patois / mrspicky

MrsPicky - An IDAPython decompiler script that helps auditing memcpy() and memmove() calls

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

MrsPicky

MrsPicky - An IDAPython decompiler script that helps auditing calls to the memcpy() and memmove() functions.

mrspicky animated gif

This example code shows how the HexRays decompiler can be scripted in order to identify potentially dangerous calls to memcpy() function calls. It is in no way meant to be a fully working script covering all possible use cases but just a few instead.

It will display a list of identified calls that can be and is meant to be searched, sorted and filtered interactively using IDA's built-in filtering features. Double clicking an entry will jump to the respective call within the currently active IDA or Decompiler view.

In cases where the "n" argument that is passed to memcpy() calls can be resolved statically, the resulting list's "max n" tab reflects the maximum number of bytes that the destination buffer "dst" can be written to (in other words: any number larger than that will corrupt whatever follows the current stack frame, which usually is a return address.

The "problems" tab may contain the following keywords:

  • "memcorr" - indicates a confirmed memory corruption
  • "argptr" - the "dst" pointer points beyond the local stack frame (this may not actually be a problem per se but...)

Feel free to adjust the script to suit your personal preferences. Relevant code is commented and explained below so that hopefully it will be easy to adapt the code to cover more use-cases as well as further functions such as malloc() whatsoever.

Requirements

This script is based on Python3 and requires IDA 7.3 to work. Python2 version is available here.

For further help, check out vds5.py that comes with the HexRays SDK.

This script is licensed under the "THE BEER-WARE LICENSE" (Revision 42) license.

About

MrsPicky - An IDAPython decompiler script that helps auditing memcpy() and memmove() calls

License:Other


Languages

Language:Python 100.0%