Mbed TLS 2.16.11

Description

This release of Mbed TLS provides bug fixes and minor enhancements. This release includes fixes for security issues.

Security Advisories

For full details, please see the following links:

https://tls.mbed.org/tech-updates/security-advisories/mbedtls-security-advisory-2021-07-1
https://tls.mbed.org/tech-updates/security-advisories/mbedtls-security-advisory-2021-07-2

Release Notes

Security

• Fix a bias in the generation of finite-field Diffie-Hellman-Merkle (DHM)
  private keys and of blinding values for DHM and elliptic curves (ECP)
  computations. Reported by FlorianF89 in #4245.
• Fix a potential side channel vulnerability in ECDSA ephemeral key generation.
  An adversary who is capable of very precise timing measurements could
  learn partial information about the leading bits of the nonce used for the
  signature, allowing the recovery of the private key after observing a
  large number of signature operations. This completes a partial fix in
  Mbed TLS 2.16.4.
• It was possible to configure MBEDTLS_ECP_MAX_BITS to a value that is
  too small, leading to buffer overflows in ECC operations. Fail the build
  in such a case.
• An adversary with access to precise enough information about memory
  accesses (typically, an untrusted operating system attacking a secure
  enclave) could recover an RSA private key after observing the victim
  performing a single private-key operation. Found and reported by
  Zili KOU, Wenjian HE, Sharad Sinha, and Wei ZHANG.
• An adversary with access to precise enough timing information (typically, a
  co-located process) could recover a Curve25519 or Curve448 static ECDH key
  after inputting a chosen public key and observing the victim performing the
  corresponding private-key operation. Found and reported by Leila Batina,
  Lukas Chmielewski, Björn Haase, Niels Samwel and Peter Schwabe.

Bugfix

• Fix premature fopen() call in mbedtls_entropy_write_seed_file which may
  lead to the seed file corruption in case if the path to the seed file is
  equal to MBEDTLS_PLATFORM_STD_NV_SEED_FILE. Contributed by Victor
  Krasnoshchok in #3616.
• Fix some cases in the bignum module where the library constructed an
  unintended representation of the value 0 which was not processed
  correctly by some bignum operations. This could happen when
  mbedtls_mpi_read_string() was called on "-0", or when
  mbedtls_mpi_mul_mpi() and mbedtls_mpi_mul_int() was called with one of
  the arguments being negative and the other being 0. Fixes #4643.
• Fix a compilation error when MBEDTLS_ECP_RANDOMIZE_MXZ_ALT is
  defined. Fixes #4217.
• Fix an incorrect error code when parsing a PKCS#8 private key.
• In a TLS client, enforce the Diffie-Hellman minimum parameter size
  set with mbedtls_ssl_conf_dhm_min_bitlen() precisely. Before, the
  minimum size was rounded down to the nearest multiple of 8.
• In library/net_sockets.c, _POSIX_C_SOURCE and _XOPEN_SOURCE are
  defined to specific values. If the code is used in a context
  where these are already defined, this can result in a compilation
  error. Instead, assume that if they are defined, the values will
  be adequate to build Mbed TLS.
• The cipher suite TLS-RSA-WITH-CAMELLIA-256-GCM-SHA384 was not available
  when SHA-1 was disabled and was offered when SHA-1 was enabled but SHA-384
  was disabled. Fix the dependency. Fixes #4472.
• Fix test suite code on platforms where int32_t is not int, such as
  Arm Cortex-M. Fixes #4530.
• Fix some issues affecting MBEDTLS_ARIA_ALT implementations: a misplaced
  directive in a header and a missing initialization in the self-test.
• Fix a missing initialization in the Camellia self-test, affecting
  MBEDTLS_CAMELLIA_ALT implementations.
• Fix a regression introduced in 2.16.8 which broke (D)TLS CBC ciphersuites
  (when the encrypt-then-MAC extension is not in use) with some ALT
  implementations of the underlying hash (SHA-1, SHA-256, SHA-384), causing
  the affected side to wrongly reject valid messages. Fixes #4118.
• Fix mbedtls_net_poll() and mbedtls_net_recv_timeout() often failing with
  MBEDTLS_ERR_NET_POLL_FAILED on Windows. Fixes #4465.
• Fix a resource leak in a test suite with an alternative AES
  implementation. Fixes #4176.
• Fix a crash in mbedtls_mpi_debug_mpi on a bignum having 0 limbs.
  Reported by lhuang04 in #4578. Fixes #4608.
• Fix a null pointer dereference when mbedtls_mpi_exp_mod() was called with
  A=0 represented with 0 limbs. This bug could not be triggered by code
  that constructed A with one of the mbedtls_mpi_read_xxx functions
  (including in particular TLS code) since those always built an mpi object
  with at least one limb. Credit to OSS-Fuzz. Fixes #4641.
• Fix mbedtls_mpi_gcd(G,A,B) when the value of B is zero. This had no
  effect on Mbed TLS's internal use of mbedtls_mpi_gcd(), but may affect
  applications that call mbedtls_mpi_gcd() directly. Fixes #4642.
• mbedtls_pk_sign() and mbedtls_pk_verify() and their extended and
  restartable variants now require at least the specified hash length if
  nonzero. Before, for RSA, hash_len was ignored in favor of the length of
  the specified hash algorithm.
• Fix which alert is sent in some cases to conform to the
  applicable RFC: on an invalid Finished message value, an
  invalid max_fragment_length extension, or an
  unsupported extension used by the server.

Changes

• Fix the setting of the read timeout in the DTLS sample programs.
• Remove the AES sample application programs/aes/aescrypt2 which shows
  bad cryptographic practice. Fix #1906.
• When building the test suites with GNU make, invoke python3 or python, not
  python2. The build still works with either Python 2.7 or 3.5+, but we
  recommend using a version of Python that is supported upstream.

Who should update

We recommend all users should update to take advantage of the bug fixes contained in this release at an appropriate point in their development lifecycle.

Checksum

The SHA256 hashes for the archives are:

c18e7e9abf95e69e425260493720470021384a1728417042060a35d0b7b18b41 mbedtls-2.16.11.tar.gz
c75ec6a654fc9ef487904172633758d6f61d88a0d53329c79657221043bdb6f4 mbedtls-2.16.11.zip

Mbed TLS 2.26.0

Description

This release of Mbed TLS provides bug fixes, minor enhancements and new features. This release includes fixes for security issues.

