pq_crypto 0.6.4 → 0.6.6

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/aarch64/src/keccak_f1600_x1_v84a_aarch64_asm.S

@@ -16,10 +16,13 @@
  */
 
 /*yaml
-  Name: keccak_f1600_x1_v84a_asm
+  Name: keccak_f1600_x1_v84a_aarch64_asm
   Description: AArch64 ARMv8.4-A implementation of Keccak-f[1600] permutation for single state
   Signature: void mld_keccak_f1600_x1_v84a_aarch64_asm(uint64_t state[25], const uint64_t rc[24])
   ABI:
+    Architecture: aarch64
+    CallingConvention: AAPCS64
+    Features: [NEON, SHA3]
     x0:
       type: buffer
       size_bytes: 200
@@ -30,7 +33,7 @@
       type: buffer
       size_bytes: 192
       permissions: read-only
-      c_parameter: const uint64_t rc[24]
+      c_parameter: uint64_t const *rc
       description: Round constants (24 x uint64_t)
   Stack:
     bytes: 64
@@ -92,7 +95,7 @@ MLD_ASM_FN_SYMBOL(keccak_f1600_x1_v84a_aarch64_asm)
         ldr d24, [x0, #0xc0]
         mov x2, #0x18               // =24
 
-Lkeccak_f1600_x1_v84a_loop:
+Lmld_keccak_f1600_x1_v84a_loop:
         eor3 v30.16b, v0.16b, v5.16b, v10.16b
         eor3 v29.16b, v1.16b, v6.16b, v11.16b
         eor3 v28.16b, v2.16b, v7.16b, v12.16b
@@ -161,7 +164,7 @@ Lkeccak_f1600_x1_v84a_loop:
         bcax v4.16b, v4.16b, v27.16b, v30.16b
         eor v0.16b, v0.16b, v31.16b
         sub x2, x2, #0x1
-        cbnz x2, Lkeccak_f1600_x1_v84a_loop
+        cbnz x2, Lmld_keccak_f1600_x1_v84a_loop
         stp d0, d1, [x0]
         stp d2, d3, [x0, #0x10]
         stp d4, d5, [x0, #0x20]

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/aarch64/src/keccak_f1600_x2_v84a_aarch64_asm.S

@@ -16,10 +16,13 @@
  */
 
 /*yaml
-  Name: keccak_f1600_x2_v84a_asm
+  Name: keccak_f1600_x2_v84a_aarch64_asm
   Description: AArch64 ARMv8.4-A implementation of Keccak-f[1600] permutation for two sequential states
   Signature: void mld_keccak_f1600_x2_v84a_aarch64_asm(uint64_t state[50], const uint64_t rc[24])
   ABI:
+    Architecture: aarch64
+    CallingConvention: AAPCS64
+    Features: [NEON, SHA3]
     x0:
       type: buffer
       size_bytes: 400
@@ -30,7 +33,7 @@
       type: buffer
       size_bytes: 192
       permissions: read-only
-      c_parameter: const uint64_t rc[24]
+      c_parameter: uint64_t const *rc
       description: Round constants (24 x uint64_t)
   Stack:
     bytes: 64
@@ -119,7 +122,7 @@ MLD_ASM_FN_SYMBOL(keccak_f1600_x2_v84a_aarch64_asm)
         trn1 v24.2d, v25.2d, v27.2d
         mov x2, #0x18               // =24
 
-Lkeccak_f1600_x2_v84a_loop:
+Lmld_keccak_f1600_x2_v84a_loop:
         eor3 v30.16b, v0.16b, v5.16b, v10.16b
         eor3 v29.16b, v1.16b, v6.16b, v11.16b
         eor3 v28.16b, v2.16b, v7.16b, v12.16b
@@ -188,7 +191,7 @@ Lkeccak_f1600_x2_v84a_loop:
         bcax v4.16b, v4.16b, v27.16b, v30.16b
         eor v0.16b, v0.16b, v31.16b
         sub x2, x2, #0x1
-        cbnz x2, Lkeccak_f1600_x2_v84a_loop
+        cbnz x2, Lmld_keccak_f1600_x2_v84a_loop
         sub x0, x0, #0xc0
         add x2, x0, #0xc8
         trn1 v25.2d, v0.2d, v1.2d

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/aarch64/src/keccak_f1600_x4_v8a_scalar_hybrid_aarch64_asm.S

