linux/meta-openembedded.git/meta-oe/recipes-crypto, branch master

botan: upgrade 3.11.1 -> 3.12.0

2026-05-13T07:30:49+00:00

Signed-off-by: Wang Mingyu Signed-off-by: Khem Raj

cryptsetup: update udev package config

2026-05-13T07:15:13+00:00

Commit 1ca8df16af411871e10f268064570146cdef54cb fixed a build problem in wrong way. Relevant rules from lvm2-udevrules were merged into libdevmapper (which was previously pulled by the rules package), however instead of that, the whole lvm2 package was added as dependency (which then pulls the needed libdevmapper). That is a huge package completely unneeded and due to that, this new dependency was later changed to recommendation. Switch to libdevmapper instead of whole lvm2. Keep recommendation for now as it gives more flexibility. Signed-off-by: Peter Marko Signed-off-by: Khem Raj

libkcapi: disable C23 support to fix configure check

2026-04-18T07:11:35+00:00

Set ac_cv_prog_cc_c23=no to prevent autoconf from detecting C23 compiler support, avoiding potential build failures as the package is not yet fully ported to support C23 standard. Signed-off-by: Khem Raj

meta-oe/recipes: disable C23 support to fix configure check

2026-04-16T19:21:07+00:00

leancrypto: upgrade 1.7.1 -> 1.7.2

2026-04-15T14:24:54+00:00

ChangeLog: https://github.com/smuellerDD/leancrypto/releases/tag/v1.7.2 Signed-off-by: Liu Yiding Signed-off-by: Khem Raj

libsodium: upgrade 1.0.21 -> 1.0.22

2026-04-14T15:59:32+00:00

0001-Fix-compilation-with-GCC-on-aarch64.patch removed since it's included in 1.0.22 Changelog: ============ - Post-quantum key encapsulation is now available. ML-KEM768, the NIST-standardized lattice-based KEM, is accessible through the 'crypto_kem_mlkem768_*()' functions. - X-Wing, a hybrid KEM combining ML-KEM768 with X25519 for protection against both classical and quantum adversaries, is available through the 'crypto_kem_*()' functions. X-Wing is the recommended KEM for most applications. - SHA-3 hash functions are now available as 'crypto_hash_sha3256_*()' and 'crypto_hash_sha3512_*()', with both one-shot and streaming APIs. - Performance: NEON optimizations for Argon2 on ARM platforms. - Performance: SHA3 (Keccak1600) now leverages ARM SHA3 instructions when available on ARM platforms. - Performance: WebAssembly SIMD implementations of Argon2 have been added. - Emscripten: LTO is now disabled. With Emscripten 4, LTO produced WebAssembly modules with functions that ran significantly slower than without it. - Emscripten: a new option allows compilation with SIMD support. - Emscripten: native ESM module generation is now supported. - JavaScript sumo builds now allow up to 80 MiB memory usage, so that 'crypto_pwhash' with the interactive settings can be used in pure JavaScript, not just WebAssembly. - XOF state alignment has been relaxed. - 'crypto_core_keccak1600_state' has been added. - Export missing 'crypto_ipcrypt_nd_keygen()' helper function. - 'crypto_auth_hmacsha256_init' and 'crypto_auth_hmacsha512_init' now accept NULL key pointers (with a zero key length), for consistency with other '_init' functions. - apple-xcframework: headers are now in a Clibsodium subdirectory to prevent module.modulemap collisions with other xcframeworks. - Fixed compilation with GCC on aarch64 and gcc 4.x. - On aarch64, aes256-gcm is now enabled even when not using clang, including MSVC. - Added compatibility with Visual Studio 2026 when toolsets do not define PlatformToolsetVersion. - Libsodium can be directly used as a dependency in a Zig project. - Performance of MSVC builds has been improved. Signed-off-by: Wang Mingyu Signed-off-by: Khem Raj

cryptsetup: upgrade 2.8.4 -> 2.8.6

2026-04-10T14:59:52+00:00

Signed-off-by: Wang Mingyu Signed-off-by: Khem Raj

leancrypto: upgrade 1.7.0 -> 1.7.1

2026-04-06T16:46:30+00:00

Contains fix for CVE-2026-34610 (which is however tracked without a version by NVD, so it is marked as patched explicitly) Changelog: - Offer a means to select the AES-C constant time / S-Box implementation via lc_init API - use the AES-C constant time implementation by default - it is about 3 times slower than the AES-C S-Box implementation, but more secure. As the leancrypto library is about secure by default, the CT implementation is just right. Furthermore, if a caller wants to have the faster AES-C S-Box, he can call lc_init(LC_INIT_AES_SBOX) at the beginning. - X.509: fix security issue (CVE-2026-34610) Signed-off-by: Gyorgy Sarvari Signed-off-by: Khem Raj

botan: mark CVE-2026-32877 and CVE-2026-32883 patched

2026-04-06T16:46:30+00:00

Both CVEs were fixed in version 3.11.0, however NVD tracks them without version/CPE info. Relevant commits: CVE-2026-32877: https://github.com/randombit/botan/commit/798a332e11949afa8b004564bb9031e66c1a4d13 CVE-2026-32883: https://github.com/randombit/botan/commit/6ecc62a4e36937d036df8c8eda6a85708abb8c37 Signed-off-by: Gyorgy Sarvari Signed-off-by: Khem Raj

botan: upgrade 3.11.0 -> 3.11.1

2026-04-06T16:46:29+00:00

Contains fixes for CVE-2026-35580 and CVE-2026-35582 Changelog: https://botan.randombit.net/news.html#version-3-11-1-2026-03-31 -CVE-2026-35580: Resolve certificate verification bypass bug introduced in 3.11.0 -CVE-2026-35582: Resolve TLS 1.3 client authentication bypass -Add optimized Argon2 implementation using AVX512 -Add optimized and constant-time Twofish implementation using AVX512/GFNI -Add optimized and constant-time SEED implementation using AVX512/GFNI -Add optimized and constant-time Whirlpool implementations using AVX2 and AVX512 -Add SSSE3/NEON and AVX2 optimized codepaths for CTR -Add constant time implementations of Camellia, ARIA, SEED and SM4 using AES-NI or ARMv8 AES instructions to implement sbox lookups -Improve performance of the AVX512 implementation of SHA-512 especially for Clang -Optimizations for the IDEA modular multiplication -Fix various minor TLS conformance issues flagged by TLS-Anvil -Fix bug in Ed25519 where an invalid signature checked with PK_Verifier might cause a later valid signature to be rejected. -Fix a bug in handling of ECDSA DER-encode signatures where an invalid signature checked with PK_Verifier might cause a later valid signature to be rejected. -Fix a problem introduced in 3.11.0 which could cause crashes on processors without SSSE3 support, particularly when compiled by GCC. -Fix various new warnings from clang-tidy 22 -Fix a compilation error introduced in 3.11.0 which prevented using ffi unless bcrypt was also enabled. -Avoid a macro collision with Microsoft headers that could cause a compilation problem in amalgamation mode. -Enable explicit_bzero, getentropy, getrandom on Hurd Signed-off-by: Gyorgy Sarvari Signed-off-by: Khem Raj