Rewrite sparse dmatrix using callbacks. (#7092)

- Reduce dependency on dmlc parsers and provide an interface for users to load data by themselves.
- Remove use of threaded iterator and IO queue.
- Remove `page_size`.
- Make sure the number of pages in memory is bounded.
- Make sure the cache can not be violated.
- Provide an interface for internal algorithms to process data asynchronously.

tests/cpp/data/test_metainfo.cu

@@ -141,7 +141,7 @@ TEST(MetaInfo, DeviceExtend) {
   lhs.num_row_ = kRows;
   rhs.num_row_ = kRows;
 
-  lhs.Extend(rhs, true);
+  lhs.Extend(rhs, true, true);
   ASSERT_EQ(lhs.num_row_, kRows * 2);
   ASSERT_FALSE(lhs.labels_.HostCanRead());