It was discovered that the AccessController class implementation in the Security component of OpenJDK failed, in certain cases, to consider the current context and correctly restrict privileges based on it. An untrusted Java application or applet could use this flaw to bypass certain Java sandbox restrictions.
It was discovered that the SMTP client implementation in the Networking component of OpenJDK failed to correctly handle sender and recipient addresses containing newline characters. A remote attacker could possibly use this flaw to manipulate an SMTP connection opened by a Java application if it could make it send an email to or from a specially crafted address.
It was discovered that the Security component of OpenJDK did not allow users to restrict the set of algorithms allowed for Jar integrity verification. This flaw could allow an attacker to modify content of the Jar file that used weak signing key or hash algorithm.
This problem was originally addressed as part of October 2016 CPU as CVE-2016-5542 (bug 1385723). In that update, the following changes were made:
- New security property jdk.jar.disabledAlgorithms was introduced, which can be used to restrict which algorithms can be used for jar verification.
- MD2 hash algorithm and RSA keys with less than 1024 bits were disabled by default.
At the same time, it was announced that the MD5 has algorithm was going to be disabled in the future updates. It was originally planned to get disabled as part of the January 2017 CPU, but the change was further postponed to the April 2017 CPU. Hence, MD5 is now becoming disabled by default.
The further details of the planned cryptography changes are available on the "Oracle JRE and JDK Cryptographic Roadmap" page:
https://www.java.com/en/jre-jdk-cryptoroadmap.html
It was discovered that the FTP client implementation in the Networking component of OpenJDK failed to correctly handle user inputs (e.g. usernames and passwords) containing newline characters. A remote attacker could possibly use this flaw to manipulate an FTP connection opened by a Java application if it could make it access a specially crafted FTP URL.
Blog posts were published that describe how to use this flaw to open ports on firewalls using active FTP connections:
http://blog.blindspotsecurity.com/2017/02/advisory-javapython-ftp-injections.html
and sending mails during XML parsing when use of XML external entities is allowed:
https://shiftordie.de/blog/2017/02/18/smtp-over-xxe/
It was discovered that the Security component of OpenJDK did not correctly perform merging of multiple sections for the same file listed in the JAR archive file manifest. An attacker could possibly use this flaw to alter certain attributes specified in the manifest without changing archive signature.
A memory disclosure flaw was found in the FileChannelImpl class in the Libraries component of OpenJDK. An untrusted Java application or applet could use this flaw leak limited amount of Java Virtual Machine memory possibly containing sensitive information, resulting in a partial bypass of Java sandbox restrictions.
An information leak flaw was found in the Networking component of OpenJDK. The HttpURLConnection class implementation could re-send HTTP headers containing sensitive data (such as Cookie or Authorization headers) to a different host when following HTTP redirects. This could lead to exposure of data to unintended HTTP hosts.
It was discovered that the Security component of OpenJDK could incorrectly use unsigned manifest attribute entries when only properly signed entries were meant to be used. This could lead to bypass of protections provided by Jar signing. An untrusted Java application or applet could use this flaw to bypass certain Java sandbox restrictions.