@@ -10,10 +10,13 @@
 // Author: Matthias Kannwischer <matthias@kannwischer.eu>
 
 /*yaml
-  Name: keccak_f1600_x4_v8a_scalar_hybrid_asm
+  Name: keccak_f1600_x4_v8a_scalar_hybrid_aarch64_asm
   Description: AArch64 hybrid scalar/vector implementation of Keccak-f[1600] permutation for four sequential states
   Signature: void mld_keccak_f1600_x4_v8a_scalar_hybrid_aarch64_asm(uint64_t state[100], const uint64_t rc[24])
   ABI:
+    Architecture: aarch64
+    CallingConvention: AAPCS64
+    Features: [NEON]
     x0:
       type: buffer
       size_bytes: 800
@@ -24,7 +27,7 @@
       type: buffer
       size_bytes: 192
       permissions: read-only
-      c_parameter: const uint64_t rc[24]
+      c_parameter: uint64_t const *rc
       description: Round constants (24 x uint64_t)
   Stack:
     bytes: 224
@@ -141,7 +144,7 @@ MLD_ASM_FN_SYMBOL(keccak_f1600_x4_v8a_scalar_hybrid_aarch64_asm)
         ldr x25, [x0, #0xc0]
         sub x0, x0, #0x190
 
-Lkeccak_f1600_x4_v8a_scalar_hybrid_initial:
+Lmld_keccak_f1600_x4_v8a_scalar_hybrid_initial:
         eor x30, x24, x25
         eor x27, x9, x10
         eor v30.16b, v0.16b, v5.16b
@@ -524,7 +527,7 @@ Lkeccak_f1600_x4_v8a_scalar_hybrid_initial:
         str x30, [sp, #0x10]
         eor v0.16b, v0.16b, v28.16b
 
-Lkeccak_f1600_x4_v8a_scalar_hybrid_loop:
+Lmld_keccak_f1600_x4_v8a_scalar_hybrid_loop:
         eor x0, x15, x11, ror #52
         eor x0, x0, x13, ror #48
         eor v30.16b, v0.16b, v5.16b
@@ -912,8 +915,8 @@ Lkeccak_f1600_x4_v8a_scalar_hybrid_loop:
         str x30, [sp, #0x10]
         eor v0.16b, v0.16b, v28.16b
 
-Lkeccak_f1600_x4_v8a_scalar_hybrid_loop_end:
-        b.le Lkeccak_f1600_x4_v8a_scalar_hybrid_loop
+Lmld_keccak_f1600_x4_v8a_scalar_hybrid_loop_end:
+        b.le Lmld_keccak_f1600_x4_v8a_scalar_hybrid_loop
         ror x2, x2, #0x3d
         ror x3, x3, #0x27
         ror x4, x4, #0x36
@@ -939,7 +942,7 @@ Lkeccak_f1600_x4_v8a_scalar_hybrid_loop_end:
         ror x25, x25, #0x9
         ldr x30, [sp, #0x20]
         cmp x30, #0x1
-        b.eq Lkeccak_f1600_x4_v8a_scalar_hybrid_done
+        b.eq Lmld_keccak_f1600_x4_v8a_scalar_hybrid_done
         mov x30, #0x1               // =1
         str x30, [sp, #0x20]
         ldr x0, [sp]
@@ -973,9 +976,9 @@ Lkeccak_f1600_x4_v8a_scalar_hybrid_loop_end:
         ldp x15, x20, [x0, #0xb0]
         ldr x25, [x0, #0xc0]
         sub x0, x0, #0x258
-        b Lkeccak_f1600_x4_v8a_scalar_hybrid_initial
+        b Lmld_keccak_f1600_x4_v8a_scalar_hybrid_initial
 
-Lkeccak_f1600_x4_v8a_scalar_hybrid_done:
+Lmld_keccak_f1600_x4_v8a_scalar_hybrid_done:
         ldr x0, [sp]
         add x0, x0, #0x258
         stp x1, x6, [x0]

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/aarch64/src/keccak_f1600_x4_v8a_v84a_scalar_hybrid_aarch64_asm.S

@@ -10,10 +10,13 @@
 // Author: Matthias Kannwischer <matthias@kannwischer.eu>
 
 /*yaml
-  Name: keccak_f1600_x4_v8a_v84a_scalar_hybrid_asm
+  Name: keccak_f1600_x4_v8a_v84a_scalar_hybrid_aarch64_asm
   Description: AArch64 hybrid scalar/vector implementation of Keccak-f[1600] permutation for four sequential states with ARMv8.4-A optimizations
   Signature: void mld_keccak_f1600_x4_v8a_v84a_scalar_hybrid_aarch64_asm(uint64_t state[100], const uint64_t rc[24])
   ABI:
+    Architecture: aarch64
+    CallingConvention: AAPCS64
+    Features: [NEON, SHA3]
     x0:
       type: buffer
       size_bytes: 800
@@ -24,7 +27,7 @@
       type: buffer
       size_bytes: 192
       permissions: read-only
-      c_parameter: const uint64_t rc[24]
+      c_parameter: uint64_t const *rc
       description: Round constants (24 x uint64_t)
   Stack:
     bytes: 224
@@ -143,7 +146,7 @@ MLD_ASM_FN_SYMBOL(keccak_f1600_x4_v8a_v84a_scalar_hybrid_aarch64_asm)
         ldr x25, [x0, #0xc0]
         sub x0, x0, #0x190
 
