[MLIR] Apply clang-tidy fixes for bugprone-argument-comment in VectorToGPU.cpp (NFC)

Author: joker-eph

mlir/lib/Conversion/VectorToGPU/VectorToGPU.cpp

@@ -741,7 +741,7 @@ creatLdMatrixCompatibleLoads(RewriterBase &rewriter, vector::TransferReadOp op,
   }
 
   // Adjust the load offset.
-  auto laneId = gpu::LaneIdOp::create(rewriter, loc, /*upperBound=*/nullptr);
+  auto laneId = gpu::LaneIdOp::create(rewriter, loc, /*upper_bound=*/nullptr);
   FailureOr<AffineMap> offsets =
       nvgpu::getLaneIdToLdMatrixMatrixCoord(rewriter, loc, *params);
   if (failed(offsets)) {
@@ -781,7 +781,7 @@ createNonLdMatrixLoads(RewriterBase &rewriter, vector::TransferReadOp op,
             "conversion to distributed non-ldmatrix compatible load");
   }
 
-  Value laneId = gpu::LaneIdOp::create(rewriter, loc, /*upperBound=*/nullptr);
+  Value laneId = gpu::LaneIdOp::create(rewriter, loc, /*upper_bound=*/nullptr);
 
   // This is the individual element type.
   Type loadedElType = regInfo->registerLLVMType;
@@ -915,7 +915,7 @@ convertTransferWriteToStores(RewriterBase &rewriter, vector::TransferWriteOp op,
     return rewriter.notifyMatchFailure(op, "not mma sync reg info");
 
   VectorType vectorType = getMmaSyncVectorOperandType(*regInfo);
-  Value laneId = gpu::LaneIdOp::create(rewriter, loc, /*upperBound=*/nullptr);
+  Value laneId = gpu::LaneIdOp::create(rewriter, loc, /*upper_bound=*/nullptr);
 
   for (unsigned i = 0; i < vectorType.getShape()[0]; i++) {
     Value logicalValueId = arith::ConstantOp::create(