DF-0853: NO PANIC OBSERVED.

The bug is type confusion (CWE-704), not memory corruption. In 3+ mount
attempts on the unpatched 6.5-DEVELOPMENT #0 kernel the guest stayed up;
the only observable effect was:
  - dmesg: "cd9660: High Sierra Format" logged for a Standard ISO9660 PVD
  - mount succeeded (rc=0) but ls /mnt/df0853 returned ENOTDIR
    (root_extent misread at Sierra offsets)

No fatal trap, no KASSERT, no Stopped-at in dfbsd-qemu/boot.log.