-Lkeccak_f1600_x4_v8a_v84a_scalar_hybrid_initial:
+Lmld_keccak_f1600_x4_v8a_v84a_scalar_hybrid_initial:
         eor x30, x24, x25
         eor x27, x9, x10
         eor3 v30.16b, v0.16b, v5.16b, v10.16b
@@ -479,7 +482,7 @@ Lkeccak_f1600_x4_v8a_v84a_scalar_hybrid_initial:
         str x30, [sp, #0x10]
         eor v0.16b, v0.16b, v28.16b
 
-Lkeccak_f1600_x4_v8a_v84a_scalar_hybrid_loop:
+Lmld_keccak_f1600_x4_v8a_v84a_scalar_hybrid_loop:
         eor x0, x15, x11, ror #52
         eor x0, x0, x13, ror #48
         eor3 v30.16b, v0.16b, v5.16b, v10.16b
@@ -820,8 +823,8 @@ Lkeccak_f1600_x4_v8a_v84a_scalar_hybrid_loop:
         str x30, [sp, #0x10]
         eor v0.16b, v0.16b, v28.16b
 
-Lkeccak_f1600_x4_v8a_v84a_scalar_hybrid_loop_end:
-        b.le Lkeccak_f1600_x4_v8a_v84a_scalar_hybrid_loop
+Lmld_keccak_f1600_x4_v8a_v84a_scalar_hybrid_loop_end:
+        b.le Lmld_keccak_f1600_x4_v8a_v84a_scalar_hybrid_loop
         ror x2, x2, #0x3d
         ror x3, x3, #0x27
         ror x4, x4, #0x36
@@ -847,7 +850,7 @@ Lkeccak_f1600_x4_v8a_v84a_scalar_hybrid_loop_end:
         ror x25, x25, #0x9
         ldr x30, [sp, #0x20]
         cmp x30, #0x1
-        b.eq Lkeccak_f1600_x4_v8a_v84a_scalar_hybrid_done
+        b.eq Lmld_keccak_f1600_x4_v8a_v84a_scalar_hybrid_done
         mov x30, #0x1               // =1
         str x30, [sp, #0x20]
         ldr x0, [sp]
@@ -881,9 +884,9 @@ Lkeccak_f1600_x4_v8a_v84a_scalar_hybrid_loop_end:
         ldp x15, x20, [x0, #0xb0]
         ldr x25, [x0, #0xc0]
         sub x0, x0, #0x258
-        b Lkeccak_f1600_x4_v8a_v84a_scalar_hybrid_initial
+        b Lmld_keccak_f1600_x4_v8a_v84a_scalar_hybrid_initial
 
-Lkeccak_f1600_x4_v8a_v84a_scalar_hybrid_done:
+Lmld_keccak_f1600_x4_v8a_v84a_scalar_hybrid_done:
         ldr x0, [sp]
         add x0, x0, #0x258
         stp x1, x6, [x0]

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/aarch64/x1_scalar.h

@@ -8,7 +8,7 @@
 #define MLD_FIPS202_NATIVE_AARCH64_X1_SCALAR_H
 
 /* Part of backend API */
-#define MLD_USE_FIPS202_X1_NATIVE
+#define MLD_USE_NATIVE_FIPS202_X1
 /* Guard for assembly file */
 #define MLD_FIPS202_AARCH64_NEED_X1_SCALAR
 

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/aarch64/x1_v84a.h

@@ -12,7 +12,7 @@
 #endif
 
 /* Part of backend API */
-#define MLD_USE_FIPS202_X1_NATIVE
+#define MLD_USE_NATIVE_FIPS202_X1
 /* Guard for assembly file */
 #define MLD_FIPS202_AARCH64_NEED_X1_V84A
 
@@ -22,7 +22,8 @@
 MLD_MUST_CHECK_RETURN_VALUE
 static MLD_INLINE int mld_keccak_f1600_x1_native(uint64_t *state)
 {
-  if (!mld_sys_check_capability(MLD_SYS_CAP_SHA3))
+  if (!mld_sys_check_capability(MLD_SYS_CAP_AARCH64_NEON) ||
+      !mld_sys_check_capability(MLD_SYS_CAP_AARCH64_SHA3))
   {
     return MLD_NATIVE_FUNC_FALLBACK;
   }

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/aarch64/x2_v84a.h

@@ -12,7 +12,7 @@
 #endif
 
 /* Part of backend API */
-#define MLD_USE_FIPS202_X4_NATIVE
+#define MLD_USE_NATIVE_FIPS202_X4
 /* Guard for assembly file */
 #define MLD_FIPS202_AARCH64_NEED_X2_V84A
 
