You are viewing an old version of this page. View the current version.

Compare with Current View Page History

Version 1 Next »

These checkers enforce the CERT C Secure Coding rules.

This code has been developed and tested on an i386 workstation running Linux (2.6.16.60) and g++ (3.4.4)

This code depends on ROSE 0.9.3a, which is available for free download from:

http://rosecompiler.org/

ROSE 0.9.3a also depends on the BOOST C++ library, version 1.3.5, which is available for free download from:

http://www.boost.org/

First make sure that the ROSE environment variable points to the build directory of ROSE:

export ROSE=/usr/local/rose/compileTree

Building Diagnose
-----------------

To build the ROSE 'diagnose' program, which runs secure coding rules:

make pgms

To test diagnose on the code samples from the CERT C Secure Coding Rules:

make tests

To build API documentation pages, you must have doxygen installed:

make doc

To clean documentation pages and build files:

make clean

Running Diagnose
----------------

To run the diagnose program on a C file, simply pass the C file as an argument:

diagnose hello.c

If the C file violates some secure coding rules, the diagnose program will print them out. If the diagnose program can not find any violations, it prints nothing.

Secure Coding Rules Enforced by Diagnose
----------------------------------------

The C Secure Coding Rules are available at:

https://www.securecoding.cert.org/confluence/display/seccode/CERT+C+Secure+Coding+Standard

Here is a breakdown of how thoroughly diagnose enforces the C Secure Coding Rules:

Complete 57

ROSE catches all violations of these rules

Partial 45

ROSE catches some, but not all violations of these rules

false-positive 9

These rules could be checked by diagnose, but they will also catch some false positives.

Potential 29

These rules are not checked by diagnose, but could be

Undoable 32

These rules could not be checked by ROSE due to various limitations in ROSE.

Unenforceable 48

These rules could not be checked by any tool that relies purely on unaided static analysis.

TOTAL 220

  • No labels