Codereview-Input Validation

Revision as of 09:07, 12 January 2009 by KirstenS (talk | contribs) (Introduction)

Jump to: navigation, search
OWASP Code Review Guide Table of Contents


Input validation is one of the most effective application security technical controls. It can mitigate numerous vulnerabilities (but not all). Input validation is more than checking form field values. The chapter on transactional analysis talks about this.

Data Validation

All external input to the system shall undergo input validation. The validation rules are defined by the business requirements for the application. If possible, an exact match validator should be implemented. Exact match only permits data that conforms to an expected value. A "Known good" approach (white-list) is a little weaker but more flexible are common. Known good only permits characters/ASCII ranges defined within a white-list. Such a range is defined by the business requirements of the input field. The other approaches to data validation are "known bad" which is a black list of "bad characters" - not future proof and would need maintenance. "Encode bad" would be very weakm as it would simply encode characters considered "bad" to a format which is deemed not to affect the functionality of the application.

Business Validation

Business validation is concerned with business logic. An understanding of the business logic is required prior to reviewing the code which performs such logic. Business validation could be used to limit the value range or a transaction inputted by a user or reject input which does not make too much business sense. Reviewing code for business validation can also include rounding errors or floating point issues which may give rise to issues such as integer overflows which can dramatically damage the bottom line.


Canonicalization is the process by which various equivalent forms of a name can be resolved to a single standard name, or the "canonical" name.

The most popular encoding are UTF-8, UTF-16, and so on (which are described in detail in RFC 2279) A single character, such as a period/full-stop (.), may be represented in many different ways, such as ASCII 2E, Unicode C0 AE and many others.

The problem is, with all of these different ways of encoding user input, a Web application's filters can be easily confused if they're not carefully built.

Bad Example

public static void main(String[] args) {
    File x = new File("/cmd/" + args[1]);
    String absPath = x.getAbsolutePath();

Good Example

public static void main(String[] args) throws IOException {
    File x = new File("/cmd/" + args[1]);
    String canonicalPath = x.getCanonicalPath();


Reviewing code for Data Validation