mbedtls-2-2.28.6-bp156.1.2

- Update to 2.28.6:
  Changes:
  * Mbed TLS is now released under a dual Apache-2.0 OR GPL-2.0-or-later
    license. Users may choose which license they take the code under.

- Update to 2.28.5:
  Features
  * The documentation of mbedtls_ecp_group now describes the optimized
    representation of A for some curves. Fixes gh#Mbed-TLS/mbedtls#8045.
  Security
  * Developers using mbedtls_pkcs5_pbes2() or mbedtls_pkcs12_pbe() should
    review the size of the output buffer passed to this function, and note that
    the output after decryption may include CBC padding. Consider moving to the
    new functions mbedtls_pkcs5_pbes2_ext() or mbedtls_pkcs12_pbe_ext() which
    checks for overflow of the output buffer and reports the actual length of
    the output.
  * Improve padding calculations in CBC decryption, NIST key unwrapping and
    RSA OAEP decryption. With the previous implementation, some compilers
    (notably recent versions of Clang and IAR) could produce non-constant time
    code, which could allow a padding oracle attack if the attacker has access
    to precise timing measurements.
  * Fix a buffer overread when parsing short TLS application data records in
    ARC4 or null-cipher cipher suites. Credit to OSS-Fuzz.
  Bugfix
  * Fix x509 certificate generation to conform to RFC 5480 / RFC 5758 when
    using ECC key. The certificate was rejected by some crypto frameworks.
    Fixes gh#Mbed-TLS/mbedtls#2924.
  * Fix some cases where mbedtls_mpi_mod_exp, RSA key construction or ECDSA
    signature can silently return an incorrect result in low memory conditions.
  * Fix IAR compiler warnings. Fixes gh#Mbed-TLS/mbedtls#7873,
    gh#Mbed-TLS/mbedtls#4300.
  * Fix an issue when parsing an otherName subject alternative name into a
    mbedtls_x509_san_other_name struct. The type-id of the otherName was not
    copied to the struct. This meant that the struct had incomplete information
    about the otherName SAN and contained uninitialized memory.
  * Fix the detection of HardwareModuleName otherName SANs. These were being
    detected by comparing the wrong field and the check was erroneously
    inverted.
  * Fix an error when MBEDTLS_ECDSA_SIGN_ALT is defined but not
    MBEDTLS_ECDSA_VERIFY_ALT, causing ecdsa verify to fail. Fixes
    gh#Mbed-TLS/mbedtls#7498.  Functions in the ssl_cache module now return a
    negative MBEDTLS_ERR_xxx error code on failure. Before, they returned 1 to
    indicate failure in some cases involving a missing entry or a full cache.
  Changes
  * In configurations with ARIA or Camellia but not AES, the value of
    MBEDTLS_CIPHER_BLKSIZE_MAX was 8, rather than 16 as the name might suggest.
    This did not affect any library code, because this macro was only used in
    relation with CMAC which does not support these ciphers.  Its value is now
    16 if ARIA or Camellia are present. This may affect application code that
    uses this macro.

- Update to 2.28.4:
  Features
  * Allow MBEDTLS_CONFIG_FILE and MBEDTLS_USER_CONFIG_FILE to be set by
    setting the CMake variable of the same name at configuration time.
  Bugfix
  * Fix crypt_and_hash decryption fail when used with a stream cipher
    mode of operation, due to the input not being a multiple of the block
    size. Resolves #7417.
  * Fix a bug where mbedtls_x509_string_to_names() would return success
    when given a invalid name string, if it did not contain '=' or ','.
  * Fix missing PSA initialization in sample programs when
    MBEDTLS_USE_PSA_CRYPTO is enabled.
  * Fix clang and armclang compilation error when targeting certain Arm
    M-class CPUs (Cortex-M0, Cortex-M0+, Cortex-M1, Cortex-M23,
    SecurCore SC000). Fixes #1077.
  * Fixed an issue that caused compile errors when using CMake and the IAR
    toolchain.
  * Fix the build with MBEDTLS_PSA_INJECT_ENTROPY. Fixes #7516.
  * Fix builds on Windows with clang.
  * Fix compilation warnings in aes.c for certain combinations
    of configuration options.
  * Fix a compilation error on some platforms when including mbedtls/ssl.h
    with all TLS support disabled. Fixes #6628.
  Changes
  * Update test data to avoid failures of unit tests after 2023-08-07, and
    update expiring certififcates in the certs

