[MLIR] Apply clang-tidy fixes for llvm-qualified-auto in ShardToMPI.cpp (NFC)

Author: joker-eph

mlir/lib/Conversion/ShardToMPI/ShardToMPI.cpp

@@ -301,7 +301,7 @@ class ConvertProcessLinearIndexOp
                   ConversionPatternRewriter &rewriter) const override {
     // Create mpi::CommRankOp
     Location loc = op.getLoc();
-    auto ctx = op.getContext();
+    auto *ctx = op.getContext();
     Value commWorld =
         mpi::CommWorldOp::create(rewriter, loc, mpi::CommType::get(ctx));
     auto rank = mpi::CommRankOp::create(
@@ -520,7 +520,7 @@ struct ConvertShardShapeOp : public OpConversionPattern<ShardShapeOp> {
 };
 
 static mpi::MPI_ReductionOpEnumAttr getMPIReductionOp(ReductionKindAttr kind) {
-  auto ctx = kind.getContext();
+  auto *ctx = kind.getContext();
   auto getReductionOp = [ctx](mpi::MPI_ReductionOpEnum redOp) {
     return mpi::MPI_ReductionOpEnumAttr::get(ctx, redOp);
   };