Libmodsecurity is one component of the ModSecurity v3 project. The library codebase serves as an interface to ModSecurity Connectors taking in web traffic and applying traditional ModSecurity processing. In general, it provides the capability to load/interpret rules written in the ModSecurity SecRules format and apply them to HTTP content provided by your application via Connectors.
If you are looking for ModSecurity for Apache (aka ModSecurity v2.x), it is still under maintenance and available: here.
- All Apache dependencies have been removed
- Higher performance
- New features
- New architecture
Libmodsecurity is a complete rewrite of the ModSecurity platform. When it was first devised the ModSecurity project started as just an Apache module. Over time the project has been extended, due to popular demand, to support other platforms including (but not limited to) Nginx and IIS. In order to provide for the growing demand for additional platform support, it has became necessary to remove the Apache dependencies underlying this project, making it more platform independent.
As a result of this goal we have rearchitected Libmodsecurity such that it is no longer dependent on the Apache web server (both at compilation and during runtime). One side effect of this is that across all platforms users can expect increased performance. Additionally, we have taken this opportunity to lay the groundwork for some new features that users have been long seeking. For example we are looking to natively support auditlogs in the JSON format, along with a host of other functionality in future versions.
The 'ModSecurity' branch no longer contains the traditional module logic (for Nginx, Apache, and IIS) that has traditionally been packaged all together. Instead, this branch only contains the library portion (libmodsecurity) for this project. This library is consumed by what we have termed 'Connectors' these connectors will interface with your webserver and provide the library with a common format that it understands. Each of these connectors is maintained as a separate GitHub project. For instance, the Nginx connector is supplied by the ModSecurity-nginx project (https://github.com/owasp-modsecurity/ModSecurity-nginx).
Keeping these connectors separated allows each project to have different release cycles, issues and development trees. Additionally, it means that when you install ModSecurity v3 you only get exactly what you need, no extras you won't be using.
Before starting the compilation process, make sure that you have all the dependencies in place. Read the subsection “Dependencies” for further information.
After the compilation make sure that there are no issues on your build/platform. We strongly recommend the utilization of the unit tests and regression tests. These test utilities are located under the subfolder ‘tests’.
As a dynamic library, don’t forget that libmodsecurity must be installed to a location (folder) where you OS will be looking for dynamic libraries.
On unix the project uses autotools to help the compilation process. Please note that if you are working with git, don't forget to initialize and update the submodules. Here's a quick how-to:
$ git clone --recursive https://github.com/owasp-modsecurity/ModSecurity ModSecurity
$ cd ModSecurityYou can then start the build process:
$ ./build.sh
$ ./configure
$ make
$ sudo make installDetails on distribution specific builds can be found in our Wiki: Compilation Recipes
Windows build information can be found here.
This library is written in C++ using the C++17 standards. It also uses Flex and Yacc to produce the “Sec Rules Language” parser. Other, mandatory dependencies include YAJL, as ModSecurity uses JSON for producing logs and its testing framework, libpcre (not yet mandatory) for processing regular expressions in SecRules, and libXML2 (not yet mandatory) which is used for parsing XML requests.
All others dependencies are related to operators specified within SecRules or configuration directives and may not be required for compilation. A short list of such dependencies is as follows:
- libinjection is needed for the operator @detectXSS and @detectSQL
- curl is needed for the directive SecRemoteRules.
If those libraries are missing ModSecurity will be compiled without the support for the operator @detectXSS and the configuration directive SecRemoteRules.
The library documentation is written within the code in Doxygen format. To generate this documentation, please use the doxygen utility with the provided configuration file, “doxygen.cfg”, located with the "doc/" subfolder. This will generate HTML formatted documentation including usage examples.
The library provides a C++ and C interface. Some resources are currently only available via the C++ interface, for instance, the capability to create custom logging mechanism (see the regression test to check for how those logging mechanism works). The objective is to have both APIs (C, C++) providing the same functionality, if you find an aspect of the API that is missing via a particular interface, please open an issue.
