Summary pgjdbc is vulnerable to a client-side denial of service during SCRAM-SHA-256 authentication.

Impact A malicious server can instruct the driver to perform SCRAM authentication with a very large iteration count. With a large enough value, the client spends an unbounded amount of CPU time inside PBKDF2 before authentication can fail. A single attempt ties up a CPU core. Repeated or concurrent attempts exhaust client CPU and can wedge connection pools.

In affected versions, loginTimeout did not fully mitigate this problem. When loginTimeout expired, the caller could stop waiting, but the worker thread performing the connection attempt could continue running and burning CPU inside the SCRAM PBKDF2 computation.

This issue affects availability. It does not provide authentication bypass, privilege escalation, or direct password disclosure.

A user is vulnerable when all of the following are true:

1. The connection uses SCRAM-SHA-256 authentication. 2. The client reaches a malicious, compromised, or attacker-controlled PostgreSQL endpoint. 3. That endpoint sends a very large SCRAM PBKDF2 iteration count in the server-first-message.

In practice, that can happen in these situations:

- the application lets end users or tenants supply their own database connection details (as in many BI, reporting, analytics, ETL, and low-code platforms), so a user can point the shared client host at a server they control - the application accepts connection strings, hostnames, or JDBC URLs from user input, configuration uploaded by users, or other untrusted sources - the application is configured to connect to a PostgreSQL server that is itself malicious or later becomes compromised - the application connects through an untrusted proxy, relay, tunnel, bastion, or connection-pooling service that can act as the PostgreSQL server - an attacker can redirect the client to a fake PostgreSQL endpoint by manipulating DNS, service discovery, Kubernetes service resolution, /etc/hosts, environment variables, or similar indirection - an active network attacker on the path can impersonate the server because the connection does not strongly verify server identity (for example, sslmode lower than verify-full, or trusting a CA that signs hosts outside the operator's control)

The issue is more damaging when the application uses connection retries, many parallel connection attempts, or loginTimeout and assumes the timeout fully stops the work.

Patches The patch introduces a new connection property, scramMaxIterations, with a default of 100K. The client now rejects SCRAM server messages that advertise more PBKDF2 iterations than the configured cap before starting the PBKDF2 computation begins.

Workarounds

Until a patched version of pgjdbc is deployed, the following measures reduce exposure:

1. Only connect to trusted PostgreSQL servers whose identity is verified. Connect only to trusted PostgreSQL servers, and verify server identity with TLS using sslmode=verify-full and a trusted CA. TLS without certificate and hostname verification is not sufficient as an active network attacker can still impersonate the server.

2. Do not rely on loginTimeout as a complete mitigation on unpatched versions. On affected versions, loginTimeout can stop the waiting caller while the worker thread continues spending CPU.

3. Avoid SCRAM on untrusted or interceptable connection paths. For those paths, use an authentication method that does not let the server choose a SCRAM PBKDF2 iteration count.

4. Reduce blast radius operationally. Limit parallel connection attempts, add retry backoff, isolate connection establishment in a separate worker or process when possible, and apply CPU or container limits where appropriate.

5. On trusted servers you control, keep SCRAM iteration counts at ordinary values. This does not defend against an attacker-controlled server, but it avoids unnecessary client cost when talking to legitimate servers.

Vulnerability

PreparedStatement.setText(int, InputStream) and

PreparedStatemet.setBytea(int, InputStream)

will create a temporary file if the InputStream is larger than 51k

Example of vulnerable code:

java String s = "some very large string greater than 51200 bytes";

PreparedStatement.setInputStream(1, new ByteArrayInputStream(s.getBytes()) ); This will create a temporary file which is readable by other users on Unix like systems, but not MacOS.

Impact On Unix like systems, the system's temporary directory is shared between all users on that system. Because of this, when files and directories are written into this directory they are, by default, readable by other users on that same system.

This vulnerability does not allow other users to overwrite the contents of these directories or files. This is purely an information disclosure vulnerability.

When analyzing the impact of this vulnerability, here are the important questions to ask:

Is the driver running in an environment where the OS has other untrusted users. If yes, and you answered 'yes' to question 1, this vulnerability impacts you. If no, this vulnerability does not impact you. Patches Because certain JDK file system APIs were only added in JDK 1.7, this this fix is dependent upon the version of the JDK you are using.

Java 1.8 and higher users: this vulnerability is fixed in 42.2.27, 42.3.8, 42.4.3, 42.5.1 Java 1.7 users: this vulnerability is fixed in 42.2.27.jre7 Java 1.6 and lower users: no patch is available; you must use the workaround below. Workarounds If you are unable to patch, or are stuck running on Java 1.6, specifying the java.io.tmpdir system environment variable to a directory that is exclusively owned by the executing user will fix this vulnerability.

