FYI, we noticed a +618 bytes kernel size regression due to commit:

commit: 045fa86d3b099e9650900425dcdebc9685c105a9 (rcu/tiny: Support reclaim for head-less object)
url: https://github.com/0day-ci/linux/commits/Joel-Fernandes-Google/kfree_rcu-improvements-for-rcu-dev/20200330-113719

Details as below (size data is obtained by `nm --size-sort vmlinux`):

7123dd03: rcu/tiny: Move kvfree_call_rcu() out of header
045fa86d: rcu/tiny: Support reclaim for head-less object

+----------------------------+----------+----------+-------+
|           symbol           | 7123dd03 | 045fa86d | delta |
+----------------------------+----------+----------+-------+
| bzImage                    | 440560   | 440848   | 288   |
| nm.T.kvfree_call_rcu       | 5        | 209      | 204   |
| nm.T.rcu_init              | 25       | 157      | 132   |
| nm.t.kvfree_rcu_monitor    | 0        | 96       | 96    |
| nm.t.kvfree_rcu_work       | 0        | 81       | 81    |
| nm.b.monitor_work          | 0        | 44       | 44    |
| nm.b.rcu_work              | 0        | 28       | 28    |
| nm.t.rcu_process_callbacks | 146      | 153      | 7     |
| nm.T.call_rcu              | 46       | 51       | 5     |
| nm.b.kvcache               | 0        | 4        | 4     |
| nm.b.kvhead                | 0        | 4        | 4     |
| nm.b.kvhead_free           | 0        | 4        | 4     |
| nm.b.rcu_init_done         | 0        | 4        | 4     |
| nm.T.rcu_qs                | 36       | 40       | 4     |
| nm.b.monitor_todo          | 0        | 1        | 1     |
+----------------------------+----------+----------+-------+



Thanks,
Kbuild test robot