Fix a page leak in the md(4) swap I/O path.

[markj]

Summary

In a last-minute change to D22885 I removed the vm_page_activate() call.
But this is of course needed to ensure that the pages end up in the page
queues in the first place.

Restore the pre-r356157 logic. Now, without the page lock, the
vm_page_active() check is racy, but this race is harmless.

Reported and tested by: pho