- Fix the use of the %{release} macro for the mbedtls-devel version

- Update to 2.28.3:
  Features
  * Use HOSTCC (if it is set) when compiling C code during generation of the
    configuration-independent files. This allows them to be generated when CC
    is set for cross compilation.
  * AES-NI is now supported with Visual Studio.
  * AES-NI is now supported in 32-bit builds, or when MBEDTLS_HAVE_ASM is
    disabled, when compiling with GCC or Clang or a compatible compiler for a
    target CPU that supports the requisite instructions (for example gcc -m32
  - msse2 -maes -mpclmul). (Generic x86 builds with GCC-like compilers still
    require MBEDTLS_HAVE_ASM and a 64-bit target.)
  Security
  * MBEDTLS_AESNI_C, which is enabled by default, was silently ignored on
    builds that couldn't compile the GCC-style assembly implementation (most
    notably builds with Visual Studio), leaving them vulnerable to timing
    side-channel attacks. There is now an intrinsics-based AES-NI
    implementation as a fallback for when the assembly one cannot be used.
  Bugfix
  * Fix a build issue on Windows where the source and build directory could
    not be on different drives (#5751).
  * Fix possible integer overflow in mbedtls_timing_hardclock(), which
    could cause a crash for certain platforms & compiler options.
  * Fix IAR compiler warnings. Fixes #6924.
  * Fix a bug in the build where directory names containing spaces were
    causing generate_errors.pl to error out resulting in a build failure.
    Fixes issue #6879.
  * Fix compile error where MBEDTLS_RSA_C and MBEDTLS_X509_CRT_WRITE_C are
    defined, but MBEDTLS_PK_RSA_ALT_SUPPORT is not defined. Fixes #3174.
  * Fix a build issue when defining MBEDTLS_TIMING_ALT and MBEDTLS_SELF_TEST.
    The library would not link if the user didn't provide an external self-test
    function. The self-test is now provided regardless of the choice of
    internal/alternative timing implementation. Fixes #6923.
  * mbedtls_x509write_crt_set_serial() now explicitly rejects serial numbers
    whose binary representation is longer than 20 bytes. This was already
    forbidden by the standard (RFC5280 - section 4.1.2.2) and now it's being
    enforced also at code level.
  * Fix potential undefined behavior in mbedtls_mpi_sub_abs(). Reported by
    Pascal Cuoq using TrustInSoft Analyzer in #6701; observed independently by
    Aaron Ucko under Valgrind.
  * Fix behavior of certain sample programs which could, when run with no
    arguments, access uninitialized memory in some cases. Fixes #6700 (which
    was found by TrustInSoft Analyzer during REDOCS'22) and #1120.
  * Fix build errors in test programs when MBEDTLS_CERTS_C is disabled.
    Fixes #6243.
  * Fix parsing of X.509 SubjectAlternativeName extension. Previously,
    malformed alternative name components were not caught during initial
    certificate parsing, but only on subsequent calls to
    mbedtls_x509_parse_subject_alt_name(). Fixes #2838.
  * Fix bug in conversion from OID to string in
    mbedtls_oid_get_numeric_string(). OIDs such as 2.40.0.25 are now printed
    correctly.
  * Reject OIDs with overlong-encoded subidentifiers when converting them to a
    string.
  * Reject OIDs with subidentifier values exceeding UINT_MAX. Such
    subidentifiers can be valid, but Mbed TLS cannot currently handle them.
  * Reject OIDs that have unterminated subidentifiers, or (equivalently) have
    the most-significant bit set in their last byte.
  * Silence a warning about an unused local variable in bignum.c on some
    architectures. Fixes #7166.
  * Silence warnings from clang -Wdocumentation about empty \retval
    descriptions, which started appearing with Clang 15. Fixes #6960.
  * Fix undefined behavior in mbedtls_ssl_read() and mbedtls_ssl_write() if
    len argument is 0 and buffer is NULL.
  Changes
  * The C code follows a new coding style. This is transparent for users but
    affects contributors and maintainers of local patches. For more
    information, see
    https://mbed-tls.readthedocs.io/en/latest/kb/how-to/rewrite-branch-for-coding-style/
  * Changed the default MBEDTLS_ECP_WINDOW_SIZE from 6 to 2.  As tested in
    issue 6790, the correlation between this define and RSA decryption
    performance has changed lately due to security fixes.  To fix the
    performance degradation when using default values the window was reduced
    from 6 to 2, a value that gives the best or close to best results when
    tested on Cortex-M4 and Intel i7.

