Canonicalization, locale and Unicode

Development Guide Table of Contents

Objective
To ensure the application is robust when subjected to encoded, internationalized and Unicode input.

Platforms Affected
All.

Relevant COBIT Topics
DS11.9 – Data processing integrity

Description
Applications are rarely tested for Unicode exploits, and yet many are vulnerable due to the same sort of issues which allows HTTP Request Smuggling to work – every browser, web server, web application firewall or HTTP inspection agent, and other device treats user locale handling in different (and usually confusing) manner.

Canonicalization deals with the way in which systems convert data from one form to another. Canonical means the simplest or most standard form of something. Canonicalization is the process of converting something from one representation to the simplest form.

Web applications have to deal with lots of canonicalization issues from URL encoding to IP address translation. When security decisions are made based on less than perfectly canonicalized data, the application itself must be able to deal with unexpected input safely.

'''NB: To be secure against canonicalization attacks does not mean that every application has to be internationalized, but all applications should be safe when Unicode and malformed representations are entered. '''

Unicode
Unicode Encoding is a method for storing characters with multiple bytes. Wherever input data is allowed, data can be entered using Unicode to disguise malicious code and permit a variety of attacks. RFC 2279 references many ways that text can be encoded.

Unicode was developed to allow a Universal Character Set (UCS) that encompasses most of the world's writing systems. Multi-octet characters, however, are not compatible with many current applications and protocols, and this has led to the development of a few UCS transformation formats (UTF) with varying characteristics. UTF-8 has the characteristic of preserving the full US-ASCII range. It is compatible with file systems, parsers and other software relying on US-ASCII values, but it is transparent to other values.

The importance of UTF-8 representation stems from the fact that web-servers/applications perform several steps on their input of this format. The order of the steps is sometimes critical to the security of the application. Basically, the steps are "URL decoding" potentially followed by "UTF-8 decoding", and intermingled with them are various security checks, which are also processing steps.

If, for example, one of the security checks is searching for "..", and it is carried out before UTF-8 decoding takes place, it is possible to inject ".." in their overlong UTF-8 format. Even if the security checks recognize some of the non-canonical format for dots, it may still be that not all formats are known to it.

Consider the ASCII character "." (dot). Its canonical representation is a dot (ASCII 2E). Yet if we think of it as a character in the second UTF-8 range (2 bytes), we get an overlong representation of it, as C0 AE. Likewise, there are more overlong representations: E0 80 AE, F0 80 80 AE, F8 80 80 80 AE and FC 80 80 80 80 AE.

Consider the representation C0 AE of a ".". Like UTF-8 encoding requires, the second octet has "10" as its two most significant bits. Now, it is possible to define 3 variants for it, by enumerating the rest of the possible 2 bit combinations ("00", "01" and "11"). Some UTF-8 decoders would treat these variants as identical to the original symbol (they simply use the least significant 6 bits, disregarding the most significant 2 bits). Thus, the 3 variants are C0 2E, C0 5E and C0 FE.

It is thus possible to form illegal UTF-8 encodings, in two senses:


 * A UTF-8 sequence for a given symbol may be longer than necessary for representing the symbol.


 * A UTF-8 sequence may contain octets that are in incorrect format (i.e. do not comply with the above 6 formats).

To further "complicate" things, each representation can be sent over HTTP in several ways:

In the raw. That is, without URL encoding at all. This usually results in sending non-ASCII octets in the path, query or body, which violates the HTTP standards. Nevertheless, most HTTP servers do get along just fine with non-ASCII characters.

Valid URL encoding. Each non-ASCII character (more precisely, all characters that require URL encoding - a superset of non ASCII characters) is URL-encoded. This results in sending, say, %C0%AE.

Invalid URL encoding. This is a variant of valid URL encoding, wherein some hexadecimal digits are replaced with non-hexadecimal digits, yet the result is still interpreted as identical to the original, under some decoding algorithms. For example, %C0 is interpreted as character number ('C'-'A'+10)*16+('0'-'0') = 192. Applying the same algorithm to %M0 yields ('M'-'A'+10)*16+('0'-'0') = 448, which, when forced into a single byte, yields (8 least significant bits) 192, just like the original. So, if the algorithm is willing to accept non-hexadecimal digits (such as 'M'), then it is possible to have variants for %C0 such as %M0 and %BG.