API changes

• Renamed the PSA Crypto API output buffer size macros to bring them in line
  with version 1.0.0 of the specification.
• The API glue function mbedtls_ecc_group_of_psa() now takes the curve size
  in bits rather than bytes, with an additional flag to indicate if the
  size may have been rounded up to a whole number of bytes.
• Renamed the PSA Crypto API AEAD tag length macros to bring them in line
  with version 1.0.0 of the specification.

Default behavior changes

• In mbedtls_rsa_context objects, the ver field was formerly documented
  as always 0. It is now reserved for internal purposes and may take
  different values.

New deprecations

• PSA_KEY_EXPORT_MAX_SIZE, PSA_HASH_SIZE, PSA_MAC_FINAL_SIZE,
  PSA_BLOCK_CIPHER_BLOCK_SIZE, PSA_MAX_BLOCK_CIPHER_BLOCK_SIZE and
  PSA_ALG_TLS12_PSK_TO_MS_MAX_PSK_LEN have been renamed, and the old names
  deprecated.
• PSA_ALG_AEAD_WITH_DEFAULT_TAG_LENGTH and PSA_ALG_AEAD_WITH_TAG_LENGTH
  have been renamed, and the old names deprecated.

Features

• The PSA crypto subsystem can now use HMAC_DRBG instead of CTR_DRBG.
  CTR_DRBG is used by default if it is available, but you can override
  this choice by setting MBEDTLS_PSA_HMAC_DRBG_MD_TYPE at compile time.
  Fix #3354.
• Automatic fallback to a software implementation of ECP when
  MBEDTLS_ECP_xxx_ALT accelerator hooks are in use can now be turned off
  through setting the new configuration flag MBEDTLS_ECP_NO_FALLBACK.
• The PSA crypto subsystem can now be configured to use less static RAM by
  tweaking the setting for the maximum amount of keys simultaneously in RAM.
  MBEDTLS_PSA_KEY_SLOT_COUNT sets the maximum number of volatile keys that
  can exist simultaneously. It has a sensible default if not overridden.
• Partial implementation of the PSA crypto driver interface: Mbed TLS can
  now use an external random generator instead of the library's own
  entropy collection and DRBG code. Enable MBEDTLS_PSA_CRYPTO_EXTERNAL_RNG
  and see the documentation of mbedtls_psa_external_get_random() for details.
• Applications using both mbedtls_xxx and psa_xxx functions (for example,
  applications using TLS and MBEDTLS_USE_PSA_CRYPTO) can now use the PSA
  random generator with mbedtls_xxx functions. See the documentation of
  mbedtls_psa_get_random() for details.
• In the PSA API, the policy for a MAC or AEAD algorithm can specify a
  minimum MAC or tag length thanks to the new wildcards
  PSA_ALG_AT_LEAST_THIS_LENGTH_MAC and
  PSA_ALG_AEAD_WITH_AT_LEAST_THIS_LENGTH_TAG.

Security

• Fix a security reduction in CTR_DRBG when the initial seeding obtained a
  nonce from entropy. Applications were affected if they called
  mbedtls_ctr_drbg_set_nonce_len(), if they called
  mbedtls_ctr_drbg_set_entropy_len() with a size that was 3/2 times the key
  length, or when the entropy module uses SHA-256 and CTR_DRBG uses AES-256.
  In such cases, a random nonce was necessary to achieve the advertised
  security strength, but the code incorrectly used a constant instead of
  entropy from the nonce.
  Found by John Stroebel in #3819 and fixed in #3973.
• Fix a buffer overflow in mbedtls_mpi_sub_abs() when calculating
  |A| - |B| where |B| is larger than |A| and has more limbs (so the
  function should return MBEDTLS_ERR_MPI_NEGATIVE_VALUE). Only
  applications calling mbedtls_mpi_sub_abs() directly are affected:
  all calls inside the library were safe since this function is
  only called with |A| >= |B|. Reported by Guido Vranken in #4042.
• Fix an errorneous estimation for an internal buffer in
  mbedtls_pk_write_key_pem(). If MBEDTLS_MPI_MAX_SIZE is set to an odd
  value the function might fail to write a private RSA keys of the largest
  supported size.
  Found by Daniel Otte, reported in #4093 and fixed in #4094.
• Fix a stack buffer overflow with mbedtls_net_poll() and
  mbedtls_net_recv_timeout() when given a file descriptor that is
  beyond FD_SETSIZE. Reported by FigBug in #4169.
• Guard against strong local side channel attack against base64 tables by
  making access aceess to them use constant flow code. (CVE-2021-24119)

Bugfix

• Fix use-after-scope error in programs/ssl/ssl_client2.c and ssl_server2.c
• Fix memory leak that occured when calling psa_close_key() on a
  wrapped key with MBEDTLS_PSA_CRYPTO_SE_C defined.
• Fix an incorrect error code if an RSA private operation glitched.
• Fix a memory leak in an error case in psa_generate_derived_key_internal().
• Fix a resource leak in CTR_DRBG and HMAC_DRBG when MBEDTLS_THREADING_C
  is enabled, on platforms where initializing a mutex allocates resources.
  This was a regression introduced in the previous release. Reported in
  #4017, #4045 and #4071.
• Ensure that calling mbedtls_rsa_free() or mbedtls_entropy_free()
  twice is safe. This happens for RSA when some Mbed TLS library functions
  fail. Such a double-free was not safe when MBEDTLS_THREADING_C was
  enabled on platforms where freeing a mutex twice is not safe.
• Fix a resource leak in a bad-arguments case of mbedtls_rsa_gen_key()
  when MBEDTLS_THREADING_C is enabled on platforms where initializing
  a mutex allocates resources.
• Fixes a bug where, if the library was configured to include support for
  both the old SE interface and the new PSA driver interface, external keys were
  not loaded from storage. This was fixed by #3996.
• This change makes 'mbedtls_x509write_crt_set_basic_constraints'
  consistent with RFC 5280 4.2.1.9 which says: "Conforming CAs MUST
  include this extension in all CA certificates that contain public keys
  used to validate digital signatures on certificates and MUST mark the
  extension as critical in such certificates." Previous to this change,
  the extension was always marked as non-critical. This was fixed by
  #3698.

