(RHSA-2013:0196) Important: JBoss Enterprise Web Platform 5.2.0 update

An attack technique against the W3C XML Encryption Standard when block
ciphers were used in CBC mode could allow a remote attacker to conduct
chosen-ciphertext attacks, leading to the recovery of the entire plain text
of a particular cryptogram. (CVE-2011-1096)

JBoss Web Services leaked side-channel data when distributing symmetric
keys (for XML encryption), allowing a remote attacker to recover the entire
plain text form of a symmetric key. (CVE-2011-2487)

Spring framework could possibly evaluate Expression Language (EL)
expressions twice, allowing a remote attacker to execute arbitrary code in
the context of the application server, or to obtain sensitive information
from the server. Manual action is required to apply this fix. Refer to the
Solution section. (CVE-2011-2730)

Apache CXF checked to ensure XML elements were signed or encrypted by a
Supporting Token, but not whether the correct token was used. A remote
attacker could transmit confidential information without the appropriate
security, and potentially circumvent access controls on web services
exposed via Apache CXF. Refer to the Solution section for details.
(CVE-2012-2379)

When an application used FORM authentication, along with another component
that calls request.setUserPrincipal() before the call to
FormAuthenticator#authenticate() (such as the Single-Sign-On valve), it was
possible to bypass the security constraint checks in the FORM authenticator
by appending “/j_security_check” to the end of a URL. (CVE-2012-3546)

The JMX Console was vulnerable to CSRF attacks, allowing a remote attacker
to hijack the authenticated JMX Console session of an administrator.
(CVE-2011-2908)

An XSS flaw allowed a remote attacker to perform an XSS attack against
victims using the JMX Console. (CVE-2011-4575)

SecurityAssociation.getCredential() returned the previous credential if
no security context was provided. Depending on the deployed applications,
this could possibly allow a remote attacker to hijack the credentials of a
previously-authenticated user. (CVE-2012-3370)

Configuring the JMX Invoker to restrict access to users with specific
roles did not actually restrict access, allowing remote attackers with
valid JMX Invoker credentials to perform JMX operations accessible to
roles they are not a member of. (CVE-2012-5478)

twiddle.sh accepted credentials as command line arguments, allowing local
users to view them via a process listing. (CVE-2009-5066)

NonManagedConnectionFactory logged the username and password in plain text
when an exception was thrown. This could lead to the exposure of
authentication credentials if local users had permissions to read the log
file. (CVE-2012-0034)

The JMXInvokerHAServlet and EJBInvokerHAServlet invoker servlets allow
unauthenticated access by default in some profiles. The security
interceptor’s second layer of authentication prevented direct exploitation
of this flaw. If the interceptor was misconfigured or inadvertently
disabled, this flaw could lead to arbitrary code execution in the context
of the user running the JBoss server. (CVE-2012-0874)

The JGroups diagnostics service was enabled with no authentication when a
JGroups channel was started, allowing attackers on the adjacent network to
read diagnostic information. (CVE-2012-2377)

CallerIdentityLoginModule retained the password from the previous call if a
null password was provided. In non-default configurations this could
possibly lead to a remote attacker hijacking a previously-authenticated
user’s session. (CVE-2012-3369)

Red Hat would like to thank Juraj Somorovsky of Ruhr-University Bochum for
reporting CVE-2011-1096 and CVE-2011-2487; the Apache CXF project for
reporting CVE-2012-2379; and Tyler Krpata for reporting CVE-2011-4575.
CVE-2012-3370 and CVE-2012-3369 were discovered by Carlo de Wolf of Red
Hat; CVE-2012-5478 discovered by Derek Horton of Red Hat; CVE-2012-0874
discovered by David Jorm of Red Hat; and CVE-2012-2377 was discovered by
Red Hat.