Multiple OpenSSL Vulnerabilities Allow Remote Attackers to Cause DoS and Corrupt Memory
OpenSSL has released patches for seven vulnerabilities, including a moderate severity heap buffer overflow and several memory exhaustion flaws, affecting widely used versions of the cryptographic library.
OpenSSL has issued a fresh security advisory disclosing seven vulnerabilities across its cryptographic library, ranging from a heap-corrupting write bug to memory-exhaustion flaws in its QUIC and DTLS implementations. The flaws affect widely deployed branches, including OpenSSL 4.0, 3.6, 3.5, 3.4, and 3.0, and, in some cases, the legacy 1.1.1 line, making patching a priority for any organization running TLS, CMS, or CMP services built on the library.
The most severe issue, tracked as CVE-2026-63072 and rated Moderate, lives in OpenSSL’s CMS decryption code. When sizing the buffer for an unwrapped key, the library queries the expected output length, but the AES-WRAP-PAD unwrap primitive can actually write and cleanse more bytes than that query reports. An attacker can take a legitimate encrypted message, flip a single byte in its key-wrap algorithm identifier, and force an 8-byte out-of-bounds heap write whenever the victim calls CMS_decrypt(). Because the write is fixed in size and value, exploitation is reliable and requires no special server configuration, though OpenSSL notes the realistic outcome is heap corruption and a denial of service rather than code execution.
A second Moderate-severity flaw, CVE-2026-63076, affects the Certificate Management Protocol (CMP). During password-based MAC verification, OpenSSL checks only that a protection algorithm parameter is non-null without confirming its actual type, allowing a crafted message to trigger an invalid pointer dereference. The bug is remotely triggerable by an unauthenticated attacker against any application running a CMP server that accepts PBM-protected messages, and can equally crash a CMP client connecting to a malicious or intercepted server.
The advisory also details five Low-severity issues. CVE-2026-14457 can cause a null pointer dereference in TLS servers or clients configured with Raw Public Keys but no accompanying certificate, when a peer sends the signature_algorithms_cert extension. CVE-2026-54874 lets an attacker amplify network traffic by roughly 1,200 times, tricking a DTLS endpoint into buffering large read buffers for small forged records tied to a future handshake epoch, straining memory on busy servers.
CVE-2026-63073 stems from CMP response validation passing an untrusted sender distinguished name directly into a logging function as a format string, letting a malicious CMP endpoint crash a client through classic format-string abuse. CVE-2026-63074 allows unbounded growth of cached extra certificates on long-lived CMP server contexts when rejected messages aren’t cleaned up, risking out-of-memory conditions. CVE-2026-63075 similarly enables memory exhaustion in OpenSSL’s QUIC stack, where a peer that withholds acknowledgments while forcing ACK-only packets can inflate per-connection memory usage.
Finally, CVE-2026-75803 affects AEAD ciphers ChaCha20-Poly1305 and AES-OCB, where calling EVP_Cipher() on empty ciphertext can report a successful decryption without actually verifying the authentication tag, opening the door to forged messages. None of the flaws touch the FIPS module boundary.
OpenSSL has released fixed versions for every affected branch: 4.0.2, 3.6.4, 3.5.8, 3.4.7, and 3.0.22, with premium support customers on 1.1.1 and 1.0.2 receiving separate backports. Given the breadth of affected branches and the remote, unauthenticated nature of several bugs, security teams should prioritize inventorying OpenSSL usage across TLS, CMS, and CMP-enabled applications and apply the relevant patched release without delay.