Patch CUB DeviceSegmentedSort and remove workaround (#12234)

Fix in CUB DeviceSegmentedSort allows for workaround to removed. The CUB fix is applied as a patch in the libcudf build process.
Reference NVIDIA/cub#594 and #12217

Authors:
  - David Wendt (https://github.com/davidwendt)

Approvers:
  - Elias Stehle (https://github.com/elstehle)
  - Robert Maynard (https://github.com/robertmaynard)

URL: #12234

cpp/cmake/thirdparty/patches/cub_segmented_sort_with_bool_key.diff

@@ -0,0 +1,14 @@
+diff --git a/dependencies/cub/cub/agent/agent_sub_warp_merge_sort.cuh b/dependencies/cub/cub/agent/agent_sub_warp_merge_sort.cuh
+index ad65f2a3..ad45a21e 100644
+--- a/dependencies/cub/cub/agent/agent_sub_warp_merge_sort.cuh
++++ b/dependencies/cub/cub/agent/agent_sub_warp_merge_sort.cuh
+@@ -221,7 +221,8 @@ public:
+       using UnsignedBitsT = typename Traits<KeyT>::UnsignedBits;
+       UnsignedBitsT default_key_bits = IS_DESCENDING ? Traits<KeyT>::LOWEST_KEY
+                                                      : Traits<KeyT>::MAX_KEY;
+-      KeyT oob_default = reinterpret_cast<KeyT &>(default_key_bits);
++      KeyT oob_default = std::is_same_v<KeyT, bool> ? !IS_DESCENDING
++                                                   : reinterpret_cast<KeyT &>(default_key_bits);
+
+       WarpLoadKeysT(storage.load_keys)
+         .Load(keys_input, keys, segment_size, oob_default);

cpp/cmake/thirdparty/patches/thrust_override.json

@@ -27,6 +27,11 @@
           "file" : "${current_json_dir}/thrust_faster_scan_compile_times.diff",
           "issue" : "Improve Thrust scan compile times by reducing the number of kernels generated [https://github.com/rapidsai/cudf/pull/8183]",
           "fixed_in" : ""
+        },
+        {
+          "file" : "${current_json_dir}/cub_segmented_sort_with_bool_key.diff",
+          "issue" : "Fix an error in CUB DeviceSegmentedSort when the keys are bool type [https://github.com/NVIDIA/cub/issues/594]",
+          "fixed_in" : "2.1"
         }
       ]
     }

cpp/src/sort/segmented_sort.cu

@@ -52,7 +52,7 @@ struct column_fast_sort_fn {
   static bool is_fast_sort_supported(column_view const& col)
   {
     return !col.has_nulls() and
-           ((cudf::is_integral(col.type()) && !cudf::is_boolean(col.type())) ||
+           (cudf::is_integral(col.type()) ||
             (cudf::is_fixed_point(col.type()) and (col.type().id() != type_id::DECIMAL128)));
   }