- Setup the mbedtls-2 package

- Build AVX2 enabled hwcaps library for x86_64-v3

- Update to 2.28.2: (bsc#1206576, CVE-2022-46393)
  Security:
  * Fix potential heap buffer overread and overwrite in DTLS if
    MBEDTLS_SSL_DTLS_CONNECTION_ID is enabled and
    MBEDTLS_SSL_CID_IN_LEN_MAX > 2 * MBEDTLS_SSL_CID_OUT_LEN_MAX.
  * 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 if the window size used for the
    exponentiation was 3 or smaller. Found and reported by Zili KOU,
    Wenjian HE, Sharad Sinha, and Wei ZHANG. See "Cache Side-channel Attacks
    and Defenses of the Sliding Window Algorithm in TEEs" - Design, Automation
    and Test in Europe 2023.
  Bugfix:
  * Fix a long-standing build failure when building x86 PIC code with old
    gcc (4.x). The code will be slower, but will compile. We do however
    recommend upgrading to a more recent compiler instead. Fixes #1910.
  * Fix support for little-endian Microblaze when MBEDTLS_HAVE_ASM is defined.
    Contributed by Kazuyuki Kimura to fix #2020.
  * Use double quotes to include private header file psa_crypto_cipher.h.
    Fixes 'file not found with include' error when building with Xcode.
  * Fix handling of broken symlinks when loading certificates using
    mbedtls_x509_crt_parse_path(). Instead of returning an error as soon as a
    broken link is encountered, skip the broken link and continue parsing
    other certificate files. Contributed by Eduardo Silva in #2602.
  * Fix a compilation error when using CMake with an IAR toolchain.
    Fixes #5964.
  * Fix bugs and missing dependencies when building and testing
    configurations with only one encryption type enabled in TLS 1.2.
  * Provide the missing definition of mbedtls_setbuf() in some configurations
    with MBEDTLS_PLATFORM_C disabled. Fixes #6118, #6196.
  * Fix compilation errors when trying to build with
    PSA drivers for AEAD (GCM, CCM, Chacha20-Poly1305).
  * Fix memory leak in ssl_parse_certificate_request() caused by
    mbedtls_x509_get_name() not freeing allocated objects in case of error.
    Change mbedtls_x509_get_name() to clean up allocated objects on error.
  * Fix checks on PK in check_config.h for builds with PSA and RSA. This does
    not change which builds actually work, only moving a link-time error to
    an early check.
  * Fix ECDSA verification, where it was not always validating the
    public key. This bug meant that it was possible to verify a
    signature with an invalid public key, in some cases. Reported by
    Guido Vranken using Cryptofuzz in #4420.
  * Fix a possible null pointer dereference if a memory allocation fails
    in TLS PRF code. Reported by Michael Madsen in #6516.
  * Fix a bug in which mbedtls_x509_crt_info() would produce non-printable
    bytes when parsing certificates containing a binary RFC 4108
    HardwareModuleName as a Subject Alternative Name extension. Hardware
    serial numbers are now rendered in hex format. Fixes #6262.
  * Fix bug in error reporting in dh_genprime.c where upon failure,
    the error code returned by mbedtls_mpi_write_file() is overwritten
    and therefore not printed.
  * In the bignum module, operations of the form (-A) - (+A) or (-A) - (-A)
    with A > 0 created an unintended representation of the value 0 which was
    not processed correctly by some bignum operations. Fix this. This had no
    consequence on cryptography code, but might affect applications that call
    bignum directly and use negative numbers.
  * Fix undefined behavior (typically harmless in practice) of
    mbedtls_mpi_add_mpi(), mbedtls_mpi_add_abs() and mbedtls_mpi_add_int()
    when both operands are 0 and the left operand is represented with 0 limbs.
  * Fix undefined behavior (typically harmless in practice) when some bignum
    functions receive the most negative value of mbedtls_mpi_sint. Credit
    to OSS-Fuzz. Fixes #6597.
  * Fix undefined behavior (typically harmless in practice) in PSA ECB
    encryption and decryption.