Changes

• A new library C file psa_crypto_client.c has been created to contain
  the PSA code needed by a PSA crypto client when the PSA crypto
  implementation is not included into the library.
• On recent enough versions of FreeBSD and DragonFlyBSD, the entropy module
  now uses the getrandom syscall instead of reading from /dev/urandom.

Who should update

We recommend all users should update to take advantage of the bug fixes contained in this release at an appropriate point in their development lifecycle.

Checksum

The SHA256 hashes for the archives are:

37949e823c7e1f6695fc56858578df355da0770c284b1c1304cfc8b396d539cd mbedtls-2.26.0.tar.gz
23a7437284fb07c136891cb981a065b9541b7f78d97e3671fe07c7ce59c2f3b3 mbedtls-2.26.0.zip

Mbed TLS 2.16.10

Description

This release of Mbed TLS provides bug fixes and minor enhancements. This release includes fixes for security issues.

Default behavior changes

• In mbedtls_rsa_context objects, the ver field was formerly documented
  as always 0. It is now reserved for internal purposes and may take
  different values.

Security

• Fix a buffer overflow in mbedtls_mpi_sub_abs() when calculating
  |A| - |B| where |B| is larger than |A| and has more limbs (so the
  function should return MBEDTLS_ERR_MPI_NEGATIVE_VALUE). Only
  applications calling mbedtls_mpi_sub_abs() directly are affected:
  all calls inside the library were safe since this function is
  only called with |A| >= |B|. Reported by Guido Vranken in #4042.
• Fix an errorneous estimation for an internal buffer in
  mbedtls_pk_write_key_pem(). If MBEDTLS_MPI_MAX_SIZE is set to an odd
  value the function might fail to write a private RSA keys of the largest
  supported size.
  Found by Daniel Otte, reported in #4093 and fixed in #4094,
  backported in #4100.
• Fix a stack buffer overflow with mbedtls_net_poll() and
  mbedtls_net_recv_timeout() when given a file descriptor that is
  beyond FD_SETSIZE. Reported by FigBug in #4169.
• Guard against strong local side channel attack against base64 tables by
  making access aceess to them use constant flow code. (CVE-2021-24119)

Bugfix

• Fix an incorrect error code if an RSA private operation glitched.
• Fix a resource leak in CTR_DRBG and HMAC_DRBG when MBEDTLS_THREADING_C
  is enabled, on platforms where initializing a mutex allocates resources.
  This was a regression introduced in the previous release. Reported in
  #4017, #4045 and #4071.
• Ensure that calling mbedtls_rsa_free() or mbedtls_entropy_free()
  twice is safe. This happens for RSA when some Mbed TLS library functions
  fail. Such a double-free was not safe when MBEDTLS_THREADING_C was
  enabled on platforms where freeing a mutex twice is not safe.
• Fix a resource leak in a bad-arguments case of mbedtls_rsa_gen_key()
  when MBEDTLS_THREADING_C is enabled on platforms where initializing
  a mutex allocates resources.
• This change makes 'mbedtls_x509write_crt_set_basic_constraints'
  consistent with RFC 5280 4.2.1.9 which says: "Conforming CAs MUST
  include this extension in all CA certificates that contain public keys
  used to validate digital signatures on certificates and MUST mark the
  extension as critical in such certificates." Previous to this change,
  the extension was always marked as non-critical. This was fixed by
  #4044.

Who should update

We recommend all users should update to take advantage of the bug fixes contained in this release at an appropriate point in their development lifecycle.

Checksum

The SHA256 hashes for the archives are:

96257bb03b30300b2f35f861ffe204ed957e9fd0329d80646fe57fc49f589b29 mbedtls-2.16.10.tar.gz
8a1537efd03dacdb0fc4bd6fddac0b0a49f91c229182268815e5980b9ffc555a mbedtls-2.16.10.zip

Mbed TLS 2.7.19

Description

This release of Mbed TLS provides bug fixes and minor enhancements. This release includes fixes for security issues.

Default behavior changes

• In mbedtls_rsa_context objects, the ver field was formerly documented
  as always 0. It is now reserved for internal purposes and may take
  different values.

Security

• Fix a buffer overflow in mbedtls_mpi_sub_abs() when calculating
  |A| - |B| where |B| is larger than |A| and has more limbs (so the
  function should return MBEDTLS_ERR_MPI_NEGATIVE_VALUE). Only
  applications calling mbedtls_mpi_sub_abs() directly are affected:
  all calls inside the library were safe since this function is
  only called with |A| >= |B|. Reported by Guido Vranken in #4042.
• Fix an errorneous estimation for an internal buffer in
  mbedtls_pk_write_key_pem(). If MBEDTLS_MPI_MAX_SIZE is set to an odd
  value the function might fail to write a private RSA keys of the largest
  supported size.
  Found by Daniel Otte, reported in #4093 and fixed in #4094,
  backported in #4099.
• Fix a stack buffer overflow with mbedtls_net_recv_timeout() when given a
  file descriptor that is beyond FD_SETSIZE. Reported by FigBug in #4169.
• Guard against strong local side channel attack against base64 tables by
  making access aceess to them use constant flow code. (CVE-2021-24119)

Bugfix

• Fix a resource leak in CTR_DRBG and HMAC_DRBG when MBEDTLS_THREADING_C
  is enabled, on platforms where initializing a mutex allocates resources.
  This was a regression introduced in the previous release. Reported in
  #4017, #4045 and #4071.
• Ensure that calling mbedtls_rsa_free() or mbedtls_entropy_free()
  twice is safe. This happens for RSA when some Mbed TLS library functions
  fail. Such a double-free was not safe when MBEDTLS_THREADING_C was
  enabled on platforms where freeing a mutex twice is not safe.
• Fix a resource leak in a bad-arguments case of mbedtls_rsa_gen_key()
  when MBEDTLS_THREADING_C is enabled on platforms where initializing
  a mutex allocates resources.
• Fix an incorrect error code if an RSA private operation glitched.
• Fix the build of sample programs when MBEDTLS_PEM_C is enabled but
  MBEDTLS_CERTS_C is disabled. Reported by Michael Schuster in #4206.

