"Out of swap" process killer endless loop

One of my machines ran out of swap and tried to kill off a process.
Unfortunately the process it picked was swapped out and seems to be
unkillable:

swap_pager_getswapspace(16): failed
pid 55455 (mtree), uid 0, was killed: out of swap space
pid 55455 (mtree), uid 0, was killed: out of swap space
pid 55455 (mtree), uid 0, was killed: out of swap space
pid 55455 (mtree), uid 0, was killed: out of swap space
pid 55455 (mtree), uid 0, was killed: out of swap space
pid 55455 (mtree), uid 0, was killed: out of swap space
KDB: enter: Break sequence on console
[thread pid 16 tid 100008 ]
Stopped at      kdb_enter+0x31: leave
db> wh 55455
Tracing pid 55455 tid 100190 td 0xffffff0089aea000
sched_switch() at sched_switch+0x130
mi_switch() at mi_switch+0x28e
sleepq_switch() at sleepq_switch+0x106
sleepq_wait() at sleepq_wait+0x67
_sleep() at _sleep+0x2db
vm_waitpfault() at vm_waitpfault+0x65
vm_fault() at vm_fault+0x148
trap_pfault() at trap_pfault+0x130
trap() at trap+0x287
calltrap() at calltrap+0x8
--- trap 0xc, rip = 0x80076a3cc, rsp = 0x7fffffff65d8, rbp = 0x780 ---

55455     1 22461     0  SLJ     pfault   0xffffffff805fba98 mtree

Actually they are all swapped out (I guess the rest of swap is in use
by a swap-backed md, although I don't know how it managed to fill up).
Is there anything that can be done in this situation?

kris