1eb8fff7e1
Previously we considered certain fields of `tar` or `zip` file headers to be sources, but this meant subsequent references to the same field were not considered sanitized. For example, at least some real-world projects used a pattern like `if isIllegalPathTraversal(hdr.Name) { return nil; } ... /* other code using hdr.Name */`. By associating a source with the field-read `.Name` rather than the header itself, we were unable to see that the subsequent read was guarded by the sanitizer function.
Relatedly, it is common to use some intermediary taint-propagating function, as in `clean(s string) { if strings.HasPrefix("..", filepath.Clean(filepath.Join(target, s))) ...`, in the implementation of a sanitizer. We now follow the taint propagation (locally) backwards towards the function parameter, marking the predecessor functions and ultimately the parameter `s` as sanitized in addition to the direct argument to `strings.HasPrefix`. Existing sanitizing-function logic can then sometimes lift this out into the caller too.
Go analysis support for CodeQL
This open-source repository contains the extractor, CodeQL libraries, and queries that power Go support in LGTM and the other CodeQL products that GitHub makes available to its customers worldwide.
It contains two major components:
- an extractor, itself written in Go, that parses Go source code and converts it into a database that can be queried using CodeQL.
- static analysis libraries and queries written in QL that can be used to analyze such a database to find coding mistakes or security vulnerabilities.
The goal of this project is to provide comprehensive static analysis support for Go in CodeQL.
For the queries and libraries that power CodeQL support for other languages, visit the CodeQL repository.
Installation
Simply clone this repository. There are no external dependencies.
If you want to use the CodeQL extension for Visual Studio Code, import this repository into your VS Code workspace.
Usage
To analyze a Go codebase, either use the CodeQL command-line interface to create a database yourself, or download a pre-built database from LGTM.com. You can then run any of the queries contained in this repository either on the command line or using the VS Code extension.
Note that the lgtm.com branch of this repository corresponds to the version of the queries that is currently deployed on LGTM.com. The main branch may contain changes that have not been deployed yet, so you may need to upgrade databases downloaded from LGTM.com before running queries on them.
Contributions
Contributions are welcome! Please see our contribution guidelines and our code of conduct for details on how to participate in our community.
Licensing
The code in this repository is licensed under the MIT license.