Who should update

We recommend all users should update to take advantage of the bug fixes contained in this release at an appropriate point in their development lifecycle.

Checksum

The SHA256 hashes for the archives are:
9a6e0b0386496fae6863e41968eb308034a74008e775a533750af483a38378d0 mbedtls-2.7.19.tar.gz
0f83d43f7de8866820d41db398b6640c8baebb358819223f9b2b3502f77db3d7 mbedtls-2.7.19.zip

Mbed TLS 2.25.0

Description

This release of Mbed TLS provides bug fixes, minor enhancements and new features. This release includes fixes for security issues.

API changes

• The numerical values of the PSA Crypto API macros have been updated to conform to version 1.0.0 of the specification.
• PSA_ALG_STREAM_CIPHER replaces PSA_ALG_CHACHA20 and PSA_ALG_ARC4. The underlying stream cipher is determined by the key type (PSA_KEY_TYPE_CHACHA20 or PSA_KEY_TYPE_ARC4).
• The functions mbedtls_cipher_auth_encrypt() and mbedtls_cipher_auth_decrypt() no longer accept NIST_KW contexts, as they have no way to check if the output buffer is large enough. Please use mbedtls_cipher_auth_encrypt_ext() and mbedtls_cipher_auth_decrypt_ext() instead. Credit to OSS-Fuzz and Cryptofuzz. Fixes #3665.

Requirement changes

• Update the minimum required CMake version to 2.8.12. This silences a warning on CMake 3.19.0. #3801

New deprecations

• PSA_KEY_TYPE_CHACHA20 and PSA_KEY_TYPE_ARC4 have been deprecated. Use PSA_ALG_STREAM_CIPHER instead.
• The functions mbedtls_cipher_auth_encrypt() and mbedtls_cipher_auth_decrypt() are deprecated in favour of the new functions mbedtls_cipher_auth_encrypt_ext() and mbedtls_cipher_auth_decrypt_ext(). Please note that with AEAD ciphers, these new functions always append the tag to the ciphertext, and include the tag in the ciphertext length.

Features

• Partial implementation of the new PSA Crypto accelerator APIs. (Symmetric ciphers, asymmetric signing/verification and key generation, validate_key entry point, and export_public_key interface.)
• Add support for ECB to the PSA cipher API.
• In PSA, allow using a key declared with a base key agreement algorithm in combined key agreement and derivation operations, as long as the key agreement algorithm in use matches the algorithm the key was declared with. This is currently non-standard behaviour, but expected to make it into a future revision of the PSA Crypto standard.
• Add MBEDTLS_TARGET_PREFIX CMake variable, which is prefixed to the mbedtls, mbedcrypto, mbedx509 and apidoc CMake target names. This can be used by external CMake projects that include this one to avoid CMake target name clashes. The default value of this variable is "", so default target names are unchanged.
• Add support for DTLS-SRTP as defined in RFC 5764. Contributed by Johan Pascal, improved by Ron Eldor.
• In the PSA API, it is no longer necessary to open persistent keys: operations now accept the key identifier. The type psa_key_handle_t is now identical to psa_key_id_t instead of being platform-defined. This bridges the last major gap to compliance with the PSA Cryptography specification version 1.0.0. Opening persistent keys is still supported for backward compatibility, but will be deprecated and later removed in future releases.
• PSA_AEAD_NONCE_LENGTH, PSA_AEAD_NONCE_MAX_SIZE, PSA_CIPHER_IV_LENGTH and PSA_CIPHER_IV_MAX_SIZE macros have been added as defined in version 1.0.0 of the PSA Crypto API specification.

Security

• The functions mbedtls_cipher_auth_encrypt() and mbedtls_cipher_auth_decrypt() would write past the minimum documented size of the output buffer when used with NIST_KW. As a result, code using those functions as documented with NIST_KW could have a buffer overwrite of up to 15 bytes, with consequences ranging up to arbitrary code execution depending on the location of the output buffer.
• Limit the size of calculations performed by mbedtls_mpi_exp_mod to MBEDTLS_MPI_MAX_SIZE to prevent a potential denial of service when generating Diffie-Hellman key pairs. Credit to OSS-Fuzz.
• A failure of the random generator was ignored in mbedtls_mpi_fill_random(), which is how most uses of randomization in asymmetric cryptography (including key generation, intermediate value randomization and blinding) are implemented. This could cause failures or the silent use of non-random values. A random generator can fail if it needs reseeding and cannot not obtain entropy, or due to an internal failure (which, for Mbed TLS's own CTR_DRBG or HMAC_DRBG, can only happen due to a misconfiguration).
• Fix a compliance issue whereby we were not checking the tag on the algorithm parameters (only the size) when comparing the signature in the description part of the cert to the real signature. This meant that a NULL algorithm parameters entry would look identical to an array of REAL (size zero) to the library and thus the certificate would be considered valid. However, if the parameters do not match in any way then the certificate should be considered invalid, and indeed OpenSSL marks these certs as invalid when mbedtls did not. Many thanks to guidovranken who found this issue via differential fuzzing and reported it in #3629.
• Zeroising of local buffers and variables which are used for calculations in mbedtls_pkcs5_pbkdf2_hmac(), mbedtls_internal_sha*_process(), mbedtls_internal_md*_process() and mbedtls_internal_ripemd160_process() functions to erase sensitive data from memory. Reported by Johan Malmgren and Johan Uppman Bruce from Sectra.