@@ -23,7 +23,8 @@
 MLD_MUST_CHECK_RETURN_VALUE
 static MLD_INLINE int mld_keccak_f1600_x4_native(uint64_t *state)
 {
-  if (!mld_sys_check_capability(MLD_SYS_CAP_SHA3))
+  if (!mld_sys_check_capability(MLD_SYS_CAP_AARCH64_NEON) ||
+      !mld_sys_check_capability(MLD_SYS_CAP_AARCH64_SHA3))
   {
     return MLD_NATIVE_FUNC_FALLBACK;
   }

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/aarch64/x4_v8a_scalar.h

@@ -8,7 +8,7 @@
 #define MLD_FIPS202_NATIVE_AARCH64_X4_V8A_SCALAR_H
 
 /* Part of backend API */
-#define MLD_USE_FIPS202_X4_NATIVE
+#define MLD_USE_NATIVE_FIPS202_X4
 /* Guard for assembly file */
 #define MLD_FIPS202_AARCH64_NEED_X4_V8A_SCALAR_HYBRID
 
@@ -18,6 +18,11 @@
 MLD_MUST_CHECK_RETURN_VALUE
 static MLD_INLINE int mld_keccak_f1600_x4_native(uint64_t *state)
 {
+  if (!mld_sys_check_capability(MLD_SYS_CAP_AARCH64_NEON))
+  {
+    return MLD_NATIVE_FUNC_FALLBACK;
+  }
+
   mld_keccak_f1600_x4_v8a_scalar_hybrid_aarch64_asm(
       state, mld_keccakf1600_round_constants);
   return MLD_NATIVE_FUNC_SUCCESS;

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/aarch64/x4_v8a_v84a_scalar.h

@@ -12,7 +12,7 @@
 #endif
 
 /* Part of backend API */
-#define MLD_USE_FIPS202_X4_NATIVE
+#define MLD_USE_NATIVE_FIPS202_X4
 /* Guard for assembly file */
 #define MLD_FIPS202_AARCH64_NEED_X4_V8A_V84A_SCALAR_HYBRID
 
@@ -22,7 +22,8 @@
 MLD_MUST_CHECK_RETURN_VALUE
 static MLD_INLINE int mld_keccak_f1600_x4_native(uint64_t *state)
 {
-  if (!mld_sys_check_capability(MLD_SYS_CAP_SHA3))
+  if (!mld_sys_check_capability(MLD_SYS_CAP_AARCH64_NEON) ||
+      !mld_sys_check_capability(MLD_SYS_CAP_AARCH64_SHA3))
   {
     return MLD_NATIVE_FUNC_FALLBACK;
   }

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/api.h

@@ -39,13 +39,13 @@
  * A _backend_ is a specific implementation of parts of this interface.
  *
  * You can replace 1-fold or 4-fold batched Keccak-F1600.
- * To enable, set MLD_USE_FIPS202_X1_NATIVE or MLD_USE_FIPS202_X4_NATIVE
+ * To enable, set MLD_USE_NATIVE_FIPS202_X1 or MLD_USE_NATIVE_FIPS202_X4
  * in your backend, and define the inline wrappers mld_keccak_f1600_x1_native()
  * and/or mld_keccak_f1600_x4_native(), respectively, to forward to your
  * implementation.
  */
 
-#if defined(MLD_USE_FIPS202_X1_NATIVE)
+#if defined(MLD_USE_NATIVE_FIPS202_X1)
 MLD_MUST_CHECK_RETURN_VALUE
 static MLD_INLINE int mld_keccak_f1600_x1_native(uint64_t *state)
 __contract__(
@@ -54,8 +54,8 @@ __contract__(
     ensures(return_value == MLD_NATIVE_FUNC_FALLBACK || return_value == MLD_NATIVE_FUNC_SUCCESS)
     ensures((return_value == MLD_NATIVE_FUNC_FALLBACK) ==> array_unchanged_u64(state, 25 * 1))
 );
-#endif /* MLD_USE_FIPS202_X1_NATIVE */
-#if defined(MLD_USE_FIPS202_X4_NATIVE)
+#endif /* MLD_USE_NATIVE_FIPS202_X1 */
+#if defined(MLD_USE_NATIVE_FIPS202_X4)
 MLD_MUST_CHECK_RETURN_VALUE
 static MLD_INLINE int mld_keccak_f1600_x4_native(uint64_t *state)
 __contract__(
@@ -64,7 +64,7 @@ __contract__(
     ensures(return_value == MLD_NATIVE_FUNC_FALLBACK || return_value == MLD_NATIVE_FUNC_SUCCESS)
     ensures((return_value == MLD_NATIVE_FUNC_FALLBACK) ==> array_unchanged_u64(state, 25 * 4))
 );
-#endif /* MLD_USE_FIPS202_X4_NATIVE */
+#endif /* MLD_USE_NATIVE_FIPS202_X4 */
 