- Update to 2.28.1: (CVE-2022-35409)
  Default behavior changes
  * mbedtls_cipher_set_iv will now fail with ChaCha20 and
    ChaCha20+Poly1305 for IV lengths other than 12. The library was
    silently overwriting this length with 12, but did not inform
    the caller about it.
    gh#Mbed-TLS/mbedtls#4301
  Features
  * When MBEDTLS_PSA_CRYPTO_CONFIG is enabled, you may list the PSA
    crypto feature requirements in the file named by the new macro
    MBEDTLS_PSA_CRYPTO_CONFIG_FILE instead of the default
    psa/crypto_config.h. Furthermore you may name an additional
    file to include after the main file with the macro
    MBEDTLS_PSA_CRYPTO_USER_CONFIG_FILE.
  Security
  * Zeroize dynamically-allocated buffers used by the PSA Crypto
    key storage module before freeing them. These buffers contain
    secret key material, and could thus potentially leak the key
    through freed heap.
  * Fix a potential heap buffer overread in TLS 1.2 server-side
    when MBEDTLS_USE_PSA_CRYPTO is enabled, an opaque key (created
    with mbedtls_pk_setup_opaque()) is provisioned, and a static
    ECDH ciphersuite is selected. This may result in an application
    crash or potentially an information leak.
  * Fix a buffer overread in DTLS ClientHello parsing in servers
    with MBEDTLS_SSL_DTLS_CLIENT_PORT_REUSE enabled.
    An unauthenticated client or a man-in-the-middle could cause a
    DTLS server to read up to 255 bytes after the end of the SSL
    input buffer. The buffer overread only happens when
    MBEDTLS_SSL_IN_CONTENT_LEN is less than a threshold that
    depends on the exact configuration: 258 bytes if using
    mbedtls_ssl_cookie_check(), and possibly up to 571 bytes with
    a custom cookie check function.
    Reported by the Cybeats PSI Team.
  Bugfix
  * Fix a memory leak if mbedtls_ssl_config_defaults() is called
    twice.
  * Fix several bugs (warnings, compiler and linker errors, test
    failures) in reduced configurations when MBEDTLS_USE_PSA_CRYPTO
    is enabled.
  * Fix a bug in (D)TLS curve negotiation: when
    MBEDTLS_USE_PSA_CRYPTO was enabled and an ECDHE-ECDSA or
    ECDHE-RSA key exchange was used, the client would fail to check
    that the curve selected by the server for ECDHE was indeed one
    that was offered. As a result, the client would accept any
    curve that it supported, even if that curve was not allowed
    according to its configuration.
    gh#Mbed-TLS/mbedtls#5291
  * Fix unit tests that used 0 as the file UID. This failed on some
    implementations of PSA ITS.
    gh#Mbed-TLS/mbedtls#3838
  * Fix API violation in mbedtls_md_process() test by adding a call
    to mbedtls_md_starts().
    gh#Mbed-TLS/mbedtls#2227
  * Fix compile errors when MBEDTLS_HAVE_TIME is not defined.
    Add tests to catch bad uses of time.h.
  * Fix bug in the alert sending function
    mbedtls_ssl_send_alert_message() potentially leading to
    corrupted alert messages being sent in case the function needs
    to be re-called after initially returning
    MBEDTLS_SSL_WANT_WRITE.
    gh#Mbed-TLS/mbedtls#1916
  * In configurations with MBEDTLS_SSL_DTLS_CONNECTION_ID enabled
    but none of MBEDTLS_SSL_HW_RECORD_ACCEL,
    MBEDTLS_SSL_EXPORT_KEYS or MBEDTLS_DEBUG_C, DTLS handshakes
    using CID would crash due to a null pointer dereference.
    Fix this.
    gh#Mbed-TLS/mbedtls#3998
  * Fix incorrect documentation of mbedtls_x509_crt_profile. The
    previous documentation stated that the allowed_pks field
    applies to signatures only, but in fact it does apply to the
    public key type of the end entity certificate, too.
    gh#Mbed-TLS/mbedtls#1992
  * Fix PSA cipher multipart operations using ARC4. Previously, an
    IV was required but discarded. Now, an IV is rejected, as it
    should be.
  * Fix undefined behavior in mbedtls_asn1_find_named_data(), where
    val is not NULL and val_len is zero. psa_raw_key_agreement()
    now returns PSA_ERROR_BUFFER_TOO_SMALL when applicable.
    gh#Mbed-TLS/mbedtls#5735
  * Fix a bug in the x25519 example program where the removal of
    MBEDTLS_ECDH_LEGACY_CONTEXT caused the program not to run.
    gh#Mbed-TLS/mbedtls#4901
    gh#Mbed-TLS/mbedtls#3191
  * Encode X.509 dates before 1/1/2000 as UTCTime rather than
    GeneralizedTime.
    gh#Mbed-TLS/mbedtls#5465
  * Fix order value of curve x448.
  * Fix string representation of DNs when outputting values
    containing commas and other special characters, conforming to
    RFC 1779.
    gh#Mbed-TLS/mbedtls#769
  * Silence a warning from GCC 12 in the selftest program.
    gh#Mbed-TLS/mbedtls#5974
  * Fix mbedtls_asn1_write_mpi() writing an incorrect encoding of
    0.
  * Fix resource leaks in mbedtls_pk_parse_public_key() in low
    memory conditions.
  * Fix server connection identifier setting for outgoing encrypted
    records on DTLS 1.2 session resumption. After DTLS 1.2 session
    resumption with connection identifier, the Mbed TLS client now
    properly sends the server connection identifier in encrypted
    record headers.
    gh#Mbed-TLS/mbedtls#5872
  * Fix a null pointer dereference when performing some operations
    on zero represented with 0 limbs (specifically
    mbedtls_mpi_mod_int() dividing by 2, and
    mbedtls_mpi_write_string() in base 2).
  * Fix record sizes larger than 16384 being sometimes accepted
    despite being non-compliant. This could not lead to a buffer
    overflow. In particular, application data size was already
    checked correctly.