Inside the subfolder examples, there are simple examples on how to use the API. Below some are illustrated:
using ModSecurity::ModSecurity;
using ModSecurity::Rules;
using ModSecurity::Transaction;
ModSecurity *modsec;
ModSecurity::Rules *rules;
modsec = new ModSecurity();
rules = new Rules();
rules->loadFromUri(rules_file);
Transaction *modsecTransaction = new Transaction(modsec, rules);
modsecTransaction->processConnection("127.0.0.1");
if (modsecTransaction->intervention()) {
std::cout << "There is an intervention" << std::endl;
}#include "modsecurity/modsecurity.h"
#include "modsecurity/transaction.h"
char main_rule_uri[] = "basic_rules.conf";
int main (int argc, char **argv)
{
ModSecurity *modsec = NULL;
Transaction *transaction = NULL;
Rules *rules = NULL;
modsec = msc_init();
rules = msc_create_rules_set();
msc_rules_add_file(rules, main_rule_uri);
transaction = msc_new_transaction(modsec, rules);
msc_process_connection(transaction, "127.0.0.1");
msc_process_uri(transaction, "http://www.modsecurity.org/test?key1=value1&key2=value2&key3=value3&test=args&test=test");
msc_process_request_headers(transaction);
msc_process_request_body(transaction);
msc_process_response_headers(transaction);
msc_process_response_body(transaction);
return 0;
}You are more than welcome to contribute to this project and look forward to growing the community around this new version of ModSecurity. Areas of interest include: New functionalities, fixes, bug report, support for beginning users, or anything that you are willing to help with.
We prefer to have your patch within the GitHub infrastructure to facilitate our review work, and our Q.A. integration. GitHub provides excellent documentation on how to perform “Pull Requests”, more information available here: https://help.github.com/articles/using-pull-requests/
Please respect the coding style. Pull requests can include various commits, so provide one fix or one piece of functionality per commit. Please do not change anything outside the scope of your target work (e.g. coding style in a function that you have passed by). For further information about the coding style used in this project, please check: https://www.chromium.org/blink/coding-style
Provides explanative commit messages. Your first line should give the highlights of your patch, 3rd and on give a more detailed explanation/technical details about your patch. Patch explanation is valuable during the review process.
Within our code there are various items marked as TODO or FIXME that may need your attention. Check the list of items by performing a grep:
$ cd /path/to/modsecurity-nginx
$ egrep -Rin "TODO|FIXME" -R *
A TODO list is also available as part of the Doxygen documentation.
Along with the manual testing, we strongly recommend you to use the our regression tests and unit tests. If you have implemented an operator, don’t forget to create unit tests for it. If you implement anything else, it is encouraged that you develop complimentary regression tests for it.
The regression test and unit test utilities are native and do not demand any external tool or script, although you need to fetch the test cases from other repositories, as they are shared with other versions of ModSecurity, those others repositories git submodules. To fetch the submodules repository and run the utilities, follow the commands listed below:
$ cd /path/to/your/ModSecurity
$ git submodule foreach git pull
$ cd test
$ ./regression-tests
$ ./unit-testsBefore start the debugging process, make sure of where your bug is. The problem could be on your connector or in libmodsecurity. In order to identify where the bug is, it is recommended that you develop a regression test that mimics the scenario where the bug is happening. If the bug is reproducible with the regression-test utility, then it will be far simpler to debug and ensure that it never occurs again. On Linux it is recommended that anyone undertaking debugging utilize gdb and/or valgrind as needed.
