands[0]->getType(), Operands.back()->getType()};
+      llvm::Function *F = CGM.getIntrinsic(ID, IntrinsicTypes);
+      return Builder.CreateCall(F, Operands, "");
+   }
+      }] in {
+          defvar T = "(Tuple:" # nf # ")";
+          def : RVVBuiltin<"v", "0Pe" # T # "v", type>;
+      }
+    }
+  }
+}
 // TODO: Extend for policy
 let UnMaskedPolicyScheme = NonePolicy,
     MaskedPolicyScheme = NonePolicy,
     IsTuple = true in {
 defm : RVVUnitStridedSegLoadTuple<"vlseg">;
 }
 
 let UnMaskedPolicyScheme = NonePolicy,
+    MaskedPolicyScheme = NonePolicy,
+    IsTuple = true in {
+defm : RVVUnitStridedSegStoreTuple<"vsseg">;
+}
+
+
+let UnMaskedPolicyScheme = NonePolicy,
     MaskedPolicyScheme = NonePolicy in {
 defm : RVVUnitStridedSegStore<"vsseg">;
 defm : RVVStridedSegStore<"vssseg">;
 defm : RVVIndexedSegStore<"vsuxseg">;
 defm : RVVIndexedSegStore<"vsoxseg">;
 }
 
 // 12. Vector Integer Arithmetic Instructions

clang/lib/CodeGen/CGBuiltin.cpp

Value *CodeGenFunction::EmitRISCVBuiltinExpr(unsigned BuiltinID,
   }
 
   if (BuiltinID == RISCV::BI__builtin_riscv_ntl_load)
     ICEArguments |= (1 << 1);
   if (BuiltinID == RISCV::BI__builtin_riscv_ntl_store)
     ICEArguments |= (1 << 2);
 
   for (unsigned i = 0, e = E->getNumArgs(); i != e; i++) {
+    // Handle aggregate argument, namely RVV tuple types in segment load/store
+    if (hasAggregateEvaluationKind(E->getArg(i)->getType())) {
+      LValue L = EmitAggExprToLValue(E->getArg(i));
+      llvm::Value *AggValue = Builder.CreateLoad(L.getAddress(*this));
+      Ops.push_back(AggValue);
+      continue;
+    }
+
     // If this is a normal argument, just emit it as a scalar.
     if ((ICEArguments & (1 << i)) == 0) {
       Ops.push_back(EmitScalarExpr(E->getArg(i)));
       continue;
     }
 
     // If this is required to be a constant, constant fold it so that we know
     // that the generated intrinsic gets a ConstantInt.

clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/non-policy/non-overloaded/vsseg2e32_tuple.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 2
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v -target-feature +zfh \
+// RUN:   -target-feature +experimental-zvfh -disable-O0-optnone  \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local void @test_vsseg2e32_v_tuple_i32m1
+// CHECK-RV64-SAME: (ptr noundef [[BASE:%.*]], <vscale x 2 x i32> [[V_TUPLE_COERCE0:%.*]], <vscale x 2 x i32> [[V_TUPLE_COERCE1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = insertvalue { <vscale x 2 x i32>, <vscale x 2 x i32> } poison, <vscale x 2 x i32> [[V_TUPLE_COERCE0]], 0
+// CHECK-RV64-NEXT:    [[TMP1:%.*]] = insertvalue { <vscale x 2 x i32>, <vscale x 2 x i32> } [[TMP0]], <vscale x 2 x i32> [[V_TUPLE_COERCE1]], 1
+// CHECK-RV64-NEXT:    [[TMP2:%.*]] = extractvalue { <vscale x 2 x i32>, <vscale x 2 x i32> } [[TMP1]], 0
+// CHECK-RV64-NEXT:    [[TMP3:%.*]] = extractvalue { <vscale x 2 x i32>, <vscale x 2 x i32> } [[TMP1]], 1
+// CHECK-RV64-NEXT:    call void @llvm.riscv.vsseg2.nxv2i32.i64(<vscale x 2 x i32> [[TMP2]], <vscale x 2 x i32> [[TMP3]], ptr [[BASE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret void
+//
+void test_vsseg2e32_v_tuple_i32m1(int32_t *base, vint32m1x2_t v_tuple, size_t vl) {
+  return __riscv_vsseg2e32_v_tuple_i32m1(base, v_tuple, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local void @test_vsseg2e32_v_i32m1_m
+// CHECK-RV64-SAME: (<vscale x 2 x i1> [[MASK:%.*]], ptr noundef [[BASE:%.*]], <vscale x 2 x i32> [[V0:%.*]], <vscale x 2 x i32> [[V1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    call void @llvm.riscv.vsseg2.mask.nxv2i32.i64(<vscale x 2 x i32> [[V0]], <vscale x 2 x i32> [[V1]], ptr [[BASE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret void
+//
+void test_vsseg2e32_v_i32m1_m(vbool32_t mask, int32_t *base, vint32m1_t v0, vint32m1_t v1, size_t vl) {
+  return __riscv_vsseg2e32_v_i32m1_m(mask, base, v0, v1, vl);
+}