pq_crypto 0.6.3 → 0.6.5

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

@@ -13,6 +13,9 @@
   Description: AArch64 hybrid scalar/vector implementation of Keccak-f[1600] permutation for four sequential states
   Signature: void mlk_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
@@ -140,7 +143,7 @@ MLK_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:
+Lmlk_keccak_f1600_x4_v8a_scalar_hybrid_initial:
         eor x30, x24, x25
         eor x27, x9, x10
         eor v30.16b, v0.16b, v5.16b
@@ -523,7 +526,7 @@ 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:
+Lmlk_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
@@ -911,8 +914,8 @@ 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
+Lmlk_keccak_f1600_x4_v8a_scalar_hybrid_loop_end:
+        b.le Lmlk_keccak_f1600_x4_v8a_scalar_hybrid_loop
         ror x2, x2, #0x3d
         ror x3, x3, #0x27
         ror x4, x4, #0x36
@@ -938,7 +941,7 @@ 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 Lmlk_keccak_f1600_x4_v8a_scalar_hybrid_done
         mov x30, #0x1               // =1
         str x30, [sp, #0x20]
         ldr x0, [sp]
@@ -972,9 +975,9 @@ 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 Lmlk_keccak_f1600_x4_v8a_scalar_hybrid_initial
 
-Lkeccak_f1600_x4_v8a_scalar_hybrid_done:
+Lmlk_keccak_f1600_x4_v8a_scalar_hybrid_done:
         ldr x0, [sp]
         add x0, x0, #0x258
         stp x1, x6, [x0]

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

@@ -13,6 +13,9 @@
   Description: AArch64 hybrid scalar/vector implementation of Keccak-f[1600] permutation for four sequential states with ARMv8.4-A optimizations
   Signature: void mlk_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
@@ -142,7 +145,7 @@ MLK_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:
+Lmlk_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
@@ -478,7 +481,7 @@ 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:
+Lmlk_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
@@ -819,8 +822,8 @@ 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
+Lmlk_keccak_f1600_x4_v8a_v84a_scalar_hybrid_loop_end:
+        b.le Lmlk_keccak_f1600_x4_v8a_v84a_scalar_hybrid_loop
         ror x2, x2, #0x3d
         ror x3, x3, #0x27
         ror x4, x4, #0x36
@@ -846,7 +849,7 @@ 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 Lmlk_keccak_f1600_x4_v8a_v84a_scalar_hybrid_done
         mov x30, #0x1               // =1
         str x30, [sp, #0x20]
         ldr x0, [sp]
@@ -880,9 +883,9 @@ 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 Lmlk_keccak_f1600_x4_v8a_v84a_scalar_hybrid_initial
 
-Lkeccak_f1600_x4_v8a_v84a_scalar_hybrid_done:
+Lmlk_keccak_f1600_x4_v8a_v84a_scalar_hybrid_done:
         ldr x0, [sp]
         add x0, x0, #0x258
         stp x1, x6, [x0]

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

@@ -21,7 +21,8 @@
 MLK_MUST_CHECK_RETURN_VALUE
 static MLK_INLINE int mlk_keccak_f1600_x1_native(uint64_t *state)
 {
-  if (!mlk_sys_check_capability(MLK_SYS_CAP_SHA3))
+  if (!mlk_sys_check_capability(MLK_SYS_CAP_NEON) ||
+      !mlk_sys_check_capability(MLK_SYS_CAP_SHA3))
   {
     return MLK_NATIVE_FUNC_FALLBACK;
   }

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

@@ -21,7 +21,8 @@
 MLK_MUST_CHECK_RETURN_VALUE
 static MLK_INLINE int mlk_keccak_f1600_x4_native(uint64_t *state)
 {
-  if (!mlk_sys_check_capability(MLK_SYS_CAP_SHA3))
+  if (!mlk_sys_check_capability(MLK_SYS_CAP_NEON) ||
+      !mlk_sys_check_capability(MLK_SYS_CAP_SHA3))
   {
     return MLK_NATIVE_FUNC_FALLBACK;
   }

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

@@ -17,6 +17,11 @@
 MLK_MUST_CHECK_RETURN_VALUE
 static MLK_INLINE int mlk_keccak_f1600_x4_native(uint64_t *state)
 {
+  if (!mlk_sys_check_capability(MLK_SYS_CAP_NEON))
+  {
+    return MLK_NATIVE_FUNC_FALLBACK;
+  }
+
   mlk_keccak_f1600_x4_v8a_scalar_hybrid_aarch64_asm(
       state, mlk_keccakf1600_round_constants);
   return MLK_NATIVE_FUNC_SUCCESS;

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

@@ -21,7 +21,8 @@
 MLK_MUST_CHECK_RETURN_VALUE
 static MLK_INLINE int mlk_keccak_f1600_x4_native(uint64_t *state)
 {
-  if (!mlk_sys_check_capability(MLK_SYS_CAP_SHA3))
+  if (!mlk_sys_check_capability(MLK_SYS_CAP_NEON) ||
+      !mlk_sys_check_capability(MLK_SYS_CAP_SHA3))
   {
     return MLK_NATIVE_FUNC_FALLBACK;
   }

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

@@ -17,6 +17,7 @@
 
 #if !defined(__ASSEMBLER__)
 #include "../api.h"
+#include "src/fips202_native_armv81m.h"
 