 /*
  * Native x4 XOR bytes and extract bytes interface.
@@ -78,12 +78,12 @@ __contract__(
  * NOTE: We assume that the custom representation of the zero state is the
  * all-zero state.
  *
- * MLD_USE_FIPS202_X4_XOR_BYTES_NATIVE: Backend provides native XOR bytes
- * MLD_USE_FIPS202_X4_EXTRACT_BYTES_NATIVE: Backend provides native extract
+ * MLD_USE_NATIVE_FIPS202_X4_XOR_BYTES: Backend provides native XOR bytes
+ * MLD_USE_NATIVE_FIPS202_X4_EXTRACT_BYTES: Backend provides native extract
  * bytes
  */
 
-#if defined(MLD_USE_FIPS202_X4_XOR_BYTES_NATIVE)
+#if defined(MLD_USE_NATIVE_FIPS202_X4_XOR_BYTES)
 MLD_MUST_CHECK_RETURN_VALUE
 static MLD_INLINE int mld_keccakf1600_xor_bytes_x4_native(
     uint64_t *state, const unsigned char *data0, const unsigned char *data1,
@@ -103,9 +103,9 @@ __contract__(
   assigns(memory_slice(state, sizeof(uint64_t) * 25 * 4))
   ensures(return_value == MLD_NATIVE_FUNC_FALLBACK || return_value == MLD_NATIVE_FUNC_SUCCESS)
   ensures((return_value == MLD_NATIVE_FUNC_FALLBACK) ==> array_unchanged_u64(state, 25 * 4)));
-#endif /* MLD_USE_FIPS202_X4_XOR_BYTES_NATIVE */
+#endif /* MLD_USE_NATIVE_FIPS202_X4_XOR_BYTES */
 
-#if defined(MLD_USE_FIPS202_X4_EXTRACT_BYTES_NATIVE)
+#if defined(MLD_USE_NATIVE_FIPS202_X4_EXTRACT_BYTES)
 MLD_MUST_CHECK_RETURN_VALUE
 static MLD_INLINE int mld_keccakf1600_extract_bytes_x4_native(
     uint64_t *state, unsigned char *data0, unsigned char *data1,
@@ -124,6 +124,6 @@ __contract__(
   assigns(memory_slice(data2, length))
   assigns(memory_slice(data3, length))
   ensures(return_value == MLD_NATIVE_FUNC_FALLBACK || return_value == MLD_NATIVE_FUNC_SUCCESS));
-#endif /* MLD_USE_FIPS202_X4_EXTRACT_BYTES_NATIVE */
+#endif /* MLD_USE_NATIVE_FIPS202_X4_EXTRACT_BYTES */
 
 #endif /* !MLD_FIPS202_NATIVE_API_H */

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/armv81m/mve.h

@@ -10,14 +10,15 @@
 #define MLD_FIPS202_NATIVE_ARMV81M
 
 /* Part of backend API */
-#define MLD_USE_FIPS202_X4_NATIVE
-#define MLD_USE_FIPS202_X4_XOR_BYTES_NATIVE
-#define MLD_USE_FIPS202_X4_EXTRACT_BYTES_NATIVE
+#define MLD_USE_NATIVE_FIPS202_X4
+#define MLD_USE_NATIVE_FIPS202_X4_XOR_BYTES
+#define MLD_USE_NATIVE_FIPS202_X4_EXTRACT_BYTES
 /* Guard for assembly file */
 #define MLD_FIPS202_ARMV81M_NEED_X4
 
 #if !defined(__ASSEMBLER__)
 #include "../api.h"
+#include "src/fips202_native_armv81m.h"
 
 /*
  * Native x4 permutation
@@ -25,6 +26,7 @@
  */
 #define mld_keccak_f1600_x4_native_impl \
   MLD_NAMESPACE(keccak_f1600_x4_native_impl)
+MLD_INTERNAL_API
 int mld_keccak_f1600_x4_native_impl(uint64_t *state);
 
 MLD_MUST_CHECK_RETURN_VALUE
@@ -36,42 +38,27 @@ static MLD_INLINE int mld_keccak_f1600_x4_native(uint64_t *state)
 /*
  * Native x4 XOR bytes (with on-the-fly bit interleaving)
  */
-#define mld_keccak_f1600_x4_state_xor_bytes \
-  MLD_NAMESPACE(keccak_f1600_x4_state_xor_bytes_asm)
-void mld_keccak_f1600_x4_state_xor_bytes(void *state, const uint8_t *data0,
-                                         const uint8_t *data1,
-                                         const uint8_t *data2,
-                                         const uint8_t *data3, unsigned offset,
-                                         unsigned length);
-
 MLD_MUST_CHECK_RETURN_VALUE
 static MLD_INLINE int mld_keccakf1600_xor_bytes_x4_native(
     uint64_t *state, const uint8_t *data0, const uint8_t *data1,
     const uint8_t *data2, const uint8_t *data3, unsigned offset,
     unsigned length)
 {
-  mld_keccak_f1600_x4_state_xor_bytes(state, data0, data1, data2, data3, offset,
-                                      length);
+  mld_keccak_f1600_x4_state_xor_bytes_asm(state, data0, data1, data2, data3,
+                                          offset, length);
   return MLD_NATIVE_FUNC_SUCCESS;
 }
 
