[BUG] Memcheck global read error in DistinctJoinTest.IntegerInnerJoin

[davidwendt]

Found in the nightly builds which run memcheck on libcudf gtests
https://github.com/rapidsai/cudf/actions/runs/8044585263/job/21968419310

[ RUN      ] DistinctJoinTest.IntegerInnerJoin
========= Invalid __global__ read of size 4 bytes
=========     at 0x22e20 in void cudf::detail::<unnamed>::distinct_join_probe_kernel<thrust::transform_iterator<cudf::detail::<unnamed>::build_keys_fn<const cudf::experimental::row::hash::device_row_hasher<cudf::hashing::detail::default_hash, cudf::nullate::DYNAMIC>, cudf::experimental::row::rhs_index_type>, thrust::counting_iterator<int, thrust::use_default, thrust::use_default, thrust::use_default>, thrust::use_default, thrust::use_default>, cuco::static_set_ref<cuco::pair<unsigned int, cudf::experimental::row::lhs_index_type>, (cuda::std::__4::__detail::thread_scope)1, cudf::detail::comparator_adapter<cudf::experimental::row::equality::strong_index_comparator_adapter<cudf::experimental::row::equality::device_row_comparator<(bool)0, cudf::nullate::DYNAMIC, cudf::experimental::row::equality::nan_equal_physical_equality_comparator>>>, cuco::linear_probing<(int)1, cudf::detail::hasher_adapter<thrust::identity<unsigned int>>>, cuco::aow_storage_ref<cuco::pair<unsigned int, cudf::experimental::row::lhs_index_type>, (int)1, cuco::window_extent<int, (unsigned long)18446744073709551615>>, cuco::op::find_tag>>(T1, int, T2, int *, int *, int *)
=========     by thread (244,0,0) in block (3,0,0)
=========     Address 0x7f06d200c68c is out of bounds
=========     and is 5 bytes after the nearest allocation at 0x7f06d2004800 of size 32392 bytes

Use the following to recreate the error

compute-sanitizer --tool memcheck gtests/JOIN_TEST --gtest_filter=DistinctJoinTest.IntegerInnerJoin --rmm_mode=cuda