Bugfix

• Fix an invalid (but nonzero) return code from mbedtls_pk_parse_subpubkey() when the input has trailing garbage. Fixes #2512.
• Fix build failure in configurations where MBEDTLS_USE_PSA_CRYPTO is enabled but ECDSA is disabled. Contributed by jdurkop. Fixes #3294.
• Include the psa_constant_names generated source code in the source tree instead of generating it at build time. Fixes #3524.
• Fix rsa_prepare_blinding() to retry when the blinding value is not invertible (mod N), instead of returning MBEDTLS_ERR_RSA_RNG_FAILED. This addresses a regression but is rare in practice (approx. 1 in 2/sqrt(N)). Found by Synopsys Coverity, fix contributed by Peter Kolbus (Garmin). Fixes #3647.
• Use socklen_t on Android and other POSIX-compliant system
• Fix the build when the macro _GNU_SOURCE is defined to a non-empty value. Fix #3432.
• Consistently return PSA_ERROR_INVALID_ARGUMENT on invalid cipher input sizes (instead of PSA_ERROR_BAD_STATE in some cases) to make the psa_cipher_* functions compliant with the PSA Crypto API specification.
• mbedtls_ecp_curve_list() now lists Curve25519 and Curve448 under the names "x25519" and "x448". These curves support ECDH but not ECDSA. If you need only the curves that support ECDSA, filter the list with mbedtls_ecdsa_can_do().
• Fix psa_generate_key() returning an error when asked to generate an ECC key pair on Curve25519 or secp244k1.
• Fix psa_key_derivation_output_key() to allow the output of a combined key agreement and subsequent key derivation operation to be used as a key inside of the PSA Crypto core.
• Fix handling of EOF against 0xff bytes and on platforms with unsigned chars. Fixes a build failure on platforms where char is unsigned. Fixes #3794.
• Fix an off-by-one error in the additional data length check for CCM, which allowed encryption with a non-standard length field. Fixes #3719.
• Correct the default IV size for mbedtls_cipher_info_t structures using MBEDTLS_MODE_ECB to 0, since ECB mode ciphers don't use IVs.
• Make arc4random_buf available on NetBSD and OpenBSD when _POSIX_C_SOURCE is defined. Fix contributed in #3571.
• Fix conditions for including string.h in error.c. Fixes #3866.
• psa_set_key_id() now also sets the lifetime to persistent for keys located in a secure element.
• Attempting to create a volatile key with a non-zero key identifier now fails. Previously the key identifier was just ignored when creating a volatile key.
• Attempting to create or register a key with a key identifier in the vendor range now fails.
• Fix build failures on GCC 11. Fixes #3782.
• Add missing arguments of debug message in mbedtls_ssl_decrypt_buf.
• Fix a memory leak in mbedtls_mpi_sub_abs() when the result was negative (an error condition) and the second operand was aliased to the result.
• Fix a case in elliptic curve arithmetic where an out-of-memory condition could go undetected, resulting in an incorrect result.
• In CTR_DRBG and HMAC_DRBG, don't reset the reseed interval in seed(). Fixes #2927.
• In PEM writing functions, fill the trailing part of the buffer with null bytes. This guarantees that the corresponding parsing function can read the buffer back, which was the case for mbedtls_x509write_{crt,csr}_pem until this property was inadvertently broken in Mbed TLS 2.19.0. Fixes #3682.
• Fix a build failure that occurred with the MBEDTLS_AES_SETKEY_DEC_ALT option on. In this configuration key management methods that are required for MBEDTLS_CIPHER_MODE_XTS were excluded from the build and made it fail. Fixes #3818. Reported by John Stroebel.

Changes

• Reduce stack usage significantly during sliding window exponentiation. Reported in #3591 and fix contributed in #3592 by Daniel Otte.
• The PSA persistent storage format is updated to always store the key bits attribute. No automatic upgrade path is provided. Previously stored keys must be erased, or manually upgraded based on the key storage format specification (docs/architecture/mbed-crypto-storage-specification.md). Fixes #3740.
• Remove the zeroization of a pointer variable in AES rounds. It was valid but spurious and misleading since it looked like a mistaken attempt to zeroize the pointed-to buffer. Reported by Antonio de la Piedra, CEA Leti, France.

Who should update

         ...

Mbed TLS 2.16.9

Description

This release of Mbed TLS provides bug fixes and minor enhancements. This release includes fixes for security issues.

Security

• Limit the size of calculations performed by mbedtls_mpi_exp_mod to MBEDTLS_MPI_MAX_SIZE to prevent a potential denial of service when generating Diffie-Hellman key pairs. Credit to OSS-Fuzz.
• A failure of the random generator was ignored in mbedtls_mpi_fill_random(), which is how most uses of randomization in asymmetric cryptography (including key generation, intermediate value randomization and blinding) are implemented. This could cause failures or the silent use of non-random values. A random generator can fail if it needs reseeding and cannot not obtain entropy, or due to an internal failure (which, for Mbed TLS's own CTR_DRBG or HMAC_DRBG, can only happen due to a misconfiguration).
• Fix a compliance issue whereby we were not checking the tag on the algorithm parameters (only the size) when comparing the signature in the description part of the cert to the real signature. This meant that a NULL algorithm parameters entry would look identical to an array of REAL (size zero) to the library and thus the certificate would be considered valid. However, if the parameters do not match in any way then the certificate should be considered invalid, and indeed OpenSSL marks these certs as invalid when mbedtls did not. Many thanks to guidovranken who found this issue via differential fuzzing and reported it in #3629.
• Zeroising of local buffers and variables which are used for calculations in mbedtls_pkcs5_pbkdf2_hmac(), mbedtls_internal_sha*_process(), mbedtls_internal_md*_process() and mbedtls_internal_ripemd160_process() functions to erase sensitive data from memory. Reported by Johan Malmgren and Johan Uppman Bruce from Sectra.

Bugfix

• Fix an invalid (but nonzero) return code from mbedtls_pk_parse_subpubkey() when the input has trailing garbage. Fixes #2512.
• Fix rsa_prepare_blinding() to retry when the blinding value is not invertible (mod N), instead of returning MBEDTLS_ERR_RSA_RNG_FAILED. This addresses a regression but is rare in practice (approx. 1 in 2/sqrt(N)). Found by Synopsys Coverity, fix contributed by Peter Kolbus (Garmin). Fixes #3647.
• Fix the build when the macro _GNU_SOURCE is defined to a non-empty value. Fix #3432.
• Correct the default IV size for mbedtls_cipher_info_t structures using MBEDTLS_MODE_ECB to 0, since ECB mode ciphers don't use IVs.
• Make arc4random_buf available on NetBSD and OpenBSD when _POSIX_C_SOURCE is defined. Fix contributed in #3571. Adopted for LTS branch 2.16 in #3602.
• Fix build failures on GCC 11. Fixes #3782.
• Fix a memory leak in mbedtls_mpi_sub_abs() when the result was negative (an error condition) and the second operand was aliased to the result.
• Fix a case in elliptic curve arithmetic where an out-of-memory condition could go undetected, resulting in an incorrect result.
• In CTR_DRBG and HMAC_DRBG, don't reset the reseed interval in seed().
  Fixes #2927.
• In PEM writing functions, fill the trailing part of the buffer with null bytes. This guarantees that the corresponding parsing function can read the buffer back, which was the case for mbedtls_x509write_{crt,csr}_pem until this property was inadvertently broken in Mbed TLS 2.19.0. Fixes #3682.
• Fix a build failure that occurred with the MBEDTLS_AES_SETKEY_DEC_ALT option on. In this configuration key management methods that are required for MBEDTLS_CIPHER_MODE_XTS were excluded from the build and made it fail. Fixes #3818. Reported by John Stroebel.

