pull update [run tests] [extended tests]

Signed-off-by: Basil Hess <[email protected]>
open-quantum-safe · Jan 30, 2025 · df95edc · df95edc
1 parent bb69318
commit df95edc
Show file tree

Hide file tree

Showing 118 changed files with 9,445 additions and 12,730 deletions.
diff --git a/docs/algorithms/kem/ml_kem.md b/docs/algorithms/kem/ml_kem.md
@@ -7,9 +7,9 @@
 - **Authors' website**: https://pq-crystals.org/kyber/ and https://csrc.nist.gov/pubs/fips/203
 - **Specification version**: ML-KEM.
 - **Primary Source**<a name="primary-source"></a>:
-  - **Source**: https://github.com/pq-code-package/mlkem-native/commit/84398e7230fa31ba4241f5eb36bdc3c1dbbd5bcd with copy_from_upstream patches
+  - **Source**: https://github.com/pq-code-package/mlkem-native/commit/d830bc22eb1613bbe38028cfefc33f1a52a40b2f with copy_from_upstream patches
   - **Implementation license (SPDX-Identifier)**: CC0-1.0 or Apache-2.0
-- **Optimized Implementation sources**: https://github.com/pq-code-package/mlkem-native/commit/84398e7230fa31ba4241f5eb36bdc3c1dbbd5bcd with copy_from_upstream patches
+- **Optimized Implementation sources**: https://github.com/pq-code-package/mlkem-native/commit/d830bc22eb1613bbe38028cfefc33f1a52a40b2f with copy_from_upstream patches
   - **cupqc-cuda**:<a name="cupqc-cuda"></a>
       - **Source**: https://github.com/praveksharma/cupqc-mlkem/commit/b026f4e5475cd9c20c2082c7d9bad80e5b0ba89e
       - **Implementation license (SPDX-Identifier)**: Apache-2.0

diff --git a/docs/algorithms/kem/ml_kem.yml b/docs/algorithms/kem/ml_kem.yml
@@ -17,7 +17,7 @@ website: https://pq-crystals.org/kyber/ and https://csrc.nist.gov/pubs/fips/203
 nist-round: FIPS203
 spec-version: ML-KEM
 primary-upstream:
-  source: https://github.com/pq-code-package/mlkem-native/commit/84398e7230fa31ba4241f5eb36bdc3c1dbbd5bcd
+  source: https://github.com/pq-code-package/mlkem-native/commit/d830bc22eb1613bbe38028cfefc33f1a52a40b2f
     with copy_from_upstream patches
   spdx-license-identifier: CC0-1.0 or Apache-2.0
 optimized-upstreams:

diff --git a/scripts/copy_from_upstream/copy_from_upstream.yml b/scripts/copy_from_upstream/copy_from_upstream.yml
@@ -34,7 +34,7 @@ upstreams:
     name: mlkem-native
     git_url: https://github.com/pq-code-package/mlkem-native.git
     git_branch: main
-    git_commit: 84398e7230fa31ba4241f5eb36bdc3c1dbbd5bcd
+    git_commit: d830bc22eb1613bbe38028cfefc33f1a52a40b2f
     kem_meta_path: 'integration/liboqs/{pretty_name_full}_META.yml'
     kem_scheme_path: '.'
     patches: [mlkem-native-ml_kem.patch]

diff --git a/src/kem/ml_kem/CMakeLists.txt b/src/kem/ml_kem/CMakeLists.txt
diff --git a/src/kem/ml_kem/mlkem-native_ml-kem-1024_aarch64/native/aarch64/README.md b/src/kem/ml_kem/mlkem-native_ml-kem-1024_aarch64/native/aarch64/README.md
@@ -10,10 +10,9 @@ works:
 - _Fast and Clean: Auditable high-performance assembly via constraint solving_, Amin Abdulrahman, Hanno Becker, Matthias
   J. Kannwischer, Fabien Klein, [https://eprint.iacr.org/2022/1303](https://eprint.iacr.org/2022/1303)
 
