DF-0930 race_ntfs run log (unpatched #0 kernel, kern.maxvnodes=50, 120s)
=========================================================================

Guest: DragonFly 6.5-DEVELOPMENT #0: Thu Jul  2 06:02:54 UTC 2026
NTFS: ntfs.ko loaded, /dev/vn0 -> /mnt/ntfs (ntfs_file.img, inode 32 = "target")

Command: ./race_ntfs /mnt/ntfs/target 120

Output:
[*] DF-0930 race: 3 lookup + 2 stat + 2 churn threads, 6 cpus, 120 sec
[*] target: /mnt/ntfs/target
[*] if the kernel panics in ntfs_ntget/lockmgr on a freed ntnode, the UAF is confirmed.
[*] race finished without panic. The UAF window may be too tight to hit in this run.
RACE_EXIT=0

Guest status after run: UP (no panic)
dmesg: no kernel warnings related to NTFS slab/lock corruption

Notes:
  The race did NOT trigger a panic. This is expected for a CVSS AC:High race
  (the window between lwkt_reltoken at ntfs_ihash.c:100 and ip->i_usecount++
  at ntfs_subr.c:348 is only 2-5 instructions). The bug is confirmed by
  code-level trace (see VERDICT.md), not by runtime panic.

  Three runs were attempted:
    Run 1: 30s, kern.maxvnodes=300  — no panic
    Run 2: 60s, kern.maxvnodes=100  — no panic
    Run 3: 120s, kern.maxvnodes=50  — no panic
  Total race time: ~210 seconds across ~6M open/close iterations.