- Fix baselib.conf

- Update to 2.28.0: (bsc#1193979, CVE-2021-45450)
  API changes
  * Some fields of mbedtls_ssl_session and mbedtls_ssl_config are in a
  different order. This only affects applications that define such
  structures directly or serialize them.
  Requirement changes
  * Sign-magnitude and one's complement representations for signed integers are
  not supported. Two's complement is the only supported representation.
  Removals
  * Remove config option MBEDTLS_TLS_DEFAULT_ALLOW_SHA1_IN_CERTIFICATES,
  which allowed SHA-1 in the default TLS configuration for certificate
  signing. It was intended to facilitate the transition in environments
  with SHA-1 certificates. SHA-1 is considered a weak message digest and
  its use constitutes a security risk.
  * Remove the partial support for running unit tests via Greentea on Mbed OS,
  which had been unmaintained since 2018.
  Features
  * The identifier of the CID TLS extension can be configured by defining
  MBEDTLS_TLS_EXT_CID at compile time.
  * Warn if errors from certain functions are ignored. This is currently
  supported on GCC-like compilers and on MSVC and can be configured through
  the macro MBEDTLS_CHECK_RETURN. The warnings are always enabled
  (where supported) for critical functions where ignoring the return
  value is almost always a bug. Enable the new configuration option
  MBEDTLS_CHECK_RETURN_WARNING to get warnings for other functions. This
  is currently implemented in the AES, DES and md modules, and will be
  extended to other modules in the future.
  * Add missing PSA macros declared by PSA Crypto API 1.0.0:
  PSA_ALG_IS_SIGN_HASH, PSA_ALG_NONE, PSA_HASH_BLOCK_LENGTH, PSA_KEY_ID_NULL.
  * Add new API mbedtls_ct_memcmp for constant time buffer comparison.
  * Add PSA API definition for ARIA.
  Security
  * Zeroize several intermediate variables used to calculate the expected
  value when verifying a MAC or AEAD tag. This hardens the library in
  case the value leaks through a memory disclosure vulnerability. For
  example, a memory disclosure vulnerability could have allowed a
  man-in-the-middle to inject fake ciphertext into a DTLS connection.
  * In psa_cipher_generate_iv() and psa_cipher_encrypt(), do not read back
  from the output buffer. This fixes a potential policy bypass or decryption
  oracle vulnerability if the output buffer is in memory that is shared with
  an untrusted application.
  * Fix a double-free that happened after mbedtls_ssl_set_session() or
  mbedtls_ssl_get_session() failed with MBEDTLS_ERR_SSL_ALLOC_FAILED
  (out of memory). After that, calling mbedtls_ssl_session_free()
  and mbedtls_ssl_free() would cause an internal session buffer to
  be free()'d twice.
  Bugfix
  * Stop using reserved identifiers as local variables. Fixes #4630.
  * The GNU makefiles invoke python3 in preference to python except on Windows.
  * The check was accidentally not performed when cross-compiling for Windows
  on Linux. Fix this. Fixes #4774.
  * Prevent divide by zero if either of PSA_CIPHER_ENCRYPT_OUTPUT_SIZE() or
  PSA_CIPHER_UPDATE_OUTPUT_SIZE() were called using an asymmetric key type.
  * Fix a parameter set but unused in psa_crypto_cipher.c. Fixes #4935.
  * Don't use the obsolete header path sys/fcntl.h in unit tests.
  These header files cause compilation errors in musl.
  Fixes #4969.
  * Fix missing constraints on x86_64 and aarch64 assembly code
  for bignum multiplication that broke some bignum operations with
  (at least) Clang 12.
  Fixes #4116, #4786, #4917, #4962.
  * Fix mbedtls_cipher_crypt: AES-ECB when MBEDTLS_USE_PSA_CRYPTO is enabled.
  * Failures of alternative implementations of AES or DES single-block
  functions enabled with MBEDTLS_AES_ENCRYPT_ALT, MBEDTLS_AES_DECRYPT_ALT,
  MBEDTLS_DES_CRYPT_ECB_ALT or MBEDTLS_DES3_CRYPT_ECB_ALT were ignored.
  This does not concern the implementation provided with Mbed TLS,
  where this function cannot fail, or full-module replacements with
  MBEDTLS_AES_ALT or MBEDTLS_DES_ALT. Reported by Armelle Duboc in #1092.
  * Some failures of HMAC operations were ignored. These failures could only
  happen with an alternative implementation of the underlying hash module.
  * Fix the error returned by psa_generate_key() for a public key. Fixes #4551.
  * Fix the build of sample programs when neither MBEDTLS_ERROR_C nor
  MBEDTLS_ERROR_STRERROR_DUMMY is enabled.
  * Fix PSA_ALG_RSA_PSS verification accepting an arbitrary salt length.
  This algorithm now accepts only the same salt length for verification
  that it produces when signing, as documented. Use the new algorithm
  PSA_ALG_RSA_PSS_ANY_SALT to accept any salt length. Fixes #4946.
  * The existing predicate macro name PSA_ALG_IS_HASH_AND_SIGN is now reserved
  for algorithm values that fully encode the hashing step, as per the PSA
  Crypto API specification. This excludes PSA_ALG_RSA_PKCS1V15_SIGN_RAW and
  PSA_ALG_ECDSA_ANY. The new predicate macro PSA_ALG_IS_SIGN_HASH covers
  all algorithms that can be used with psa_{sign,verify}_hash(), including
  these two.
  * Fix issue in Makefile on Linux with SHARED=1, that caused shared libraries
  not to list other shared libraries they need.
  * Fix a bug in mbedtls_gcm_starts() when the bit length of the iv
  exceeds 2^32. Fixes #4884.
  * Fix an uninitialized variable warning in test_suite_ssl.function with GCC
  version 11.
  * Fix the build when no SHA2 module is included. Fixes #4930.
  * Fix the build when only the bignum module is included. Fixes #4929.
  * Fix a potential invalid pointer dereference and infinite loop bugs in
  pkcs12 functions when the password is empty. Fix the documentation to
  better describe the inputs to these functions and their possible values.
  Fixes #5136.
  * The key usage flags PSA_KEY_USAGE_SIGN_MESSAGE now allows the MAC
  operations psa_mac_compute() and psa_mac_sign_setup().
  * The key usage flags PSA_KEY_USAGE_VERIFY_MESSAGE now allows the MAC
  operations psa_mac_verify() and psa_mac_verify_setup().
  Changes
  * Set config option MBEDTLS_TLS_DEFAULT_ALLOW_SHA1_IN_KEY_EXCHANGE to be
  disabled by default.
  * Improve the performance of base64 constant-flow code. The result is still
  slower than the original non-constant-flow implementation, but much faster
  than the previous constant-flow implementation. Fixes #4814.
  * Indicate in the error returned if the nonce length used with
  ChaCha20-Poly1305 is invalid, and not just unsupported.
  * The mbedcrypto library includes a new source code module constant_time.c,
  containing various functions meant to resist timing side channel attacks.
  * This module does not have a separate configuration option, and functions
  from this module will be included in the build as required. Currently
  most of the interface of this module is private and may change at any
  time.