-## Profiles
 
-This backend comes with two profiles: "clean" and optimized. The "clean" backend is handwritten and meant to be easy to
-read and modify; for example, is heavily leverages register aliases and assembly macros. The optimized profile is
-automatically generated from the clean profile via [SLOTHY](https://github.com/slothy-optimizer/slothy). Currently, the
+## Variants
+
+This backend comes in two versions: "clean" and optimized. The "clean" backend is handwritten and meant to be easy to read and modify; for example, is heavily leverages register aliases and assembly macros. This directory contains the optimized version, which is automatically generated from the clean one via [SLOTHY](https://github.com/slothy-optimizer/slothy). Currently, the
 target architecture is Cortex-A55, but you can easily re-optimize the code for a different microarchitecture supported
-by SLOTHY, by adjusting the parameters in [optimize.sh](src/optimize.sh).
+by SLOTHY, by adjusting the parameters in [optimize.sh](../../../test/aarch64_clean/src/optimize.sh).
diff --git a/src/kem/ml_kem/mlkem-native_ml-kem-1024_aarch64/native/aarch64/clean.h b/src/kem/ml_kem/mlkem-native_ml-kem-1024_aarch64/native/aarch64/clean.h
diff --git a/src/kem/ml_kem/mlkem-native_ml-kem-1024_aarch64/native/aarch64/opt.h b/src/kem/ml_kem/mlkem-native_ml-kem-1024_aarch64/native/aarch64/opt.h
@@ -3,8 +3,6 @@
  * SPDX-License-Identifier: Apache-2.0
  */
 
-/* ML-KEM arithmetic native profile for clean assembly */
-
 #ifdef MLKEM_NATIVE_ARITH_PROFILE_H
 #error Only one MLKEM_ARITH assembly profile can be defined -- did you include multiple profiles?
 #else

diff --git a/src/kem/ml_kem/mlkem-native_ml-kem-1024_aarch64/native/aarch64/src/aarch64_zetas.c b/src/kem/ml_kem/mlkem-native_ml-kem-1024_aarch64/native/aarch64/src/aarch64_zetas.c
@@ -10,8 +10,7 @@
 
 #include "../../../common.h"
 
-#if defined(MLKEM_NATIVE_ARITH_BACKEND_AARCH64_CLEAN) || \
-    defined(MLKEM_NATIVE_ARITH_BACKEND_AARCH64_OPT)
+#if defined(MLKEM_NATIVE_ARITH_BACKEND_AARCH64_OPT)
 
 #include <stdint.h>
 #include "arith_native_aarch64.h"

diff --git a/src/kem/ml_kem/mlkem-native_ml-kem-1024_aarch64/native/aarch64/src/arith_native_aarch64.h b/src/kem/ml_kem/mlkem-native_ml-kem-1024_aarch64/native/aarch64/src/arith_native_aarch64.h
@@ -29,62 +29,49 @@ extern const int16_t aarch64_zetas_mulcache_native[];
 extern const int16_t aarch64_zetas_mulcache_twisted_native[];
 extern const uint8_t rej_uniform_table[];
 
-#define ntt_asm_clean MLKEM_NAMESPACE(ntt_asm_clean)
-void ntt_asm_clean(int16_t *, const int16_t *, const int16_t *);
-
 #define ntt_asm_opt MLKEM_NAMESPACE(ntt_asm_opt)
 void ntt_asm_opt(int16_t *, const int16_t *, const int16_t *);
 
-#define intt_asm_clean MLKEM_NAMESPACE(intt_asm_clean)
-void intt_asm_clean(int16_t *, const int16_t *, const int16_t *);
-
 #define intt_asm_opt MLKEM_NAMESPACE(intt_asm_opt)
 void intt_asm_opt(int16_t *, const int16_t *, const int16_t *);
 
-#define rej_uniform_asm_clean MLKEM_NAMESPACE(rej_uniform_asm_clean)
-unsigned rej_uniform_asm_clean(int16_t *r, const uint8_t *buf, unsigned buflen,
-                               const uint8_t *table);
-
-#define poly_reduce_asm_clean MLKEM_NAMESPACE(poly_reduce_asm_clean)
-void poly_reduce_asm_clean(int16_t *);
-
 #define poly_reduce_asm_opt MLKEM_NAMESPACE(poly_reduce_asm_opt)
 void poly_reduce_asm_opt(int16_t *);
 
