Skip to main content
This is a pre-production deployment of Warehouse. Changes made here affect the production instance of PyPI (pypi.python.org).
Help us improve Python packaging - Donate today!

Search C Structures in a process' memory

Project Description

Quick Start:

Quick usage guide in the docs/ folder.

Haystack-reverse CLI in the docs/ folder.

More documentation

Introduction:

python-haystack is an heap analysis framework, focused on searching and reversing of C structure in allocated memory.

The first function/API is the SEARCH function. It gives the ability to search for known record types in a process memory dump or live process’s memory.

The second function/API is the REVERSE function in the extension python-haystack-reverse It aims at helping an analyst in reverse engineering the memory records types present in a process heap. It focuses on reconstruction, classification of classic C structures from memory. It attempts to recreate types definition.

Scripts & Memory handler format:

A few entry points exists to handle the format your memory dump.

Memory dump folder produced by haystack-live-dump

  • haystack-find-heap allows to show details on Windows HEAP.
  • haystack-search search CLI
  • haystack-show show CLI for specific record type at a specific address

You can use the following URL to designate your memory handler/dump:

  • dir:///path/to/my/haystack/fump/folder to use the haystack dump format
  • dmp:///path/to/my/minidump/file use the minidump format (microsoft?)
  • frida://name_or_pid_of_process_to_attach_to use frida to access a live process memory
  • live://name_or_pid_of_process_to_attach_to ptrace a live process
  • rekall:// load a rekall image
  • volatility:// load a volatility image
  • cuckoo:// load a memory dump produced by Cuckoo (beta might need patch)

How to get a memory dump:

On Windows, the most straightforward is to get a Minidump. The Microsoft Sysinternals suite of tools provide either a CLI (procdump.exe) or a GUI (Process explorer). Using one of these (with full memory dump option) you will produce a file that can be used with the haystack-minidump-xxx list of entry points.

While technically you could use many third party tool, haystack actually need memory mapping information to work with. So there is a dumping tool included haystack-live-dump:

# haystack-live-dump <pid> myproc.dump

For live processes

  • haystack-live-dump capture a process memory dump to a folder (haystack format)

For a Rekall memory dump

  • haystack-rekall-dump dump a specific process to a haystack process dump

For a Volatility memory dump

  • haystack-volatility-dump dump a specific process to a haystack process dump

You can easily reproduce the format of the dump, its a folder/archive containing each memory map in a separate file :

  • memory content in a file named after it’s start/end addresses ( 0x000700000-0x000800000 )
  • ‘mappings’ file containing memory mappings metadata. ( mappings )

Or you can code a haystack.abc.IMemoryMapping implementation for your favorite format.

Otherwise, if you already have a system memory dump from Volatility or Rekall, you can use the haystack-rekall-xxx or haystack-volatility-xxx families of entry points to extract a specific process memory into a file.

Verifying Windows Heap attributes:

The entry point haystack-find-heap allows to show details on Windows HEAP. It should support:

  • Windows XP 32 bits
  • Windows XP 64 bits
  • Windows 7 32 bits
  • Windows 7 64 bits

and show details of the Look Aside List (LAL) and Low Fragmentation Heap (LFH) frontend.

You might be surprised to see that sometimes, a single process can mix the two types of HEAP (32 & 64).

Search for known structures:

To search for a specific record, you will first need to define that record type. A [quick usage guide](docs/Haystack basic usage.ipynb) is available to go over the basic steps to go from a C Header file to a Python ctypes definition. Or you can do it yourself, with traditional Python ctypes records.

The search api is available through the haystack-xxx-search family of scripts but also in an API so that you can embed that search in your own code.

In short, the haystack search will iterate over every offset of the program’s memory to try and find ‘valid’ offset for that specific record type.

The validity of the record is determined by type constraints such as: - pointer field should have valid address space values - user-defined type constraints (see ‘Constraints file’ section below) - etc..

$ python haystack/cli.py dir:///home/user/project/python-haystack/test/src/test-ctypes6.32.dump ctypes6_gen32.struct_usual

Constraints file:

The following constraints are supported:
  • IgnoreMember: The value of this field will be ignored. Useful to Ignore pointer fields.
  • NotNull: The value of this field must not be 0.
  • RangeValue(x,y): the field must have a value between x and y.
  • PerfectMatch(‘hello world’): the field (a string) must match ‘hello world’
  • [1,2,3]: A list of values that the fields should have
  • [1, RangeValue(12,16), 42]: The field value should be 1, 12-16 or 42.