- Update to 2.27.0:
  API changes:
  * Update AEAD output size macros to bring them in line with the PSA Crypto
    API version 1.0 spec. This version of the spec parameterizes them on the
    key type used, as well as the key bit-size in the case of
    PSA_AEAD_TAG_LENGTH.
    The old versions of these macros were renamed and deprecated as follows:
  - PSA_AEAD_TAG_LENGTH          -> PSA_AEAD_TAG_LENGTH_1_ARG
  - PSA_AEAD_ENCRYPT_OUTPUT_SIZE -> PSA_AEAD_ENCRYPT_OUTPUT_SIZE_2_ARG
  - PSA_AEAD_DECRYPT_OUTPUT_SIZE -> PSA_AEAD_DECRYPT_OUTPUT_SIZE_2_ARG
  - PSA_AEAD_UPDATE_OUTPUT_SIZE  -> PSA_AEAD_UPDATE_OUTPUT_SIZE_2_ARG
  - PSA_AEAD_FINISH_OUTPUT_SIZE  -> PSA_AEAD_FINISH_OUTPUT_SIZE_1_ARG
  - PSA_AEAD_VERIFY_OUTPUT_SIZE  -> PSA_AEAD_VERIFY_OUTPUT_SIZE_1_ARG
  * Implement one-shot cipher functions, psa_cipher_encrypt and
    psa_cipher_decrypt, according to the PSA Crypto API 1.0.0
    specification.
  Features:
  * Add mbedtls_rsa_rsassa_pss_sign_ext() function allowing to generate a
    signature with a specific salt length. This function allows to validate
    test cases provided in the NIST's CAVP test suite.
  * Added support for built-in driver keys through the PSA opaque crypto
    driver interface. Refer to the documentation of
    MBEDTLS_PSA_CRYPTO_BUILTIN_KEYS for more information.
  * Implement psa_sign_message() and psa_verify_message().
  * The new function mbedtls_mpi_random() generates a random value in a
    given range uniformly.
  * Implement psa_mac_compute() and psa_mac_verify() as defined in the
    PSA Cryptograpy API 1.0.0 specification.
  * MBEDTLS_ECP_MAX_BITS is now determined automatically from the configured
    curves and no longer needs to be configured explicitly to save RAM.
  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.
  * 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 2.20.0.
  * 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.
  * 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.
  Bugfix
  * Add printf function attributes to mbedtls_debug_print_msg to ensure we
    get printf format specifier warnings.
  * Fix premature fopen() call in mbedtls_entropy_write_seed_file which may
    lead to seed file corruption in the case where the path to the seed file is
    equal to MBEDTLS_PLATFORM_STD_NV_SEED_FILE.
  * PSA functions other than psa_open_key now return PSA_ERROR_INVALID_HANDLE
    rather than PSA_ERROR_DOES_NOT_EXIST for an invalid handle, bringing them
    in line with version 1.0.0 of the specification.
  * PSA functions creating a key now return PSA_ERROR_INVALID_ARGUMENT rather
    than PSA_ERROR_INVALID_HANDLE when the identifier specified for the key
    to create is not valid, bringing them in line with version 1.0.0 of the
    specification.
  * 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.
  * Fix a bug in ECDSA that would cause it to fail when the hash is all-bits
    zero.
  * Fix a compilation error when MBEDTLS_ECP_RANDOMIZE_MXZ_ALT is
    defined.
  * 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.
  * Do not offer SHA384 cipher suites when SHA-384 is disabled.
  * With MBEDTLS_PSA_CRYPTO_C disabled, some functions were getting built
    nonetheless, resulting in undefined reference errors when building a
    shared library.
  * Fix test suite code on platforms where int32_t is not int, such as
    Arm Cortex-M.
  * 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.
  * Restore the ability to configure PSA via Mbed TLS options to support RSA
    key pair operations but exclude RSA key generation. When MBEDTLS_GENPRIME
    is not defined PSA will no longer attempt to use mbedtls_rsa_gen_key().
  * Fix a regression introduced in 2.24.0 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.
  * Remove outdated check-config.h check that prevented implementing the
    timing module on Mbed OS.
  * Fix PSA_ALG_TLS12_PRF and PSA_ALG_TLS12_PSK_TO_MS being too permissive
    about missing inputs.