-#define poly_tomont_asm_clean MLKEM_NAMESPACE(poly_tomont_asm_clean)
-void poly_tomont_asm_clean(int16_t *);
-
 #define poly_tomont_asm_opt MLKEM_NAMESPACE(poly_tomont_asm_opt)
 void poly_tomont_asm_opt(int16_t *);
 
-#define poly_mulcache_compute_asm_clean \
-  MLKEM_NAMESPACE(poly_mulcache_compute_asm_clean)
-void poly_mulcache_compute_asm_clean(int16_t *, const int16_t *,
-                                     const int16_t *, const int16_t *);
-
-
 #define poly_mulcache_compute_asm_opt \
   MLKEM_NAMESPACE(poly_mulcache_compute_asm_opt)
 void poly_mulcache_compute_asm_opt(int16_t *, const int16_t *, const int16_t *,
                                    const int16_t *);
 
-#define poly_tobytes_asm_clean MLKEM_NAMESPACE(poly_tobytes_asm_clean)
-void poly_tobytes_asm_clean(uint8_t *r, const int16_t *a);
-
 #define poly_tobytes_asm_opt MLKEM_NAMESPACE(poly_tobytes_asm_opt)
 void poly_tobytes_asm_opt(uint8_t *r, const int16_t *a);
 
-#define polyvec_basemul_acc_montgomery_cached_asm_clean \
-  MLKEM_NAMESPACE_K(polyvec_basemul_acc_montgomery_cached_asm_clean)
-void polyvec_basemul_acc_montgomery_cached_asm_clean(int16_t *r,
-                                                     const int16_t *a,
-                                                     const int16_t *b,
-                                                     const int16_t *b_cache);
+#define polyvec_basemul_acc_montgomery_cached_asm_k2_opt \
+  MLKEM_NAMESPACE(polyvec_basemul_acc_montgomery_cached_asm_k2_opt)
+void polyvec_basemul_acc_montgomery_cached_asm_k2_opt(int16_t *r,
+                                                      const int16_t *a,
+                                                      const int16_t *b,
+                                                      const int16_t *b_cache);
+
+#define polyvec_basemul_acc_montgomery_cached_asm_k3_opt \
+  MLKEM_NAMESPACE(polyvec_basemul_acc_montgomery_cached_asm_k3_opt)
+void polyvec_basemul_acc_montgomery_cached_asm_k3_opt(int16_t *r,
+                                                      const int16_t *a,
+                                                      const int16_t *b,
+                                                      const int16_t *b_cache);
+
+#define polyvec_basemul_acc_montgomery_cached_asm_k4_opt \
+  MLKEM_NAMESPACE(polyvec_basemul_acc_montgomery_cached_asm_k4_opt)
+void polyvec_basemul_acc_montgomery_cached_asm_k4_opt(int16_t *r,
+                                                      const int16_t *a,
+                                                      const int16_t *b,
+                                                      const int16_t *b_cache);
 
-#define polyvec_basemul_acc_montgomery_cached_asm_opt \
-  MLKEM_NAMESPACE_K(polyvec_basemul_acc_montgomery_cached_asm_opt)
-void polyvec_basemul_acc_montgomery_cached_asm_opt(int16_t *r, const int16_t *a,
-                                                   const int16_t *b,
-                                                   const int16_t *b_cache);
+#define rej_uniform_asm_clean MLKEM_NAMESPACE(rej_uniform_asm_clean)
+unsigned rej_uniform_asm_clean(int16_t *r, const uint8_t *buf, unsigned buflen,
+                               const uint8_t *table);
 
 #endif /* MLKEM_AARCH64_NATIVE_H */
diff --git a/src/kem/ml_kem/mlkem-native_ml-kem-1024_aarch64/native/aarch64/src/clean_impl.h b/src/kem/ml_kem/mlkem-native_ml-kem-1024_aarch64/native/aarch64/src/clean_impl.h