Changes

• Reduce stack usage significantly during sliding window exponentiation. Reported in #3591 and fix contributed in #3592 by Daniel Otte.
• Remove the zeroization of a pointer variable in AES rounds. It was valid but spurious and misleading since it looked like a mistaken attempt to zeroize the pointed-to buffer. Reported by Antonio de la Piedra, CEA Leti, France.

Who should update

We recommend all affected users should update to take advantage of the bug fixes contained in this release at an appropriate point in their development lifecycle.

Checksum

The SHA256 hashes for the archives are:

fc17ff7d8c11d08f23ae2800a18269408ad2c24ea6bb8b9363e41a01c2425697  mbedtls-2.16.9.tar.gz
111300d0d307d738fb20596c358fda7129dfd5f5c682f055793262080ac50950  mbedtls-2.16.9.zip

Mbed TLS 2.7.18

Description

This release of Mbed TLS provides bug fixes and minor enhancements. This release includes fixes for security issues.

Security

• Limit the size of calculations performed by mbedtls_mpi_exp_mod to MBEDTLS_MPI_MAX_SIZE to prevent a potential denial of service when generating Diffie-Hellman key pairs. Credit to OSS-Fuzz.
• Fix a compliance issue whereby we were not checking the tag on the algorithm parameters (only the size) when comparing the signature in the description part of the cert to the real signature. This meant that a NULL algorithm parameters entry would look identical to an array of REAL (size zero) to the library and thus the certificate would be considered valid. However, if the parameters do not match in any way then the certificate should be considered invalid, and indeed OpenSSL marks these certs as invalid when mbedtls did not. Many thanks to guidovranken who found this issue via differential fuzzing and reported it in #3629.
• Zeroising of local buffers and variables which are used for calculations in mbedtls_pkcs5_pbkdf2_hmac(), mbedtls_internal_sha*_process(), mbedtls_internal_md*_process() and mbedtls_internal_ripemd160_process() functions to erase sensitive data from memory. Reported by Johan Malmgren and Johan Uppman Bruce from Sectra.

Bugfix

• Fix an invalid (but nonzero) return code from mbedtls_pk_parse_subpubkey() when the input has trailing garbage. Fixes #2512.
• Fix rsa_prepare_blinding() to retry when the blinding value is not invertible (mod N), instead of returning MBEDTLS_ERR_RSA_RNG_FAILED. This addresses a regression but is rare in practice (approx. 1 in 2/sqrt(N)). Found by Synopsys Coverity, fix contributed by Peter Kolbus (Garmin). Fixes #3647.
• Correct the default IV size for mbedtls_cipher_info_t structures using MBEDTLS_MODE_ECB to 0, since ECB mode ciphers don't use IVs.
• Fix build failures on GCC 11. Fixes #3782.
• Fix a memory leak in mbedtls_mpi_sub_abs() when the result was negative (an error condition) and the second operand was aliased to the result.
• Fix a case in elliptic curve arithmetic where an out-of-memory condition could go undetected, resulting in an incorrect result.
• In CTR_DRBG and HMAC_DRBG, don't reset the reseed interval in seed(). Fixes #2927.

Changes

• Reduce stack usage significantly during sliding window exponentiation. Reported in #3591 and fix contributed in #3592 by Daniel Otte.
• Remove the zeroization of a pointer variable in AES rounds. It was valid but spurious and misleading since it looked like a mistaken attempt to zeroize the pointed-to buffer. Reported by Antonio de la Piedra, CEA Leti, France.

Who should update

We recommend all affected users should update to take advantage of the bug fixes contained in this release at an appropriate point in their development lifecycle.

Checksum

The SHA256 hashes for the archives are:

f4a0f0c62e7016fa96f6ee78e2b8ba05d1f653c3aa2734fd73fbec02c3637e23  mbedtls-2.7.18.tar.gz
2b169bbf8c04ef9c9776a7ddb487b3387d4be2f3220a112c6d52111f643b9a22  mbedtls-2.7.18.zip

Mbed TLS 2.24.0

Description

This release of Mbed TLS provides bug fixes and minor enhancements. This release includes fixes for security issues and the most notable of them are described in more detail in the security advisories.

API changes

• In the PSA API, rename the types of elliptic curve and Diffie-Hellman group families to psa_ecc_family_t and psa_dh_family_t, in line with the PSA Crypto API specification version 1.0.0. Rename associated macros as well:
  PSA_ECC_CURVE_xxx renamed to PSA_ECC_FAMILY_xxx
  PSA_DH_GROUP_xxx renamed to PSA_DH_FAMILY_xxx
  PSA_KEY_TYPE_GET_CURVE renamed to to PSA_KEY_TYPE_ECC_GET_FAMILY
  PSA_KEY_TYPE_GET_GROUP renamed to PSA_KEY_TYPE_DH_GET_FAMILY

Default behavior changes

• Stop storing persistent information about externally stored keys created through PSA Crypto with a volatile lifetime. Reported in #3288 and contributed by Steven Cooreman in #3382.

Features

• The new function mbedtls_ecp_write_key() exports private ECC keys back to a byte buffer. It is the inverse of the existing mbedtls_ecp_read_key().
• Support building on e2k (Elbrus) architecture: correctly enable -Wformat-signedness, and fix the code that causes signed-one-bit-field and sign-compare warnings. Contributed by makise-homura (Igor Molchanov) akemi_homura@kurisa.ch.

