Add builtins for bitreverse intrinsic

[arsenm]

Summary

Follow the naming convention that bswap uses since it's a
similar sort of operation.

[hfinkel]

LGTM.

[arsenm]

r259671.

Should I put some documentation for these somewhere? docs/LanguageExtensions.rst seems to be the place, but it only has a random subset of the current builtins.

[hfinkel]

In D16482#343177, @arsenm wrote:

r259671.

Should I put some documentation for these somewhere? docs/LanguageExtensions.rst seems to be the place, but it only has a random subset of the current builtins.

My impression is that we prioritize documenting the builtins that don't have GCC equivalents. Thus, documenting them there along with the other assorted Clang-specific builtins seems reasonable.

include/clang/Basic/Builtins.def

 BUILTIN(__builtin_popcountll, "iULLi", "nc")
 
 // FIXME: These type signatures are not correct for targets with int != 32-bits
 // or with ULL != 64-bits.
 BUILTIN(__builtin_bswap16, "UsUs", "nc")
 BUILTIN(__builtin_bswap32, "UiUi", "nc")
 BUILTIN(__builtin_bswap64, "ULLiULLi", "nc")
 
+BUILTIN(__builtin_bitreverse16, "UsUs", "nc")
+BUILTIN(__builtin_bitreverse32, "UiUi", "nc")
+BUILTIN(__builtin_bitreverse64, "ULLiULLi", "nc")
+
 // Random GCC builtins
 BUILTIN(__builtin_constant_p, "i.", "nctu")
 BUILTIN(__builtin_classify_type, "i.", "nctu")
 BUILTIN(__builtin___CFStringMakeConstantString, "FC*cC*", "nc")
 BUILTIN(__builtin___NSStringMakeConstantString, "FC*cC*", "nc")
 BUILTIN(__builtin_va_start, "vA.", "nt")
 BUILTIN(__builtin_va_end, "vA", "n")
 BUILTIN(__builtin_va_copy, "vAA", "n")

lib/CodeGen/CGBuiltin.cpp

assert(X->getType() == Y->getType() &&
          "arguments have the same integer width?)");
 
   llvm::Value *Callee = CGF.CGM.getIntrinsic(IntrinsicID, X->getType());
   llvm::Value *Tmp = CGF.Builder.CreateCall(Callee, {X, Y});
   Carry = CGF.Builder.CreateExtractValue(Tmp, 1);
   return CGF.Builder.CreateExtractValue(Tmp, 0);
 }
 
+// Emit a simple mangled intrinsic that has 1 argument and a return type
+// matching the argument type.
+static Value *emitUnaryBuiltin(CodeGenFunction &CGF,
+                               const CallExpr *E,
+                               unsigned IntrinsicID) {
+  llvm::Value *Src0 = CGF.EmitScalarExpr(E->getArg(0));
+
+  Value *F = CGF.CGM.getIntrinsic(IntrinsicID, Src0->getType());
+  return CGF.Builder.CreateCall(F, Src0);
+}
+
+// Emit an intrinsic that has 3 float or double operands.
+static Value *emitTernaryFPBuiltin(CodeGenFunction &CGF,
+                                   const CallExpr *E,
+                                   unsigned IntrinsicID) {
+  llvm::Value *Src0 = CGF.EmitScalarExpr(E->getArg(0));
+  llvm::Value *Src1 = CGF.EmitScalarExpr(E->getArg(1));
+  llvm::Value *Src2 = CGF.EmitScalarExpr(E->getArg(2));
+
+  Value *F = CGF.CGM.getIntrinsic(IntrinsicID, Src0->getType());
+  return CGF.Builder.CreateCall(F, {Src0, Src1, Src2});
+}
+
+// Emit an intrinsic that has 1 float or double operand, and 1 integer.
+static Value *emitFPIntBuiltin(CodeGenFunction &CGF,
+                               const CallExpr *E,
+                               unsigned IntrinsicID) {
+  llvm::Value *Src0 = CGF.EmitScalarExpr(E->getArg(0));
+  llvm::Value *Src1 = CGF.EmitScalarExpr(E->getArg(1));
+
+  Value *F = CGF.CGM.getIntrinsic(IntrinsicID, Src0->getType());
+  return CGF.Builder.CreateCall(F, {Src0, Src1});
+}
+
 namespace {
   struct WidthAndSignedness {
     unsigned Width;
     bool Signed;
   };
 }
 
