In Java data is stored in big-endian format (also called network order), that is, all data is represented sequentially starting from the most significant bit (MSB) to the least significant. Prior to JDK 1.4, custom methods had to be defined to be compatible with little-endian systems that use the reverse byte order. Handling byte order related issues is critical when data is to be exchanged in a networked environment that consists of machines varying in endianness. Failure to handle such cases can cause misinterpretations and unexpected program behavior.
Noncompliant Code Example
The java.io.DataInputStream
class defines read methods (readByte, readShort, readInt, readLong, readFloat and readDouble
) and the corresponding write methods. All these methods work with big-endian data only. The use of these methods can be unsafe while interfacing with traditional languages such as C or C++, that do not provide any guarantees on endianness. This noncompliant example shows such a discrepancy.
DataInputStream dis = new DataInputStream( new FileInputStream("data.txt")); // Little-endian data might be read as big-endian int serialNumber = dis.readInt();
Compliant Solution (1)
This compliant solution wraps the byte array containing the integer bytes read-in into a ByteBuffer
and sets the byte order to little-endian. The result is stored in the integer serialNumber
.
DataInputStream dis = new DataInputStream( new FileInputStream("data.txt")); byte[] buffer= new byte[4]; int bytesRead = dis.read(buffer); // Bytes are read into buffer int serialNumber = ByteBuffer.wrap(buffer).order(ByteOrder.LITTLE_ENDIAN).getInt();
Compliant Solution (2)
Assuming that an integer value is to be read from the file, read and write methods can be defined for handling little-endian data. The readLittleEndianInteger
method reads data into a byte buffer and then pieces together the integer in the right order. Long
values can be handled by defining a byte buffer of size 8
. The writeLittleEndianInteger
method obtains bytes by repeatedly casting the integer so that the most significant byte is extracted on successive right shifts.
// read method public static int readLittleEndianInteger(InputStream ips) throws IOException { int result; byte[] buffer = new byte[4]; int check = ips.read(buffer); if (check != 4) throw new IOException("Unexpected End of Stream"); result = (buffer[3] << 24) | (buffer[2] << 16) | (buffer[1] << 8) | buffer[0]; return result; } // write method public static void writeLittleEndianInteger(int i, OutputStream ops) throws IOException { int result; byte[] buffer = new byte[4]; buffer[0] = (byte) i; buffer[1] = (byte) (i >> 8); buffer[2] = (byte) (i >> 16); buffer[3] = (byte) (i >> 24); ops.write(buffer); }
Compliant Solution (3)
In JDK 1.5+, the Integer.reverseBytes()
method can be used to reverse the order of the bytes constituting the integer. Note that there is no such method for float
and double
values.
public static int reverse(int i){ return Integer.reverseBytes(i); }
Risk Assessment
Reading and writing data without considering endianness may lead to serious misinterpretations about magnitude and sign, alike.
Rule |
Severity |
Likelihood |
Remediation Cost |
Priority |
Level |
---|---|---|---|---|---|
SEC01-J |
low |
unlikely |
low |
P3 |
L3 |
Automated Detection
TODO
Related Vulnerabilities
Search for vulnerabilities resulting from the violation of this rule on the CERT website.
References
[[API 06]] Class ByteBuffer: Methods wrap
and order
. Class Integer: method reverseBytes
[[Harold 97]] Chapter 2: Primitive Data Types, Cross Platform issues
[[MITRE 09]] CWE ID 198 "Use of Incorrect Byte Ordering"
04. Integers (INT) 04. Integers (INT) INT01-J. Provide mechanisms to handle unsigned data when required