haproxy/contrib/modsecurity
Willy Tarreau aeed4a85d6 REORG: include: move log.h to haproxy/log{,-t}.h
The current state of the logging is a real mess. The main problem is
that almost all files include log.h just in order to have access to
the alert/warning functions like ha_alert() etc, and don't care about
logs. But log.h also deals with real logging as well as log-format and
depends on stream.h and various other things. As such it forces a few
heavy files like stream.h to be loaded early and to hide missing
dependencies depending where it's loaded. Among the missing ones is
syslog.h which was often automatically included resulting in no less
than 3 users missing it.

Among 76 users, only 5 could be removed, and probably 70 don't need the
full set of dependencies.

A good approach would consist in splitting that file in 3 parts:
  - one for error output ("errors" ?).
  - one for log_format processing
  - and one for actual logging.
2020-06-11 10:18:58 +02:00
..
Makefile REORG: ebtree: clean up remains of the ebtree/ directory 2020-06-11 09:31:11 +02:00
modsec_wrapper.c REORG: include: move log.h to haproxy/log{,-t}.h 2020-06-11 10:18:58 +02:00
modsec_wrapper.h
README
spoa.c REORG: include: move common/chunk.h to haproxy/chunk.h 2020-06-11 10:18:57 +02:00
spoa.h

ModSecurity for HAProxy
-----------------------

This is a third party daemon which speaks SPOE. It gives requests send by HAProxy
to ModSecurity and returns the verdict.

  Compilation
---------------

You must compile ModSecurity in standalone mode. Below an example for
ModSecurity-2.9.1. Note that ModSecurity depends the Apache APR. I assume that
the Apache dependencies are installed on the system.

   ./configure \
      --prefix=$PWD/INSTALL \
		--disable-apache2-module \
      --enable-standalone-module \
      --enable-pcre-study \
      --without-lua \
      --enable-pcre-jit
   make
	make -C standalone install
	mkdir -p $PWD/INSTALL/include
	cp standalone/*.h $PWD/INSTALL/include
	cp apache2/*.h $PWD/INSTALL/include

Note that this compilation method works, but is a little bit rustic. I can't
deal with Lua, I supposed that is a dependencies problem on my computer.

  Start the service
---------------------

After you have compiled it, to start the service, you just need to use "spoa"
binary:

    $> ./modsecurity  -h
    Usage: ./spoa [-h] [-d] [-p <port>] [-n <num-workers>] [-f <config-file>]
        -h                  Print this message
        -d                  Enable the debug mode
        -f <config-file>    Modsecurity configuration file
        -m <max-frame-size> Specify the maximum frame size (default : 16384)
        -p <port>           Specify the port to listen on (default: 12345)
        -n <num-workers>    Specify the number of workers (default: 5)
        -c <capability>     Enable the support of the specified capability
        -t <time>           Set a delay to process a message (default: 0)
                            The value is specified in milliseconds by default,
                            but can be in any other unit if the number is suffixed
                            by a unit (us, ms, s)

Note: A worker is a thread.


  Configure a SPOE to use the service
---------------------------------------

All information about SPOE configuration can be found in "doc/SPOE.txt". Here is
the configuration template to use for your SPOE with ModSecurity module:

   [modsecurity]

   spoe-agent modsecurity-agent
      messages check-request
      option var-prefix modsec
      timeout hello      100ms
      timeout idle       30s
      timeout processing 15ms
      use-backend spoe-modsecurity

   spoe-message check-request
      args unique-id method path query req.ver req.hdrs_bin req.body_size req.body
      event on-frontend-http-request

The engine is in the scope "modsecurity". So to enable it, you must set the
following line in a frontend/listener section:

   frontend my-front
      ...
      filter spoe engine modsecurity config spoe-modsecurity.conf
      ...


Because, in SPOE configuration file, we declare to use the backend
"spoe-modsecurity" to communicate with the service, you must define it in
HAProxy configuration. For example:

   backend spoe-modsecurity
      mode tcp
      balance roundrobin
      timeout connect 5s
      timeout server  3m
      server modsec1 127.0.0.1:12345

The modsecurity action is returned in a variable called txn.modsec.code. It
contains the HTTP returned code. If the variable contains 0, the request is
clean.

   http-request deny if { var(txn.modsec.code) -m int gt 0 }

With this rule, all the request not clean are rejected.


  Known bugs, limitations and TODO list
-----------------------------------------

Modsecurity bugs:
-----------------

* When the audit_log is used with the directive "SecAuditLogType Serial", in
  some systems, the APR mutex initialisation silently fails, this causes a
  segmentation fault. For my own usage, I have a patched version of modsec where
  I use another mutex than "APR_LOCK_DEFAULT" like "APR_LOCK_PROC_PTHREAD"

   -    rc = apr_global_mutex_create(&msce->auditlog_lock, NULL, APR_LOCK_DEFAULT, mp);
   +    rc = apr_global_mutex_create(&msce->auditlog_lock, NULL, APR_LOCK_PROC_PTHREAD, mp);

* Configuration file loaded with wildcard (eg. Include rules/*.conf), are loaded
  in reverse alphabetical order. You can found a patch below. The ModSecurity
  team ignored this patch.

  https://github.com/SpiderLabs/ModSecurity/issues/1285
  http://www.arpalert.org/0001-Fix-bug-when-load-files.patch

  Or insert includes without wildcards.

Todo:
-----

* Clarify the partial body analysis.
* The response body is not yet analyzed.
* ModSecurity can't modify the response body.
* Implements real log management. Actually, the log are sent on stderr.
* Implements daemon things (forks, write a pid, etc.).