During the configuration/compilation time, you may want to disable the compiler optimization making your “back traces” populated with readable data. Use the CFLAGS to disable the compilation optimization parameters:
$ export CFLAGS="-g -O0"
$ ./build.sh
$ ./configure --enable-assertions=yes
$ make
$ sudo make install"Assertions allow us to document assumptions and to spot violations early in the development process. What is more, assertions allow us to spot violations with a minimum of effort." https://dl.acm.org/doi/pdf/10.1145/240964.240969
It is recommended to use assertions where applicable, and to enable them with '--enable-assertions=yes' during the testing and debugging workflow.
The source tree includes a Benchmark tool that can help measure library performance. The tool is located in the test/benchmark/ directory. The build process also creates the binary here, so you will have the tool after the compilation is finished.
To run, just type:
cd test/benchmark
$ ./benchmark
Doing 1000000 transactions...
You can also pass a lower value:
$ ./benchmark 1000
Doing 1000 transactions...To measure the time:
$ time ./benchmark 1000
Doing 1000 transactions...
real 0m0.351s
user 0m0.337s
sys 0m0.022sThis is very fast because the benchmark uses the minimal modsecurity.conf.default configuration, which doesn't include too many rules:
$ cat basic_rules.conf
Include "../../modsecurity.conf-recommended"
To measure with real rules, run one of the download scripts in the same directory:
$ ./download-owasp-v3-rules.sh
Cloning into 'owasp-v3'...
remote: Enumerating objects: 33007, done.
remote: Counting objects: 100% (2581/2581), done.
remote: Compressing objects: 100% (907/907), done.
remote: Total 33007 (delta 2151), reused 2004 (delta 1638), pack-reused 30426
Receiving objects: 100% (33007/33007), 9.02 MiB | 16.21 MiB/s, done.
Resolving deltas: 100% (25927/25927), done.
Switched to a new branch 'tag3.0.2'
/path/to/ModSecurity/test/benchmark
Done.
$ cat basic_rules.conf
Include "../../modsecurity.conf-recommended"
Include "owasp-v3/crs-setup.conf.example"
Include "owasp-v3/rules/*.conf"Now the command will give much higher value.
The tool is a straightforward wrapper application that utilizes the library. It creates a ModSecurity instance and a RuleSet instance, then runs a loop based on the specified number. Within this loop, it creates a Transaction object to emulate real HTTP transactions.
Each transaction is an HTTP/1.1 GET request with some GET parameters. Common headers are added, followed by the response headers and an XML body. Between phases, the tool checks whether an intervention has occurred. All transactions are created with the same data.
Note that the tool does not call the last phase (logging).
Please remember to reset basic_rules.conf if you want to try with a different ruleset.
If you are facing a configuration issue or something is not working as you expected to be, please use the ModSecurity user’s mailing list. Issues on GitHub are also welcomed, but we prefer to have user ask questions on the mailing list first so that you can reach an entire community. Also don’t forget to look for existing issues before open a new one.
If you are going to open a new issue on GitHub, don’t forget to tell us the version of your libmodsecurity and the version of a specific connector if there is one.
Please do not make public any security issue. Contact us at: modsecurity@owasp.org reporting the issue. Once the problem is fixed your credit will be given.
We are open to discussing any new feature request with the community via the mailing lists. You can alternativly, feel free to open GitHub issues requesting new features. Before opening a new issue, please check if there is one already opened on the same topic.
The libModSecurity design allows the integration with bindings. There is an effort to avoid breaking API [binary] compatibility to make an easy integration with possible bindings. Currently, there are a few notable projects maintained by the community:
- Python - https://github.com/actions-security/pymodsecurity
- Rust - https://github.com/rkrishn7/rust-modsecurity
- Varnish - https://github.com/xdecock/vmod-modsecurity
Having our packages in distros on time is a desire that we have, so let us know if there is anything we can do to facilitate your work as a packager.
Development of ModSecurity is sponsored by Trustwave. Sponsorship will end July 1, 2024. Additional information can be found here https://www.trustwave.com/en-us/resources/security-resources/software-updates/end-of-sale-and-trustwave-support-for-modsecurity-web-application-firewall/