 /*
  * Native x4 extract bytes (with on-the-fly bit de-interleaving)
  */
-#define mld_keccak_f1600_x4_state_extract_bytes \
-  MLD_NAMESPACE(keccak_f1600_x4_state_extract_bytes_asm)
-void mld_keccak_f1600_x4_state_extract_bytes(void *state, uint8_t *data0,
-                                             uint8_t *data1, uint8_t *data2,
-                                             uint8_t *data3, unsigned offset,
-                                             unsigned length);
-
 MLD_MUST_CHECK_RETURN_VALUE
 static MLD_INLINE int mld_keccakf1600_extract_bytes_x4_native(
     uint64_t *state, uint8_t *data0, uint8_t *data1, uint8_t *data2,
     uint8_t *data3, unsigned offset, unsigned length)
 {
-  mld_keccak_f1600_x4_state_extract_bytes(state, data0, data1, data2, data3,
-                                          offset, length);
+  mld_keccak_f1600_x4_state_extract_bytes_asm(state, data0, data1, data2, data3,
+                                              offset, length);
   return MLD_NATIVE_FUNC_SUCCESS;
 }
 

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/armv81m/src/keccak_f1600_x4_mve.S

@@ -10,6 +10,9 @@
   Description: Armv8.1-M MVE implementation of batched (x4) Keccak-f[1600] permutation using bit-interleaved state
   Signature: void mld_keccak_f1600_x4_mve_asm(void *state, void *tmpstate, const uint32_t *rc)
   ABI:
+    Architecture: armv81m
+    CallingConvention: AAPCS32
+    Features: [MVE]
     r0:
       type: buffer
       size_bytes: 800
@@ -112,9 +115,9 @@ MLD_ASM_FN_SYMBOL(keccak_f1600_x4_mve_asm)
         vldrw.u32 q0, [r3]
         vldrw.u32 q1, [r2]
         vldrw.u32 q2, [r2, #32]
-        wls lr, lr, Lkeccak_f1600_x4_mve_asm_roundend @ imm = #0x8c0
+        wls lr, lr, Lmld_keccak_f1600_x4_mve_asm_roundend @ imm = #0x8c0
 
-Lkeccak_f1600_x4_mve_asm_roundstart:
+Lmld_keccak_f1600_x4_mve_asm_roundstart:
         vldrw.u32 q6, [r2, #112]
         veor q7, q6, q2
         vldrw.u32 q2, [r2, #80]
@@ -675,10 +678,10 @@ Lkeccak_f1600_x4_mve_asm_roundstart:
         veor q0, q4, q6
         vstrw.32 q0, [r5]
 
-Lkeccak_f1600_x4_mve_asm_roundend_pre:
-        le lr, Lkeccak_f1600_x4_mve_asm_roundstart @ imm = #-0x8c0
+Lmld_keccak_f1600_x4_mve_asm_roundend_pre:
+        le lr, Lmld_keccak_f1600_x4_mve_asm_roundstart @ imm = #-0x8c0
 
-Lkeccak_f1600_x4_mve_asm_roundend:
+Lmld_keccak_f1600_x4_mve_asm_roundend:
         add sp, #0x80
         .cfi_adjust_cfa_offset -0x80
         vpop {d8, d9, d10, d11, d12, d13, d14, d15}

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/armv81m/src/keccak_f1600_x4_mve.c

@@ -19,6 +19,7 @@
  */
 #define mld_keccak_f1600_x4_native_impl \
   MLD_NAMESPACE(keccak_f1600_x4_native_impl)
+MLD_INTERNAL_API
 int mld_keccak_f1600_x4_native_impl(uint64_t *state)
 {
   MLD_ALIGN uint64_t state_tmp[100];

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/armv81m/src/{state_extract_bytes_x4_mve.S → keccak_f1600_x4_state_extract_bytes_mve.S}

@@ -9,7 +9,7 @@
 // Overview
 // ---------------------------------------------------------------------------
 // MVE/Helium implementation of KeccakF1600x4_StateExtractBytes
-// (inverse of state_xor_bytes_x4_mve.S).
+// (inverse of keccak_f1600_x4_state_xor_bytes_mve.S).
 //
 // void KeccakF1600x4_StateExtractBytes(state, d0, d1, d2, d3, offset, length)
 //
@@ -62,7 +62,7 @@
 
 /*
  * WARNING: This file is auto-derived from the mldsa-native source file
- *   dev/fips202/armv81m/src/state_extract_bytes_x4_mve.S using scripts/simpasm. Do not modify it directly.
+ *   dev/fips202/armv81m/src/keccak_f1600_x4_state_extract_bytes_mve.S using scripts/simpasm. Do not modify it directly.
  */
 