  * Fix a resource leak in a test suite with an alternative AES
    implementation.
  * Fix a crash in mbedtls_mpi_debug_mpi on a bignum having 0 limbs. This
    could notably be triggered by setting the TLS debug level to 3 or above
    and using a Montgomery curve for the key exchange.
  * psa_verify_hash() was relying on implementation-specific behavior of
    mbedtls_rsa_rsassa_pss_verify() and was causing failures in some _ALT
    implementations. This reliance is now removed.
  * Disallow inputs of length different from the corresponding hash when
    signing or verifying with PSA_ALG_RSA_PSS (The PSA Crypto API mandates
    that PSA_ALG_RSA_PSS uses the same hash throughout the algorithm.)
  * Fix a null pointer dereference when mbedtls_mpi_exp_mod() was called with
    A=0 represented with 0 limbs. Up to and including Mbed TLS 2.26, 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.
  * 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.
  * The PSA API no longer allows the creation or destruction of keys with a
    read-only lifetime. The persistence level PSA_KEY_PERSISTENCE_READ_ONLY
    can now only be used as intended, for keys that cannot be modified through
    normal use of the API.
  * When MBEDTLS_PSA_CRYPTO_SPM is enabled, crypto_spe.h was not included
    in all the right places. Include it from crypto_platform.h, which is
    the natural place.
  * mbedtls_pk_sign() and mbedtls_pk_verify() and their extended and
    restartable variants now always honor 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.
  * Correct (change from 12 to 13 bytes) the value of the macro describing the
    maximum nonce length returned by psa_aead_generate_nonce().
  Changes
  * Add extra printf compiler warning flags to builds.
  * Fix memsan build false positive in x509_crt.c with Clang 11
  * 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.
  * Alternative implementations of CMAC may now opt to not support 3DES as a
    CMAC block cipher, and still pass the CMAC self test.
  * Remove configs/config-psa-crypto.h, which was identical to the default
    configuration except for having some extra cryptographic mechanisms
    enabled and for unintended differences. This configuration was primarily
    intended to demonstrate the PSA API, and lost most of its usefulness when
    MBEDTLS_PSA_CRYPTO_C became enabled by default.
  * When building the test suites with GNU make, invoke python3 or python, not
    python2, which is no longer supported upstream.
  * When using session cache based session resumption on the server,
    double-check that custom session cache implementations return
    sessions which are consistent with the negotiated ciphersuite
    and compression method.
  * Fix build failure on MinGW toolchain when __USE_MING_ANSI_STDIO is on.
    When that flag is on, standard GNU C printf format specifiers
    should be used.
  * Reduce the default value of MBEDTLS_ECP_WINDOW_SIZE. This reduces RAM usage
    during ECC operations at a negligible performance cost.
  * mbedtls_mpi_read_binary(), mbedtls_mpi_read_binary_le() and
    mbedtls_mpi_read_string() now construct an mbedtls_mpi object with 0 limbs
    when their input has length 0. Note that this is an implementation detail
    and can change at any time, so this change should be transparent, but it
    may result in mbedtls_mpi_write_binary() or mbedtls_mpi_write_string()
    now writing an empty string where it previously wrote one or more
    zero digits when operating from values constructed with an mpi_read
    function and some mpi operations.
  * Implicitly add PSA_KEY_USAGE_SIGN_MESSAGE key usage policy flag when
    PSA_KEY_USAGE_SIGN_HASH flag is set and PSA_KEY_USAGE_VERIFY_MESSAGE flag
    when PSA_KEY_USAGE_VERIFY_HASH flag is set. This usage flag extension
    is also applied when loading a key from storage.
  * Remove mbedtls-4237.patch upstream
  * Library soname bump to libmbedcrypto7