 static WidthAndSignedness
case Builtin::BI__builtin_assume: {
 
     Value *ArgValue = EmitScalarExpr(E->getArg(0));
     Value *FnAssume = CGM.getIntrinsic(Intrinsic::assume);
     return RValue::get(Builder.CreateCall(FnAssume, ArgValue));
   }
   case Builtin::BI__builtin_bswap16:
   case Builtin::BI__builtin_bswap32:
   case Builtin::BI__builtin_bswap64: {
-    Value *ArgValue = EmitScalarExpr(E->getArg(0));
-    llvm::Type *ArgType = ArgValue->getType();
-    Value *F = CGM.getIntrinsic(Intrinsic::bswap, ArgType);
-    return RValue::get(Builder.CreateCall(F, ArgValue));
+    return RValue::get(emitUnaryBuiltin(*this, E, Intrinsic::bswap));
+  }
+  case Builtin::BI__builtin_bitreverse16:
+  case Builtin::BI__builtin_bitreverse32:
+  case Builtin::BI__builtin_bitreverse64: {
+    return RValue::get(emitUnaryBuiltin(*this, E, Intrinsic::bitreverse));
   }
   case Builtin::BI__builtin_object_size: {
     unsigned Type =
         E->getArg(1)->EvaluateKnownConstInt(getContext()).getZExtValue();
     auto *ResType = cast<llvm::IntegerType>(ConvertType(E->getType()));
 
     // We pass this builtin onto the optimizer so that it can figure out the
     // object size in more complex cases.
switch (BuiltinID) {
         return Builder.CreateFSub(Zero, FsubRes, "sub");
     }
     llvm_unreachable("Unknown FMA operation");
     return nullptr; // Suppress no-return warning
   }
   }
 }
 
-// Emit an intrinsic that has 1 float or double.
-static Value *emitUnaryFPBuiltin(CodeGenFunction &CGF,
-                                 const CallExpr *E,
-                                 unsigned IntrinsicID) {
-  llvm::Value *Src0 = CGF.EmitScalarExpr(E->getArg(0));
-
-  Value *F = CGF.CGM.getIntrinsic(IntrinsicID, Src0->getType());
-  return CGF.Builder.CreateCall(F, Src0);
-}
-
-// Emit an intrinsic that has 3 float or double operands.
-static Value *emitTernaryFPBuiltin(CodeGenFunction &CGF,
-                                   const CallExpr *E,
-                                   unsigned IntrinsicID) {
-  llvm::Value *Src0 = CGF.EmitScalarExpr(E->getArg(0));
-  llvm::Value *Src1 = CGF.EmitScalarExpr(E->getArg(1));
-  llvm::Value *Src2 = CGF.EmitScalarExpr(E->getArg(2));
-
-  Value *F = CGF.CGM.getIntrinsic(IntrinsicID, Src0->getType());
-  return CGF.Builder.CreateCall(F, {Src0, Src1, Src2});
-}
-
-// Emit an intrinsic that has 1 float or double operand, and 1 integer.
-static Value *emitFPIntBuiltin(CodeGenFunction &CGF,
-                               const CallExpr *E,
-                               unsigned IntrinsicID) {
-  llvm::Value *Src0 = CGF.EmitScalarExpr(E->getArg(0));
-  llvm::Value *Src1 = CGF.EmitScalarExpr(E->getArg(1));
-
-  Value *F = CGF.CGM.getIntrinsic(IntrinsicID, Src0->getType());
-  return CGF.Builder.CreateCall(F, {Src0, Src1});
-}
-
 Value *CodeGenFunction::EmitAMDGPUBuiltinExpr(unsigned BuiltinID,
                                               const CallExpr *E) {
   switch (BuiltinID) {
   case AMDGPU::BI__builtin_amdgcn_div_scale:
   case AMDGPU::BI__builtin_amdgcn_div_scalef: {
     // Translate from the intrinsics's struct return to the builtin's out
     // argument.
 