Security

• Fix a vulnerability in the verification of X.509 certificates when matching the expected common name (the cn argument of mbedtls_x509_crt_verify()) with the actual certificate name: when the subjecAltName extension is present, the expected name was compared to any name in that extension regardless of its type. This means that an attacker could for example impersonate a 4-bytes or 16-byte domain by getting a certificate for the corresponding IPv4 or IPv6 (this would require the attacker to control that IP address, though). Similar attacks using other subjectAltName name types might be possible. Found and reported by kFYatek in #3498.
• When checking X.509 CRLs, a certificate was only considered as revoked if its revocationDate was in the past according to the local clock if available. In particular, on builds without MBEDTLS_HAVE_TIME_DATE, certificates were never considered as revoked. On builds with MBEDTLS_HAVE_TIME_DATE, an attacker able to control the local clock (for example, an untrusted OS attacking a secure enclave) could prevent revocation of certificates via CRLs. Fixed by no longer checking the revocationDate field, in accordance with RFC 5280. Reported by yuemonangong in #3340. Reported independently and fixed by Raoul Strackx and Jethro Beekman in #3433.
• In (D)TLS record decryption, when using a CBC ciphersuites without the Encrypt-then-Mac extension, use constant code flow memory access patterns to extract and check the MAC. This is an improvement to the existing countermeasure against Lucky 13 attacks. The previous countermeasure was effective against network-based attackers, but less so against local attackers. The new countermeasure defends against local attackers, even if they have access to fine-grained measurements. In particular, this fixes a local Lucky 13 cache attack found and reported by Tuba Yavuz, Farhaan Fowze, Ken (Yihan) Bai, Grant Hernandez, and Kevin Butler (University of Florida) and Dave Tian (Purdue University).
• Fix side channel in RSA private key operations and static (finite-field) Diffie-Hellman. An adversary with precise enough timing and memory access information (typically an untrusted operating system attacking a secure enclave) could bypass an existing counter-measure (base blinding) and potentially fully recover the private key.
• Fix a 1-byte buffer overread in mbedtls_x509_crl_parse_der(). Credit to OSS-Fuzz for detecting the problem and to Philippe Antoine for pinpointing the problematic code.
• Zeroising of plaintext buffers in mbedtls_ssl_read() to erase unused application data from memory. Reported in #689 by Johan Uppman Bruce of Sectra.

Bugfix

• Library files installed after a CMake build no longer have execute permission.
• Use local labels in mbedtls_padlock_has_support() to fix an invalid symbol redefinition if the function is inlined. Reported in #3451 and fix contributed in #3452 by okhowang.
• Fix the endianness of Curve25519 keys imported/exported through the PSA APIs. psa_import_key and psa_export_key will now correctly expect/output Montgomery keys in little-endian as defined by RFC7748. Contributed by Steven Cooreman in #3425.
• Fix build errors when the only enabled elliptic curves are Montgomery curves. Raised by signpainter in #941 and by Taiki-San in #1412. This also fixes missing declarations reported by Steven Cooreman in #1147.
• Fix self-test failure when the only enabled short Weierstrass elliptic curve is secp192k1. Fixes #2017.
• PSA key import will now correctly import a Curve25519/Curve448 public key instead of erroring out. Contributed by Steven Cooreman in #3492.
• Use arc4random_buf on NetBSD instead of rand implementation with cyclical lower bits. Fix contributed in #3540.
• Fix a memory leak in mbedtls_md_setup() when using HMAC under low memory conditions. Reported and fix suggested by Guido Vranken in #3486.
• Fix bug in redirection of unit test outputs on platforms where stdout is defined as a macro. First reported in #2311 and fix contributed in #3528.

Changes

• Only pass -Wformat-signedness to versions of GCC that support it. Reported in #3478 and fix contributed in #3479 by okhowang.
• Reduce the stack consumption of mbedtls_x509write_csr_der() which previously could lead to stack overflow on constrained devices. Contributed by Doru Gucea and Simon Leet in #3464.
• Undefine the ASSERT macro before defining it locally, in case it is defined in a platform header. Contributed by Abdelatif Guettouche in #3557.
• Update copyright notices to use Linux Foundation guidance. As a result, the copyright of contributors other than Arm is now acknowledged, and the years of publishing are no longer tracked in the source files. This also eliminates the need for the lines declaring the files to be part of MbedTLS. Fixes #3457.
• Add the command line parameter key_pwd to the ssl_client2 and ssl_server2 example applications which allows to provide a password for the key file specified through the existing key_file argument. This allows the use of these applications with password-protected key files. Analogously but for ssl_server2 only, add the command line parameter key_pwd2 which allows to set a password for the key file provided through the existing key_file2 argument.

Who should update

We recommend all affected users should update to take advantage of the bug fixes contained in this release at an appropriate point in their development lifecycle.

Checksum

The SHA256 hashes for the archives are:

d436ae4892bd80329ca18a3960052fbb42d3f1f46c7519711d6763621ca6cfa0  mbedtls-2.24.0.tar.gz
d7b129288a957b1936ade3921504d5524e34a0dbe9f4d965df3416db9cfe0dae  mbedtls-2.24.0.zip

Mbed TLS 2.16.8

Description

This release of Mbed TLS provides bug fixes and minor enhancements. This release includes fixes for security issues and the most notable of them are described in more detail in the security advisories.

Features

• Support building on e2k (Elbrus) architecture: correctly enable -Wformat-signedness, and fix the code that causes signed-one-bit-field and sign-compare warnings. Contributed by makise-homura (Igor Molchanov) akemi_homura@kurisa.ch.