 .thumb
@@ -99,13 +99,13 @@ MLD_ASM_FN_SYMBOL(keccak_f1600_x4_state_extract_bytes_asm)
         ldr.w r10, [sp, #0x6c]
         ldr r6, [sp, #0x70]
         cmp r6, #0x0
-        beq.w Lkeccak_f1600_x4_state_extract_bytes_asm_exit @ imm = #0x2ea
+        beq.w Lmld_keccak_f1600_x4_state_extract_bytes_asm_exit @ imm = #0x2ea
         and r5, r10, #0x7
         bic r9, r10, #0x7
         add.w r8, r0, r9, lsl #1
         add.w r7, r8, #0x190
         cmp r5, #0x0
-        beq.w Lkeccak_f1600_x4_state_extract_bytes_asm_pre_main @ imm = #0x112
+        beq.w Lmld_keccak_f1600_x4_state_extract_bytes_asm_pre_main @ imm = #0x112
         vldrw.u32 q0, [r8], #16
         vldrw.u32 q1, [r7], #16
         vrev32.16 q2, q0
@@ -175,22 +175,22 @@ MLD_ASM_FN_SYMBOL(keccak_f1600_x4_state_extract_bytes_asm)
         vstrbt.8 q3, [r4], #4
         subs.w r6, r6, lr
         cmp r6, #0x0
-        beq.w Lkeccak_f1600_x4_state_extract_bytes_asm_exit @ imm = #0x1cc
+        beq.w Lmld_keccak_f1600_x4_state_extract_bytes_asm_exit @ imm = #0x1cc
         vmov q7[2], q7[0], r1, r3
         vmov q7[3], q7[1], r2, r4
-        b Lkeccak_f1600_x4_state_extract_bytes_asm_main_body @ imm = #0xe
+        b Lmld_keccak_f1600_x4_state_extract_bytes_asm_main_body @ imm = #0xe
 
-Lkeccak_f1600_x4_state_extract_bytes_asm_pre_main:
+Lmld_keccak_f1600_x4_state_extract_bytes_asm_pre_main:
         vmov q7[2], q7[0], r1, r3
         vmov q7[3], q7[1], r2, r4
         mov.w r12, #0x4
         vsub.i32 q7, q7, r12
 
-Lkeccak_f1600_x4_state_extract_bytes_asm_main_body:
+Lmld_keccak_f1600_x4_state_extract_bytes_asm_main_body:
         lsr.w lr, r6, #0x3
-        wls lr, lr, Lkeccak_f1600_x4_state_extract_bytes_asm_main_loop_end @ imm = #0xb4
+        wls lr, lr, Lmld_keccak_f1600_x4_state_extract_bytes_asm_main_loop_end @ imm = #0xb4
 
-Lkeccak_f1600_x4_state_extract_bytes_asm_main_loop_start:
+Lmld_keccak_f1600_x4_state_extract_bytes_asm_main_loop_start:
         vldrw.u32 q0, [r8], #16
         vldrw.u32 q1, [r7], #16
         vrev32.16 q2, q0
@@ -235,11 +235,11 @@ Lkeccak_f1600_x4_state_extract_bytes_asm_main_loop_start:
         vorr q1, q1, q3
         vstrw.32 q0, [q7, #4]!
         vstrw.32 q1, [q7, #4]!
-        le lr, Lkeccak_f1600_x4_state_extract_bytes_asm_main_loop_start @ imm = #-0xb4
+        le lr, Lmld_keccak_f1600_x4_state_extract_bytes_asm_main_loop_start @ imm = #-0xb4
 
-Lkeccak_f1600_x4_state_extract_bytes_asm_main_loop_end:
+Lmld_keccak_f1600_x4_state_extract_bytes_asm_main_loop_end:
         ands r6, r6, #0x7
-        beq Lkeccak_f1600_x4_state_extract_bytes_asm_exit @ imm = #0xee
+        beq Lmld_keccak_f1600_x4_state_extract_bytes_asm_exit @ imm = #0xee
         mov.w r12, #0x4
         vadd.i32 q7, q7, r12
         vmov r1, r3, q7[2], q7[0]
@@ -301,7 +301,7 @@ Lkeccak_f1600_x4_state_extract_bytes_asm_main_loop_end:
         vstrbt.8 q2, [r3], #4
         vstrbt.8 q3, [r4], #4
 
-Lkeccak_f1600_x4_state_extract_bytes_asm_exit:
+Lmld_keccak_f1600_x4_state_extract_bytes_asm_exit:
         vpop {d8, d9, d10, d11, d12, d13, d14, d15}
         .cfi_restore d8
         .cfi_restore d9

data/ext/pqcrypto/vendor/mldsa-native/mldsa/src/fips202/native/armv81m/src/{state_xor_bytes_x4_mve.S → keccak_f1600_x4_state_xor_bytes_mve.S}