It should be kept in mind that these techniques are not directly related to Unicode, and they can be used in non-Unicode attacks as well.

http://www.example.com/cgi-bin/bad.cgi?foo=../../bin/ls%20-al

URL Encoding of the example attack:

http://www.example.com/cgi-bin/bad.cgi?foo=..%2F../bin/ls%20-al

Unicode encoding of the example attack:

http://www.example.com/cgi-bin/bad.cgi?foo=..%c0%af../bin/ls%20-al

http://www.example.com/cgi-bin/bad.cgi?foo=..%c1%9c../bin/ls%20-al

http://www.example.com/cgi-bin/bad.cgi?foo=..%c1%pc../bin/ls%20-al

http://www.example.com/cgi-bin/bad.cgi?foo=..%c0%9v../bin/ls%20-al

http://www.example.com/cgi-bin/bad.cgi?foo=..%c0%qf../bin/ls%20-al

http://www.example.com/cgi-bin/bad.cgi?foo=..%c1%8s../bin/ls%20-al

http://www.example.com/cgi-bin/bad.cgi?foo=..%c1%1c../bin/ls%20-al

http://www.example.com/cgi-bin/bad.cgi?foo=..%c1%9c../bin/ls%20-al

http://www.example.com/cgi-bin/bad.cgi?foo=..%c1%af../bin/ls%20-al

http://www.example.com/cgi-bin/bad.cgi?foo=..%e0%80%af../bin/ls%20-al

http://www.example.com/cgi-bin/bad.cgi?foo=..%f0%80%80%af../bin/ls%20-al

http://www.example.com/cgi-bin/bad.cgi?foo=..%f8%80%80%80%af../bin/ls%20-al

How to protect yourself
A suitable canonical form should be chosen and all user input canonicalized into that form before any authorization decisions are performed. Security checks should be carried out after UTF-8 decoding is completed. Moreover, it is recommended to check that the UTF-8 encoding is a valid canonical encoding for the symbol it represents.

http://www.ietf.org/rfc/rfc2279.txt?number=2279

Input Formats
Web applications usually operate internally as one of ASCII, ISO 8859-1, or Unicode (Java programs are UTF-16 example). Your users may be using another locale, and attackers can choose their locale and character set with impunity.

How to determine if you are vulnerable
Investigate the web application to determine if it asserts an internal code page, locale, or culture.

If the default character set, locale is not asserted it will be one of the following:


 * HTTP Posts. Interesting tidbit: All HTTP posts are required to be ISO 8859-1, which will lose data for most double byte character sets. You must test your application with your supported browsers to determine if they pass in fully encoded double byte characters safely


 * HTTP Gets. Depends on the previously rendered page and per-browser implementations, but URL encoding is not properly defined for double byte character sets. IE can be optionally forced to do all submits as UTF-8 which is then properly canonicalized on the server


 * .NET: Unicode (little endian)


 * JSP implementations, such as Tomcat: UTF8 - see “javaEncoding” in web.xml by many servlet containers


 * Java: Unicode (UTF-16, big endian, or depends on the OS during JVM startup)


 * PHP: Set in php.ini, ISO 8859-1.

'''NB: Many PHP functions make (invalid) assumptions as to character set and may not work properly when changed to another character set. Test your application with the new character set thoroughly!'''

How to protect yourself

 * Determine your application’s needs, and set both the asserted language locale and character set appropriately.

Locale assertion
The web server should always assert a locale and preferably a country code, such as “en_US”, “fr_FR”, “zh_CN”

How to determine if you are vulnerable
Use a HTTP header sniffer or even just telnet against your web server:

HEAD / HTTP1.0

Should display something like this:

HTTP/1.1 200 OK

Date: Sun, 24 Jul 2005 08:13:17 GMT

Server: Apache/1.3.29

Connection: close

Content-Type: text/html;' charset=iso-8859-1'

How to protect yourself
Review and implement these guidelines:

http://www.w3.org/International/technique-index

At a minimum, select the correct output locale and character set.

Double (or n-) encoding
Most web applications only check once to determine if the input is has been de-encoded into the correct Unicode values. However, an attacker may have doubly encoded the attack string.

How to determine if you are vulnerable

 * Use XSS Cheat Sheet double encoder utility to double encode a XSS string

http://ha.ckers.org/xss.html


 * If the resultant injection is a successful XSS output, then your application is vulnerable


 * This attack may also work against:


 * 1) Filenames
 * 2) Non-obvious items like report types, and language selectors
 * 3) Theme names

How to protect yourself

 * Assert the correct locale and character set for your application


 * Use HTML entities, URL encoding, and so on to prevent Unicode characters being treated improperly by the many divergent browser, server, and application combinations


 * Test your code and overall solution extensively

HTTP Request Smuggling
HTTP Request Smuggling (HRS) is an issue detailed in depth by Klein, Linhart, Heled, and Orrin in a whitepaper found in the references section. The basics of HTTP Request Smuggling is that many larger solutions use many components to provide a web application. The differences between the firewall, web application firewall, load balancers, SSL accelerators, reverse proxies, and web servers allow a specially crafted attack to bypass all the controls in the front-end systems and directly attack the web server.

The types of attack they describe are:


 * Web cache poisoning


 * Firewall/IDS/IPS evasion


 * Forward and backward HRS techniques


 * Request hijacking


 * Request credential hijacking

Since the whitepaper, several examples of real life HRS have been discovered.

How to determine if you are vulnerable

 * Review the whitepaper


 * Review your infrastructure for vulnerable components

How to protect yourself

 * Minimize the total number of components that may interpret the inbound HTTP request


 * Keep your infrastructure up to date with patches