Example:

[struct_name]
myfield: [1,0xff]
ptr_field: NotNull

You can take a look a haystack/allocators/win32/winxpheap32.constraints, where the constraints of a Windows XP HEAP x32 are defined.

Obviously, the more constraints, the better the results will be.

Dynamic constraints definition:

You can also create more complex constraints using python code by implementing a haystack.abc.interface.IRecordTypeDynamicConstraintsValidator class and feeding it to the ModuleConstraints.set_dynamic_constraints

Command line example:

sslsnoop repository needs an update to be compatible with releases > v0.30 - pending

For example, this will dump the session_state structures + pointed children structures as an python object that we can play with. Lets assume we have an ssh client or server as pid 4042:

$ sudo haystack-live-search --pickled 4042 sslsnoop.ctypes_openssh.session_state search > instance.pickled
$ sudo haystack-live-search --pickled 4042 sslsnoop.ctypes_openssh.session_state refresh 0xb8b70d18 > instance.pickled
$ sudo haystack-live-search --pickled <pid> <your ctypes Structure> search

Graphic User Interface :

This is not working right now

There is also an attempt at a Graphical UI python-haystack-gui

python API example:

See the quick usage guide

How to define your own structures:

The most easy way is to use ctypeslib to generate ctypes records from C Headers.

Or define your python ctypes record by hand.

Extension examples :

@ see sslsnoop in the Pypi repo. openssl and nss structures are generated.

@ see ctypes-kernel on my github. Linux kernel structure are generated from a build kernel tree. (VMM is abitch)

not so FAQ :

What does it do ?:

The basic functionality is to search in a process’ memory for a specific C Record.

The extended reverse engineering functionality aims at reversing structures from memory/heap analysis.

How do it knows that the structures is valid ? :

You add some constraints on the record fields expected values. Pointers are always constrained to valid memory space.

Where does the idea comes from ? :

passe-partout originally. since I started in March 2011, I have uncovered several other related previous work.

Most of them are in the docs/ folder.

Other related work are mona.py from Immunity, some other Mandiant stuff…

In a nutshell, this is probably not an original idea. But yet, I could not find a operational standalone lib for live memory extraction for my sslsnoop PoC, so….

Related work

What are the dependencies ? :

  • python-ptrace on linux
  • winappdbg on win32 ( not sure if working, feedback welcome)
  • python-numpy
  • python-networkx
  • python-levenshtein
  • several others…

Others

http://ntinfo.biz/ xntsv32

Release History

Release History

This version
History Node

0.42

History Node

0.41

History Node

0.40

History Node

0.36

History Node

0.35

History Node

0.34

History Node

0.33

History Node

0.32

History Node

0.31

History Node

0.30

History Node

0.29

History Node

0.28

History Node

0.27

History Node

0.26

History Node

0.25

History Node

0.24

History Node

0.23

History Node

0.22

History Node

0.21

History Node

0.20

History Node

0.17

History Node

0.16

History Node

0.15

History Node

0.13

History Node

0.12

History Node

0.11

History Node

0.10

History Node

0.9

History Node

0.8

History Node

0.7

History Node

0.6

History Node

0.5

History Node

0.4

History Node

0.3

History Node

0.2

Download Files

Download Files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

File Name & Checksum SHA256 Checksum Help Version File Type Upload Date
haystack-0.42-py2.py3-none-any.whl (179.6 kB) Copy SHA256 Checksum SHA256 2.7 Wheel Jul 3, 2017
haystack-0.42.tar.gz (158.5 kB) Copy SHA256 Checksum SHA256 Source Jul 3, 2017

Supported By

WebFaction WebFaction Technical Writing Elastic Elastic Search Pingdom Pingdom Monitoring Dyn Dyn DNS Sentry Sentry Error Logging CloudAMQP CloudAMQP RabbitMQ Heroku Heroku PaaS Kabu Creative Kabu Creative UX & Design Fastly Fastly CDN DigiCert DigiCert EV Certificate Rackspace Rackspace Cloud Servers DreamHost DreamHost Log Hosting