@@ -61,7 +61,7 @@
 
 /*
  * WARNING: This file is auto-derived from the mldsa-native source file
- *   dev/fips202/armv81m/src/state_xor_bytes_x4_mve.S using scripts/simpasm. Do not modify it directly.
+ *   dev/fips202/armv81m/src/keccak_f1600_x4_state_xor_bytes_mve.S using scripts/simpasm. Do not modify it directly.
  */
 
 .thumb
@@ -98,13 +98,13 @@ MLD_ASM_FN_SYMBOL(keccak_f1600_x4_state_xor_bytes_asm)
         ldr.w r10, [sp, #0x6c]
         ldr r6, [sp, #0x70]
         cmp r6, #0x0
-        beq.w Lkeccak_f1600_x4_state_xor_bytes_asm_exit @ imm = #0x346
+        beq.w Lmld_keccak_f1600_x4_state_xor_bytes_asm_exit @ imm = #0x346
         and r5, r10, #0x7
         bic r9, r10, #0x7
         add.w r8, r0, r9, lsl #1
         add.w r7, r8, #0x190
         cmp r5, #0x0
-        beq.w Lkeccak_f1600_x4_state_xor_bytes_asm_pre_main @ imm = #0x12c
+        beq.w Lmld_keccak_f1600_x4_state_xor_bytes_asm_pre_main @ imm = #0x12c
         subs r1, r1, r5
         subs r2, r2, r5
         subs r3, r3, r5
@@ -183,19 +183,19 @@ MLD_ASM_FN_SYMBOL(keccak_f1600_x4_state_xor_bytes_asm)
         vstrw.32 q5, [r7], #16
         vmov q7[2], q7[0], r1, r3
         vmov q7[3], q7[1], r2, r4
-        b Lkeccak_f1600_x4_state_xor_bytes_asm_main_body @ imm = #0xe
+        b Lmld_keccak_f1600_x4_state_xor_bytes_asm_main_body @ imm = #0xe
 
-Lkeccak_f1600_x4_state_xor_bytes_asm_pre_main:
+Lmld_keccak_f1600_x4_state_xor_bytes_asm_pre_main:
         vmov q7[2], q7[0], r1, r3
         vmov q7[3], q7[1], r2, r4
         mov.w r0, #0x4
         vsub.i32 q7, q7, r0
 
-Lkeccak_f1600_x4_state_xor_bytes_asm_main_body:
+Lmld_keccak_f1600_x4_state_xor_bytes_asm_main_body:
         lsr.w lr, r6, #0x3
-        wls lr, lr, Lkeccak_f1600_x4_state_xor_bytes_asm_main_loop_end @ imm = #0xd4
+        wls lr, lr, Lmld_keccak_f1600_x4_state_xor_bytes_asm_main_loop_end @ imm = #0xd4
 
-Lkeccak_f1600_x4_state_xor_bytes_asm_main_loop_start:
+Lmld_keccak_f1600_x4_state_xor_bytes_asm_main_loop_start:
         vldrw.u32 q0, [q7, #4]!
         vldrw.u32 q1, [q7, #4]!
         vmov q2, q0
@@ -248,11 +248,11 @@ Lkeccak_f1600_x4_state_xor_bytes_asm_main_loop_start:
         veor q5, q5, q1
         vstrw.32 q4, [r8], #16
         vstrw.32 q5, [r7], #16
-        le lr, Lkeccak_f1600_x4_state_xor_bytes_asm_main_loop_start @ imm = #-0xd4
+        le lr, Lmld_keccak_f1600_x4_state_xor_bytes_asm_main_loop_start @ imm = #-0xd4
 
-Lkeccak_f1600_x4_state_xor_bytes_asm_main_loop_end:
+Lmld_keccak_f1600_x4_state_xor_bytes_asm_main_loop_end:
         ands r6, r6, #0x7
-        beq.w Lkeccak_f1600_x4_state_xor_bytes_asm_exit @ imm = #0x110
+        beq.w Lmld_keccak_f1600_x4_state_xor_bytes_asm_exit @ imm = #0x110
         mov.w r0, #0x4
         vadd.i32 q7, q7, r0
         vmov r1, r3, q7[2], q7[0]
@@ -322,7 +322,7 @@ Lkeccak_f1600_x4_state_xor_bytes_asm_main_loop_end:
         vstrw.32 q4, [r8], #16
         vstrw.32 q5, [r7], #16
 
-Lkeccak_f1600_x4_state_xor_bytes_asm_exit:
+Lmld_keccak_f1600_x4_state_xor_bytes_asm_exit:
         vpop {d8, d9, d10, d11, d12, d13, d14, d15}
         .cfi_restore d8
         .cfi_restore d9