Mbed TLS

An issue was discovered in Mbed TLS versions from 2.19.0 up to 3.6.5, Mbed TLS 4.0.0. Insufficient protection of serialized SSL context or session structures allows an attacker who can modify the serialized structures to induce memory corruption, leading to arbitrary code execution. This is caused by Incorrect Use of Privileged APIs.

An issue was discovered in Mbed TLS through 3.6.5 and TF-PSA-Crypto 1.0.0. A buffer overflow can occur in public key export for FFDH keys.

An issue was discovered in Mbed TLS 3.5.0 through 4.0.0. Client impersonation can occur while resuming a TLS 1.3 session.

An issue was discovered in Mbed TLS 3.5.x and 3.6.x through 3.6.5 and TF-PSA-Crypto 1.0. There is a lack of contributory behavior in FFDH due to improper input validation. Using finite-field Diffie-Hellman, the other party can force the shared secret into a small set of values (lack of contributory behavior). This is a problem for protocols that depend on contributory behavior (which is not the case for TLS). The attack can be carried by the peer, or depending on the protocol by an active network attacker (person in the middle).

Mbed TLS before 3.6.6 and TF-PSA-Crypto before 1.1.0 misuse seeds in a Pseudo-Random Number Generator (PRNG).

An issue was discovered in Mbed TLS 3.x before 3.6.6. An out-of-bounds read vulnerability in mbedtls_ccm_finish() in library/ccm.c allows attackers to obtain adjacent CCM context data via invocation of the multipart CCM API with an oversized tag_len parameter. This is caused by missing validation of the tag_len parameter against the size of the internal 16-byte authentication buffer. The issue affects the public multipart CCM API in Mbed TLS 3.x, where mbedtls_ccm_finish() can be invoked directly by applications. In Mbed TLS 4.x versions prior to the fix, the same missing validation exists in the internal implementation; however, the function is not exposed as part of the public API. Exploitation requires application-level invocation of the multipart CCM API.

An issue was discovered in Mbed TLS through 3.6.5 and 4.x through 4.0.0. There is a NULL pointer dereference in distinguished name parsing that allows an attacker to write to address 0.

Mbed TLS 3.5.0 to 3.6.5 fixed in 3.6.6 and 4.1.0 has a buffer overflow in the x509_inet_pton_ipv6() function

An issue was discovered in Mbed TLS before 3.6.6 and 4.x before 4.1.0 and TF-PSA-Crypto before 1.1.0. There is a Predictable Seed in a Pseudo-Random Number Generator (PRNG).

Mbed TLS v3.3.0 up to 3.6.5 and 4.0.0 allows Algorithm Downgrade.

In Mbed TLS through 4.0.0, there is a compiler-induced timing side channel (in RSA and CBC/ECB decryption) that only occurs with LLVM's select-optimize feature. TF-PSA-Crypto through 1.0.0 is also affected.

Mbed TLS before 3.6.4 allows a use-after-free in certain situations of applications that are developed in accordance with the documentation. The function mbedtls_x509_string_to_names() takes a head argument that is documented as an output argument. The documentation does not suggest that the function will free that pointer; however, the function does call mbedtls_asn1_free_named_data_list() on that argument, which performs a deep free(). As a result, application code that uses this function (relying only on documented behavior) is likely to still hold pointers to the memory blocks that were freed, resulting in a high risk of use-after-free or double-free. In particular, the two sample programs x509/cert_write and x509/cert_req are affected (use-after-free if the san string contains more than one DN).

Mbed TLS through 3.6.4 has an Observable Timing Discrepancy.

Mbed TLS before 3.6.5 allows a local timing attack against certain RSA operations, and direct calls to mbedtls_mpi_mod_inv or mbedtls_mpi_gcd.

Mbed TLS before 3.6.4 has a NULL pointer dereference because mbedtls_asn1_store_named_data can trigger conflicting data with val.p of NULL but val.len greater than zero.

In MbedTLS 3.3.0 before 3.6.4, mbedtls_lms_verify may accept invalid signatures if hash computation fails and internal errors go unchecked, enabling LMS (Leighton-Micali Signature) forgery in a fault scenario. Specifically, unchecked return values in mbedtls_lms_verify allow an attacker (who can induce a hardware hash accelerator fault) to bypass LMS signature verification by reusing stale stack data, resulting in acceptance of an invalid signature. In mbedtls_lms_verify, the return values of the internal Merkle tree functions create_merkle_leaf_value and create_merkle_internal_value are not checked. These functions return an integer that indicates whether the call succeeded or not. If a failure occurs, the output buffer (Tc_candidate_root_node) may remain uninitialized, and the result of the signature verification is unpredictable. When the software implementation of SHA-256 is used, these functions will not fail. However, with hardware-accelerated hashing, an attacker could use fault injection against the accelerator to bypass verification.

In MbedTLS 3.3.0 before 3.6.4, mbedtls_lms_import_public_key does not check that the input buffer is at least 4 bytes before reading a 32-bit field, allowing a possible out-of-bounds read on truncated input. Specifically, an out-of-bounds read in mbedtls_lms_import_public_key allows context-dependent attackers to trigger a crash or limited adjacent-memory disclosure by supplying a truncated LMS (Leighton-Micali Signature) public-key buffer under four bytes. An LMS public key starts with a 4-byte type indicator. The function mbedtls_lms_import_public_key reads this type indicator before validating the size of its input.

Mbed TLS before 3.6.4 has a race condition in AESNI detection if certain compiler optimizations occur. An attacker may be able to extract an AES key from a multithreaded program, or perform a GCM forgery.

Mbed TLS before 3.6.4 has a PEM parsing one-byte heap-based buffer underflow, in mbedtls_pem_read_buffer and two mbedtls_pk_parse functions, via untrusted PEM input.