-  Workaround for building with gcc-11 boo#1181876

- Update to 2.26.0: [bsc#1189589, CVE-2021-24119]
  * * This release of Mbed TLS provides bug fixes, minor enhancements and new
  features. This release includes fixes for security issues.
  * see https://github.com/ARMmbed/mbedtls/releases/tag/v2.26.0
- Fix build with patch from https://github.com/ARMmbed/mbedtls/pull/4237
  mbedtls-4237.patch

- Fix build for Leap targets
- Use upstream tarball name

- update to 2.25.0:
  * This release of Mbed TLS provides bug fixes, minor enhancements and new
  features. This release includes fixes for security issues.
  * see https://github.com/ARMmbed/mbedtls/releases/tag/v2.25.0
  * 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.

- update to 2.24.0:
  * see https://github.com/ARMmbed/mbedtls/releases/tag/v2.24.0
  * 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.
  * 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
  * 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.

- Add workaround for failing builds (Python not found) due to
  https://gitlab.kitware.com/cmake/cmake/-/issues/21168

- Do not run testsuite in parallel - its not reliable

- update to 2.23.0:
  a lot of changes see https://github.com/ARMmbed/mbedtls/releases/tag/v2.23.0
  * Fix a side channel vulnerability in modular exponentiation that could reveal an RSA private key used in a secure enclave. Noticed by Sangho Lee, Ming-Wei Shih, Prasun Gera, Taesoo Kim and Hyesoon Kim (Georgia Institute of Technology); and Marcus Peinado (Microsoft Research). Reported by Raoul Strackx (Fortanix) in #3394.
  * Fix side channel in mbedtls_ecp_check_pub_priv() and mbedtls_pk_parse_key() / mbedtls_pk_parse_keyfile() (when loading a private key that didn't include the uncompressed public key), as well as mbedtls_ecp_mul() / mbedtls_ecp_mul_restartable() when called with a NULL f_rng argument. An attacker with access to precise enough timing and memory access information (typically an untrusted operating system attacking a secure enclave) could fully recover the ECC private key. Found and reported by Alejandro Cabrera Aldaya and Billy Brumley.
  * Fix issue in Lucky 13 counter-measure that could make it ineffective when hardware accelerators were used (using one of the MBEDTLS_SHAxxx_ALT macros). This would cause the original Lucky 13 attack to be possible in those configurations, allowing an active network attacker to recover plaintext after repeated timing measurements under some conditions. Reported and fix suggested by Luc Perneel in #3246.