Vulnerabilities > CVE-2023-41337 - Improper Verification of Cryptographic Signature vulnerability in Dena H2O
Summary
h2o is an HTTP server with support for HTTP/1.x, HTTP/2 and HTTP/3. In version 2.3.0-beta2 and prior, when h2o is configured to listen to multiple addresses or ports with each of them using different backend servers managed by multiple entities, a malicious backend entity that also has the opportunity to observe or inject packets exchanged between the client and h2o may misdirect HTTPS requests going to other backends and observe the contents of that HTTPS request being sent. The attack involves a victim client trying to resume a TLS connection and an attacker redirecting the packets to a different address or port than that intended by the client. The attacker must already have been configured by the administrator of h2o to act as a backend to one of the addresses or ports that the h2o instance listens to. Session IDs and tickets generated by h2o are not bound to information specific to the server address, port, or the X.509 certificate, and therefore it is possible for an attacker to force the victim connection to wrongfully resume against a different server address or port on which the same h2o instance is listening. Once a TLS session is misdirected to resume to a server address / port that is configured to use an attacker-controlled server as the backend, depending on the configuration, HTTPS requests from the victim client may be forwarded to the attacker's server. An H2O instance is vulnerable to this attack only if the instance is configured to listen to different addresses or ports using the listen directive at the host level and the instance is configured to connect to backend servers managed by multiple entities. A patch is available at commit 35760540337a47e5150da0f4a66a609fad2ef0ab. As a workaround, one may stop using using host-level listen directives in favor of global-level ones.
Vulnerable Configurations
Common Weakness Enumeration (CWE)
Common Attack Pattern Enumeration and Classification (CAPEC)
- Padding Oracle Crypto Attack An attacker is able to efficiently decrypt data without knowing the decryption key if a target system leaks data on whether or not a padding error happened while decrypting the ciphertext. A target system that leaks this type of information becomes the padding oracle and an attacker is able to make use of that oracle to efficiently decrypt data without knowing the decryption key by issuing on average 128*b calls to the padding oracle (where b is the number of bytes in the ciphertext block). In addition to performing decryption, an attacker is also able to produce valid ciphertexts (i.e., perform encryption) by using the padding oracle, all without knowing the encryption key. Any cryptosystem can be vulnerable to padding oracle attacks if the encrypted messages are not authenticated to ensure their validity prior to decryption, and then the information about padding error is leaked to the attacker. This attack technique may be used, for instance, to break CAPTCHA systems or decrypt/modify state information stored in client side objects (e.g., hidden fields or cookies). This attack technique is a side-channel attack on the cryptosystem that uses a data leak from an improperly implemented decryption routine to completely subvert the cryptosystem. The one bit of information that tells the attacker whether a padding error during decryption has occurred, in whatever form it comes, is sufficient for the attacker to break the cryptosystem. That bit of information can come in a form of an explicit error message about a padding error, a returned blank page, or even the server taking longer to respond (a timing attack). This attack can be launched cross domain where an attacker is able to use cross-domain information leaks to get the bits of information from the padding oracle from a target system / service with which the victim is communicating. To do so an attacker sends a request containing ciphertext to the target system. Due to the browser's same origin policy, the attacker is not able to see the response directly, but can use cross-domain information leak techniques to still get the information needed (i.e., information on whether or not a padding error has occurred). For instance, this can be done using "img" tag plus the onerror()/onload() events. The attacker's JavaScript can make web browsers to load an image on the target site, and know if the image is loaded or not. This is 1-bit information needed for the padding oracle attack to work: if the image is loaded, then it is valid padding, otherwise it is not.