What I don't understand is how this is a regression from the indicated commit.
We were previously calling `bus_event_queue_input_report(buf, size)` with `buf = report_buffer + 1` and `size = 10`, which ended up in `process_hid_report` with `report_buf = buf` and `report_len = 10`.
Now, the fixup happens in `process_hid_report`, but the code does the same thing as before: we offset report data to `report_buf + 1`, and set report length to 10. The checks have been moved from the unix side too.