Creating New Symbol Tables

This page details how symbol tables are located and used by Volatility, and documents the tools and methods that can be used to make new symbol tables.

How Volatility finds symbol tables

All files are stored as JSON data, they can be in pure JSON files as .json, or compressed as .json.gz or .json.xz. Volatility will automatically decompress them on use. It will also cache their contents (compressed) when used, located under the user’s home directory, in .cache/volatility3, along with other useful data. The cache directory currently cannot be altered.

Symbol table JSON files live, by default, under the volatility3/symbols, underneath an operating system directory (currently one of windows, mac or linux). The symbols directory is configurable within the framework and can usually be set within the user interface.

These files can also be compressed into ZIP files, which Volatility will process in order to locate symbol files. The ZIP file must be named after the appropriate operating system (such as linux.zip, mac.zip or windows.zip). Inside the ZIP file, the directory structure should match the uncompressed operating system directory.

Windows symbol tables

For Windows systems, Volatility accepts a string made up of the GUID and Age of the required PDB file. It then searches all files under the configured symbol directories under the windows subdirectory. Any that match the filename pattern of <pdb-name>/<GUID>-<AGE>.json (or any compressed variant) will be used. If such a symbol table cannot be found, then the associated PDB file will be downloaded from Microsoft’s Symbol Server and converted into the appropriate JSON format, and will be saved in the correct location.

Windows symbol tables can be manually constructed from an appropriate PDB file. The primary tool for doing this is built into Volatility 3, called pdbconv.py. It can be run from the top-level Volatility path, using the following command:

PYTHONPATH="." python volatility3/framework/symbols/windows/pdbconv.py

The PYTHONPATH environment variable is not required if the Volatility library is installed in the system’s library path or a virtual environment.

Mac/Linux symbol tables

For Mac/Linux systems, both use the same mechanism for identification. JSON files live under the symbol directories, under either the linux or mac directories. The generated files contain an identifying string (the operating system banner), which Volatility’s automagic can detect. Volatility caches the mapping between the strings and the symbol tables they come from, meaning the precise file names don’t matter and can be organized under any necessary hierarchy under the operating system directory.

Linux and Mac symbol tables can be generated from a DWARF file using a tool called dwarf2json. Currently a kernel with debugging symbols is the only suitable means for recovering all the information required by most Volatility plugins. Note that in most linux distributions, the standard kernel is stripped of debugging information and the kernel with debugging information is stored in a package that must be acquired separately.

A generic table isn’t guaranteed to produce accurate results, and would reduce the number of structures that all plugins could rely on. As such, and because linux kernels with different configurations can produce different structures, volatility 3 requires that the banners in the JSON file match the banners found in the image exactly, not just the version number. This can include elements such as the compilation time and even the version of gcc used for the compilation. The exact match is required to ensure that the results volatility returns are accurate, therefore there is no simple means provided to get the wrong JSON ISF file to easily match.

To determine the string for a particular memory image, use the banners plugin. Once the specific banner is known, try to locate that exact kernel debugging package for the operating system. Unfortunately each distribution provides its debugging packages under different package names and there are so many that the distribution may not keep all old versions of the debugging symbols, and therefore it may not be possible to find the right symbols to analyze a linux memory image with volatlity. With Macs there are far fewer kernels and only one distribution, making it easier to ensure that the right symbols can be found.

Once a kernel with debugging symbols/appropriate DWARF file has been located, dwarf2json will convert it into an appropriate JSON file. Example code for automatically creating a JSON from URLs for the kernel debugging package and the package containing the System.map, can be found in stock-linux-json.py . The System.map file is recommended for completeness, but a kernel with debugging information often contains the same symbol offsets within the DWARF data, which dwarf2json can extract into the JSON ISF file.

The banners available for volatility to use can be found using the isfinfo plugin, but this will potentially take a long time to run depending on the number of JSON files available. This will list all the JSON (ISF) files that volatility3 is aware of, and for linux/mac systems what banner string they search for. For volatility to use the JSON file, the banners must match exactly (down to the compilation date).