References CWE-200: Exposure of Sensitive Information to an Unauthorized Actor Fix commit https://github.com/pgjdbc/pgjdbc/commit/9008dc9aade6dbfe4efafcd6872ebc55f4699cf5 Similar Vulnerabilities Google Guava - https://github.com/google/guava/issues/4011 Apache Ant - https://nvd.nist.gov/vuln/detail/CVE-2020-1945 JetBrains Kotlin Compiler - https://nvd.nist.gov/vuln/detail/CVE-2020-15824

Impact What kind of vulnerability is it? Who is impacted?

The PGJDBC implementation of the java.sql.ResultRow.refreshRow() method is not performing escaping of column names so a malicious column name that contains a statement terminator, e.g. ;, could lead to SQL injection. This could lead to executing additional SQL commands as the application's JDBC user.

User applications that do not invoke the ResultSet.refreshRow() method are not impacted.

User application that do invoke that method are impacted if the underlying database that they are querying via their JDBC application may be under the control of an attacker. The attack requires the attacker to trick the user into executing SQL against a table name who's column names would contain the malicious SQL and subsequently invoke the refreshRow() method on the ResultSet.

For example:

sql CREATE TABLE refreshrowexample ( id int PRIMARY KEY, "1 FROM refreshrowexample; SELECT pgsleep(10); SELECT " int );

This example has a table with two columns. The name of the second column is crafted to contain a statement terminator followed by additional SQL. Invoking the ResultSet.refreshRow() on a ResultSet that queried this table, e.g. SELECT FROM refreshrow, would cause the additional SQL commands such as the SELECT pgsleep(10) invocation to be executed.

As the multi statement command would contain multiple results, it would not be possible for the attacker to get data directly out of this approach as the ResultSet.refreshRow() method would throw an exception. However, the attacker could execute any arbitrary SQL including inserting the data into another table that could then be read or any other DML / DDL statement.

Note that the application's JDBC user and the schema owner need not be the same. A JDBC application that executes as a privileged user querying database schemas owned by potentially malicious less-privileged users would be vulnerable. In that situation it may be possible for the malicious user to craft a schema that causes the application to execute commands as the privileged user.

Patches Has the problem been patched? What versions should users upgrade to?

Yes, versions 42.2.26, 42.3.7, and 42.4.1 have been released with a fix.

Workarounds Is there a way for users to fix or remediate the vulnerability without upgrading?

Check that you are not using the ResultSet.refreshRow() method.

If you are, ensure that the code that executes that method does not connect to a database that is controlled by an unauthenticated or malicious user. If your application only connects to its own database with a fixed schema with no DDL permissions, then you will not be affected by this vulnerability as it requires a maliciously crafted schema.

DISPUTED In pgjdbc before 42.3.3, an attacker (who controls the jdbc URL or properties) can call java.util.logging.FileHandler to write to arbitrary files through the loggerFile and loggerLevel connection properties. An example situation is that an attacker could create an executable JSP file under a Tomcat web root. NOTE: the vendor's position is that there is no pgjdbc vulnerability; instead, it is a vulnerability for any application to use the pgjdbc driver with untrusted connection properties.

pgjdbc is the offical PostgreSQL JDBC Driver. A security hole was found in the jdbc driver for postgresql database while doing security research. The system using the postgresql library will be attacked when attacker control the jdbc url or properties. pgjdbc instantiates plugin instances based on class names provided via authenticationPluginClassName, sslhostnameverifier, socketFactory, sslfactory, sslpasswordcallback connection properties. However, the driver did not verify if the class implements the expected interface before instantiating the class. This can lead to code execution loaded via arbitrary classes. Users using plugins are advised to upgrade. There are no known workarounds for this issue.

Impact SQL injection is possible when using the non-default connection property preferQueryMode=simple in combination with application code that has a vulnerable SQL that negates a parameter value.

There is no vulnerability in the driver when using the default query mode. Users that do not override the query mode are not impacted.

Exploitation

To exploit this behavior the following conditions must be met:

1. A placeholder for a numeric value must be immediately preceded by a minus (i.e. -) 1. There must be a second placeholder for a string value after the first placeholder on the same line. 1. Both parameters must be user controlled.

The prior behavior of the driver when operating in simple query mode would inline the negative value of the first parameter and cause the resulting line to be treated as a -- SQL comment. That would extend to the beginning of the next parameter and cause the quoting of that parameter to be consumed by the comment line. If that string parameter includes a newline, the resulting text would appear unescaped in the resulting SQL.

When operating in the default extended query mode this would not be an issue as the parameter values are sent separately to the server. Only in simple query mode the parameter values are inlined into the executed SQL causing this issue.

