[flang] Added lowering for hlfir.cshift operation. (#118918)

Author: vzakhari

flang/lib/Optimizer/HLFIR/Transforms/LowerHLFIRIntrinsics.cpp

@@ -463,6 +463,37 @@ struct MatmulTransposeOpConversion
   }
 };
 
+class CShiftOpConversion : public HlfirIntrinsicConversion<hlfir::CShiftOp> {
+  using HlfirIntrinsicConversion<hlfir::CShiftOp>::HlfirIntrinsicConversion;
+
+  llvm::LogicalResult
+  matchAndRewrite(hlfir::CShiftOp cshift,
+                  mlir::PatternRewriter &rewriter) const override {
+    fir::FirOpBuilder builder{rewriter, cshift.getOperation()};
+    const mlir::Location &loc = cshift->getLoc();
+
+    llvm::SmallVector<IntrinsicArgument, 3> inArgs;
+    mlir::Value array = cshift.getArray();
+    inArgs.push_back({array, array.getType()});
+    mlir::Value shift = cshift.getShift();
+    inArgs.push_back({shift, shift.getType()});
+    inArgs.push_back({cshift.getDim(), builder.getI32Type()});
+
+    auto *argLowering = fir::getIntrinsicArgumentLowering("cshift");
+    llvm::SmallVector<fir::ExtendedValue, 3> args =
+        lowerArguments(cshift, inArgs, rewriter, argLowering);
+
+    mlir::Type scalarResultType =
+        hlfir::getFortranElementType(cshift.getType());
+
+    auto [resultExv, mustBeFreed] =
+        fir::genIntrinsicCall(builder, loc, "cshift", scalarResultType, args);
+
+    processReturnValue(cshift, resultExv, mustBeFreed, builder, rewriter);
+    return mlir::success();
+  }
+};
+
 class LowerHLFIRIntrinsics
     : public hlfir::impl::LowerHLFIRIntrinsicsBase<LowerHLFIRIntrinsics> {
 public:
@@ -475,7 +506,8 @@ class LowerHLFIRIntrinsics
                 AllOpConversion, AnyOpConversion, SumOpConversion,
                 ProductOpConversion, TransposeOpConversion, CountOpConversion,
                 DotProductOpConversion, MaxvalOpConversion, MinvalOpConversion,
-                MinlocOpConversion, MaxlocOpConversion>(context);
+                MinlocOpConversion, MaxlocOpConversion, CShiftOpConversion>(
+            context);
 
     // While conceptually this pass is performing dialect conversion, we use
     // pattern rewrites here instead of dialect conversion because this pass