Skip to content

2026.10.1

Latest

Choose a tag to compare

@EricZimmerman EricZimmerman released this 16 Oct 13:50
· 39 commits to master since this release

New

  • $LogFile: -f $LogFile exports every log record to CSV and/or JSON (--csv, --json). The output decodes:

    • FILE records written, attributes created and deleted, and file name index entries added and deleted (name, parent, timestamps, size);
    • directory entry updates, and $STANDARD_INFORMATION timestamp changes such as time stomping, with the previous values;
    • the target file and attribute of each operation, including index buffers, $UsnJrnl:$J and bitmaps, from the log's checkpoints;
    • transaction grouping (TransactionFirstLsn).

    With -m $MFT, target and parent paths are filled in. The cluster size is worked out from the log.

  • $J with $MFT (#56): with -f $J -m $MFT, ParentPath is the path each record's parent directory had when the record was written. The journal is rewound from newest to oldest (Yogesh Khatri's technique, CyberCX), so paths through directories that were later deleted, reused, renamed or moved resolve. On the DFIR Artifact Museum samples, this resolved all 14,031 PathUnknown paths in one $J. It works in USN order, so clock changes don't break it.

  • ParseErrors column at the end of the $MFT CSV, with each damaged record's problems.

  • --lowmem: about 30% less peak memory for about 1.4x the run time (workstation instead of server garbage collection).

Improvements and fixes

  • MFT libraries 2026.10.2: a loaded $MFT takes about 20% less memory, and loads are slightly faster (2 GB $MFT: 5.2 → 4.7 GB peak). The output is the same over a 22 file $MFT corpus.
  • 48 bit entry numbers. --de shows each index entry's own file and parent.
  • Exit codes: 0, or 1 for invalid arguments, files that can't be used and failed exports.
  • $J: --csv and --json together write both (JSON was skipped). JSON files are truncated before writing (a shorter run left old data at the end).
  • With -m, --csvf/--jsonf name only the -f output; the $MFT export used to overwrite it.

Reminder from 2026.10.0

  • The $SDS CSV has one row per security id (every entry used to be output twice, once from the mirror), plus a FromMirror column at the end for mirror copies that differ.

Full Changelog: 2026.10.0...2026.10.1