Example

java PreparedStatement stmt = conn.prepareStatement("SELECT -?, ?"); stmt.setInt(1, -1); stmt.setString(2, "\nWHERE false --"); ResultSet rs = stmt.executeQuery();

The resulting SQL when operating in simple query mode would be:

sql SELECT --1,' WHERE false --'

The contents of the second parameter get injected into the command. Note how both the number of result columns and the WHERE clause of the command have changed. A more elaborate example could execute arbitrary other SQL commands.

Patch Problem will be patched upgrade to 42.7.2, 42.6.1, 42.5.5, 42.4.4, 42.3.9, 42.2.28, 42.2.28.jre7

The patch fixes the inlining of parameters by forcing them all to be serialized as wrapped literals. The SQL in the prior example would be transformed into:

sql SELECT -('-1'::int4), (' WHERE false --')

Workarounds Do not use the connection propertypreferQueryMode=simple. (NOTE: If you do not explicitly specify a query mode then you are using the default of extended and are not impacted by this issue.)

Impact

channelBinding=require connections can be silently downgraded from SCRAM-SHA-256-PLUS (with channel binding) to plain SCRAM-SHA-256 (without it), losing the man-in-the-middle protection the setting is meant to guarantee. An attacker who can intercept the TLS connection triggers the downgrade with a certificate whose signature algorithm has no tls-server-end-point channel-binding hash. Examples are Ed25519, Ed448, and post-quantum algorithms.

Two issues combine in releases 42.7.4 through 42.7.11:

1. The bundled com.ongres.scram:scram-client (3.1 or 3.2) returns an empty byte array instead of failing when it cannot derive the binding hash for such a certificate. This is the library issue tracked as GHSA-p9jg-fcr6-3mhf. 2. pgJDBC does not enforce channelBinding=require where it matters. ScramAuthenticator checks only that the server advertised a -PLUS mechanism; it neither rejects the empty binding nor checks that the negotiated mechanism uses channel binding. The connection therefore downgrades silently, and would do so even against a fixed scram-client, because the missing enforcement is in pgJDBC's own code.

Only connections that set channelBinding=require are affected. Under the default prefer policy, and under allow or disable, falling back to plain SCRAM is the documented behaviour. Releases before 42.7.4 are unaffected, because they do not support channel binding.

Patches

Fixed in pgJDBC 42.7.12. pgJDBC now enforces channel binding in its own code, independently of the scram-client version:

- Under channelBinding=require, it fails the connection when no channel-binding data can be extracted from the server certificate, instead of passing an empty value to the SCRAM client. The error names the certificate signature algorithm. - After negotiation, it requires the selected mechanism to use channel binding (a -PLUS mechanism) whenever channelBinding=require is set, regardless of how negotiation resolved.

Upgrade to 42.7.12 or later.

Workarounds

No pgJDBC setting restores channel-binding enforcement on an affected release; upgrading is the fix.

If you cannot upgrade immediately, verify the server certificate at the TLS layer so that a man-in-the-middle cannot present a substitute certificate. Set sslmode=verify-full with a truststore that contains only your server's CA. This defence is independent of channel binding and blocks the same attacker. Connections that rely on channelBinding=require in place of certificate verification have no equivalent workaround and should upgrade.

References

- GHSA-p9jg-fcr6-3mhf — the related com.ongres.scram:scram-client issue (root cause of the empty channel-binding value). - scram-client 3.3 release (library fix): https://github.com/ongres/scram/releases/tag/3.3 - pgJDBC fix in 42.7.12: commit

A flaw was found in PostgreSQL JDBC in versions prior to 42.2.13. An XML External Entity (XXE) weakness was found in PostgreSQL JDBC. The highest threat from this vulnerability is to data confidentiality and system availability.

Impact When the PostgreSQL JDBC driver is configured with channel binding set to required (default value is prefer), the driver would incorrectly allow connections to proceed with authentication methods that do not support channel binding (such as password, MD5, GSS, or SSPI authentication). This could allow a man-in-the-middle attacker to intercept connections that users believed were protected by channel binding requirements.

Patches TBD

Workarounds

Configure sslMode=verify-full to prevent MITM attacks.

References

https://www.postgresql.org/docs/current/sasl-authentication.html#SASL-SCRAM-SHA-256 https://datatracker.ietf.org/doc/html/rfc7677 https://datatracker.ietf.org/doc/html/rfc5802

Interaction error in the PostgreSQL JDBC driver before 8.2, when used with a PostgreSQL server with the "standardconformingstrings" option enabled, such as the default configuration of PostgreSQL 9.1, does not properly escape unspecified JDBC statement parameters, which allows remote attackers to perform SQL injection attacks. NOTE: as of 20120330, it was claimed that the upstream developer planned to dispute this issue, but an official dispute has not been posted as of 20121005.