Value *CodeGenFunction::EmitAMDGPUBuiltinExpr(unsigned BuiltinID,
   case AMDGPU::BI__builtin_amdgcn_div_fixup:
   case AMDGPU::BI__builtin_amdgcn_div_fixupf:
     return emitTernaryFPBuiltin(*this, E, Intrinsic::amdgcn_div_fixup);
   case AMDGPU::BI__builtin_amdgcn_trig_preop:
   case AMDGPU::BI__builtin_amdgcn_trig_preopf:
     return emitFPIntBuiltin(*this, E, Intrinsic::amdgcn_trig_preop);
   case AMDGPU::BI__builtin_amdgcn_rcp:
   case AMDGPU::BI__builtin_amdgcn_rcpf:
-    return emitUnaryFPBuiltin(*this, E, Intrinsic::amdgcn_rcp);
+    return emitUnaryBuiltin(*this, E, Intrinsic::amdgcn_rcp);
   case AMDGPU::BI__builtin_amdgcn_rsq:
   case AMDGPU::BI__builtin_amdgcn_rsqf:
-    return emitUnaryFPBuiltin(*this, E, Intrinsic::amdgcn_rsq);
+    return emitUnaryBuiltin(*this, E, Intrinsic::amdgcn_rsq);
   case AMDGPU::BI__builtin_amdgcn_rsq_clamped:
   case AMDGPU::BI__builtin_amdgcn_rsq_clampedf:
-    return emitUnaryFPBuiltin(*this, E, Intrinsic::amdgcn_rsq_clamped);
+    return emitUnaryBuiltin(*this, E, Intrinsic::amdgcn_rsq_clamped);
   case AMDGPU::BI__builtin_amdgcn_ldexp:
   case AMDGPU::BI__builtin_amdgcn_ldexpf:
     return emitFPIntBuiltin(*this, E, Intrinsic::amdgcn_ldexp);
   case AMDGPU::BI__builtin_amdgcn_class:
   case AMDGPU::BI__builtin_amdgcn_classf:
     return emitFPIntBuiltin(*this, E, Intrinsic::amdgcn_class);
 
     // Legacy amdgpu prefix
   case AMDGPU::BI__builtin_amdgpu_rsq:
   case AMDGPU::BI__builtin_amdgpu_rsqf: {
     if (getTarget().getTriple().getArch() == Triple::amdgcn)
-      return emitUnaryFPBuiltin(*this, E, Intrinsic::amdgcn_rsq);
-    return emitUnaryFPBuiltin(*this, E, Intrinsic::r600_rsq);
+      return emitUnaryBuiltin(*this, E, Intrinsic::amdgcn_rsq);
+    return emitUnaryBuiltin(*this, E, Intrinsic::r600_rsq);
   }
   case AMDGPU::BI__builtin_amdgpu_ldexp:
   case AMDGPU::BI__builtin_amdgpu_ldexpf: {
     if (getTarget().getTriple().getArch() == Triple::amdgcn)
       return emitFPIntBuiltin(*this, E, Intrinsic::amdgcn_ldexp);
     return emitFPIntBuiltin(*this, E, Intrinsic::AMDGPU_ldexp);
   }
   default:

test/CodeGen/builtins.c

#define V(n,args) p(#n #args, (__builtin_##n args, 0))
   P(object_size, (s0, 2));
   P(object_size, (s0, 3));
 
   // Whatever
 
   P(bswap16, (N));
   P(bswap32, (N));
   P(bswap64, (N));
+
+  // CHECK: @llvm.bitreverse.i16
+  // CHECK: @llvm.bitreverse.i32
+  // CHECK: @llvm.bitreverse.i64
+  P(bitreverse16, (N));
+  P(bitreverse32, (N));
+  P(bitreverse64, (N));
+
   // FIXME
   // V(clear_cache, (&N, &N+1));
   V(trap, ());
   R(extract_return_addr, (&N));
   P(signbit, (1.0));
 
   return 0;
 }