Security

• When checking X.509 CRLs, a certificate was only considered as revoked if its revocationDate was in the past according to the local clock if available. In particular, on builds without MBEDTLS_HAVE_TIME_DATE, certificates were never considered as revoked. On builds with MBEDTLS_HAVE_TIME_DATE, an attacker able to control the local clock (for example, an untrusted OS attacking a secure enclave) could prevent revocation of certificates via CRLs. Fixed by no longer checking the revocationDate field, in accordance with RFC 5280. Reported by yuemonangong in #3340. Reported independently and fixed by Raoul Strackx and Jethro Beekman in #3433.
• In (D)TLS record decryption, when using a CBC ciphersuites without the Encrypt-then-Mac extension, use constant code flow memory access patterns to extract and check the MAC. This is an improvement to the existing countermeasure against Lucky 13 attacks. The previous countermeasure was effective against network-based attackers, but less so against local attackers. The new countermeasure defends against local attackers, even if they have access to fine-grained measurements. In particular, this fixes a local Lucky 13 cache attack found and reported by Tuba Yavuz, Farhaan Fowze, Ken (Yihan) Bai, Grant Hernandez, and Kevin Butler (University of Florida) and Dave Tian (Purdue University).
• Fix side channel in RSA private key operations and static (finite-field) Diffie-Hellman. An adversary with precise enough timing and memory access information (typically an untrusted operating system attacking a secure enclave) could bypass an existing counter-measure (base blinding) and potentially fully recover the private key.
• Fix a 1-byte buffer overread in mbedtls_x509_crl_parse_der(). Credit to OSS-Fuzz for detecting the problem and to Philippe Antoine for pinpointing the problematic code.
• Zeroising of plaintext buffers in mbedtls_ssl_read() to erase unused application data from memory. Reported in #689 by Johan Uppman Bruce of Sectra.

Bugfix

• Avoid use of statically sized stack buffers for certificate writing. This previously limited the maximum size of DER encoded certificates in mbedtls_x509write_crt_der() to 2Kb. Reported by soccerGB in #2631.
• Reduce the stack consumption of mbedtls_x509write_csr_der() which previously could lead to stack overflow on constrained devices. Contributed by Doru Gucea and Simon Leet in #3464.
• Use arc4random_buf on NetBSD instead of rand implementation with cyclical lower bits. Fix contributed in #3540.
• Fix building library/net_sockets.c and the ssl_mail_client program on NetBSD. NetBSD conditionals were added for the backport to avoid the risk of breaking a platform. Original fix contributed by Nia Alarie in #3422. Adapted for long-term support branch 2.16 in #3558.
• Fix bug in redirection of unit test outputs on platforms where stdout is defined as a macro. First reported in #2311 and fix contributed in #3528. Adopted for LTS branch 2.16 in #3601.

Changes

• Update copyright notices to use Linux Foundation guidance. As a result, the copyright of contributors other than Arm is now acknowledged, and the years of publishing are no longer tracked in the source files. This also eliminates the need for the lines declaring the files to be part of MbedTLS. Fixes #3457.

Who should update

We recommend all affected users should update to take advantage of the bug fixes contained in this release at an appropriate point in their development lifecycle.

Checksum

The SHA256 hashes for the archives are:

fe9e3b15c3375943bdfebbbb20dd6b4f1147b3b5d926248bd835d73247407430  mbedtls-2.16.8.tar.gz
a2904aa4f8b23a3e3972f87ff1c7e450a128c38d00ac28ad183296607fe2d9d6  mbedtls-2.16.8.zip

Mbed TLS 2.7.17

Description

This release of Mbed TLS provides bug fixes and minor enhancements. This release includes fixes for security issues and the most notable of them are described in more detail in the security advisories.

Features

• Support building on e2k (Elbrus) architecture: correctly enable -Wformat-signedness, and fix the code that causes signed-one-bit-field and sign-compare warnings. Contributed by makise-homura (Igor Molchanov) akemi_homura@kurisa.ch.

Security

• When checking X.509 CRLs, a certificate was only considered as revoked if its revocationDate was in the past according to the local clock if available. In particular, on builds without MBEDTLS_HAVE_TIME_DATE, certificates were never considered as revoked. On builds with MBEDTLS_HAVE_TIME_DATE, an attacker able to control the local clock (for example, an untrusted OS attacking a secure enclave) could prevent revocation of certificates via CRLs. Fixed by no longer checking the revocationDate field, in accordance with RFC 5280. Reported by yuemonangong in #3340. Reported independently and fixed by Raoul Strackx and Jethro Beekman in #3433.
• In (D)TLS record decryption, when using a CBC ciphersuites without the Encrypt-then-Mac extension, use constant code flow memory access patterns to extract and check the MAC. This is an improvement to the existing countermeasure against Lucky 13 attacks. The previous countermeasure was effective against network-based attackers, but less so against local attackers. The new countermeasure defends against local attackers, even if they have access to fine-grained measurements. In particular, this fixes a local Lucky 13 cache attack found and reported by Tuba Yavuz, Farhaan Fowze, Ken (Yihan) Bai, Grant Hernandez, and Kevin Butler (University of Florida) and Dave Tian (Purdue University).
• Fix side channel in RSA private key operations and static (finite-field) Diffie-Hellman. An adversary with precise enough timing and memory access information (typically an untrusted operating system attacking a secure enclave) could bypass an existing counter-measure (base blinding) and potentially fully recover the private key.
• Fix a 1-byte buffer overread in mbedtls_x509_crl_parse_der(). Credit to OSS-Fuzz for detecting the problem and to Philippe Antoine for pinpointing the problematic code.
• Zeroising of plaintext buffers in mbedtls_ssl_read() to erase unused application data from memory. Reported in #689 by Johan Uppman Bruce of Sectra.

Bugfix

• Use arc4random_buf on NetBSD instead of rand implementation with cyclical lower bits. Fix contributed in #3540.
• Fix building library/net_sockets.c on NetBSD. NetBSD conditionals were added for the backport to avoid the risk of breaking a platform. Original fix contributed by Nia Alarie in #3422. Adopted for long-term support branch 2.7 in #3570.
• Fix bug in redirection of unit test outputs on platforms where stdout is defined as a macro. First reported in #2311 and fix contributed in #3528. Adopted for LTS branch 2.7 in #3600.

Changes

• Update copyright notices to use Linux Foundation guidance. As a result, the copyright of contributors other than Arm is now acknowledged, and the years of publishing are no longer tracked in the source files. This also eliminates the need for the lines declaring the files to be part of MbedTLS. Fixes #3457.

Who should update

We recommend all affected users should update to take advantage of the bug fixes contained in this release at an appropriate point in their development lifecycle.

Checksum

The SHA256 hashes for the archives are:

fae5441631a8696d1eae71f61b8276d728b41a905954326019830cdaf0afeb42  mbedtls-2.7.17.tar.gz
a009059b0b4b882b884e8ef7013ff068b1254d8a2d98243e000c67b1737956b6  mbedtls-2.7.17.zip