 /*
  * Native x4 permutation
@@ -35,42 +36,27 @@ static MLK_INLINE int mlk_keccak_f1600_x4_native(uint64_t *state)
 /*
  * Native x4 XOR bytes (with on-the-fly bit interleaving)
  */
-#define mlk_keccak_f1600_x4_state_xor_bytes \
-  MLK_NAMESPACE(keccak_f1600_x4_state_xor_bytes_asm)
-void mlk_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);
-
 MLK_MUST_CHECK_RETURN_VALUE
 static MLK_INLINE int mlk_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)
 {
-  mlk_keccak_f1600_x4_state_xor_bytes(state, data0, data1, data2, data3, offset,
-                                      length);
+  mlk_keccak_f1600_x4_state_xor_bytes_asm(state, data0, data1, data2, data3,
+                                          offset, length);
   return MLK_NATIVE_FUNC_SUCCESS;
 }
 
 /*
  * Native x4 extract bytes (with on-the-fly bit de-interleaving)
  */
-#define mlk_keccak_f1600_x4_state_extract_bytes \
-  MLK_NAMESPACE(keccak_f1600_x4_state_extract_bytes_asm)
-void mlk_keccak_f1600_x4_state_extract_bytes(void *state, uint8_t *data0,
-                                             uint8_t *data1, uint8_t *data2,
-                                             uint8_t *data3, unsigned offset,
-                                             unsigned length);
-
 MLK_MUST_CHECK_RETURN_VALUE
 static MLK_INLINE int mlk_keccakf1600_extract_bytes_x4_native(
     uint64_t *state, uint8_t *data0, uint8_t *data1, uint8_t *data2,
     uint8_t *data3, unsigned offset, unsigned length)
 {
-  mlk_keccak_f1600_x4_state_extract_bytes(state, data0, data1, data2, data3,
-                                          offset, length);
+  mlk_keccak_f1600_x4_state_extract_bytes_asm(state, data0, data1, data2, data3,
+                                              offset, length);
   return MLK_NATIVE_FUNC_SUCCESS;
 }
 

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

@@ -9,6 +9,9 @@
   Description: Armv8.1-M MVE implementation of batched (x4) Keccak-f[1600] permutation using bit-interleaved state
   Signature: void mlk_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
@@ -111,9 +114,9 @@ MLK_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, Lmlk_keccak_f1600_x4_mve_asm_roundend @ imm = #0x8c0
 
-Lkeccak_f1600_x4_mve_asm_roundstart:
+Lmlk_keccak_f1600_x4_mve_asm_roundstart:
         vldrw.u32 q6, [r2, #112]
         veor q7, q6, q2
         vldrw.u32 q2, [r2, #80]
@@ -674,10 +677,10 @@ 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
+Lmlk_keccak_f1600_x4_mve_asm_roundend_pre:
+        le lr, Lmlk_keccak_f1600_x4_mve_asm_roundstart @ imm = #-0x8c0
 
-Lkeccak_f1600_x4_mve_asm_roundend:
+Lmlk_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/mlkem-native/mlkem/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 mlkem-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 @@ MLK_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 Lmlk_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 Lmlk_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 @@ MLK_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 Lmlk_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 Lmlk_keccak_f1600_x4_state_extract_bytes_asm_main_body @ imm = #0xe
 
-Lkeccak_f1600_x4_state_extract_bytes_asm_pre_main:
+Lmlk_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:
+Lmlk_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, Lmlk_keccak_f1600_x4_state_extract_bytes_asm_main_loop_end @ imm = #0xb4
 
-Lkeccak_f1600_x4_state_extract_bytes_asm_main_loop_start:
+Lmlk_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, Lmlk_keccak_f1600_x4_state_extract_bytes_asm_main_loop_start @ imm = #-0xb4
 
-Lkeccak_f1600_x4_state_extract_bytes_asm_main_loop_end:
+Lmlk_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 Lmlk_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:
+Lmlk_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/mlkem-native/mlkem/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 mlkem-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 @@ MLK_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 Lmlk_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 Lmlk_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 @@ MLK_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 Lmlk_keccak_f1600_x4_state_xor_bytes_asm_main_body @ imm = #0xe
 
-Lkeccak_f1600_x4_state_xor_bytes_asm_pre_main:
+Lmlk_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:
+Lmlk_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, Lmlk_keccak_f1600_x4_state_xor_bytes_asm_main_loop_end @ imm = #0xd4
 
-Lkeccak_f1600_x4_state_xor_bytes_asm_main_loop_start:
+Lmlk_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, Lmlk_keccak_f1600_x4_state_xor_bytes_asm_main_loop_start @ imm = #-0xd4
 
-Lkeccak_f1600_x4_state_xor_bytes_asm_main_loop_end:
+Lmlk_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 Lmlk_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:
+Lmlk_keccak_f1600_x4_state_xor_bytes_asm_exit:
         vpop {d8, d9, d10, d11, d12, d13, d14, d15}
         .cfi_restore d8
         .cfi_restore d9

data/ext/pqcrypto/vendor/mlkem-native/mlkem/src/fips202/native/x86_64/src/keccak_f1600_x4_avx2_asm.S

@@ -6,6 +6,40 @@
 
 #if defined(MLK_FIPS202_X86_64_NEED_X4_AVX2) && \
     !defined(MLK_CONFIG_MULTILEVEL_NO_SHARED)
+/*yaml
+  Name: keccak_f1600_x4_avx2_asm
+  Description: x86_64 AVX2 Keccak-f[1600] permutation for four sequential states
+  Signature: void mlk_keccak_f1600_x4_avx2_asm(uint64_t states[100], const uint64_t rc[24], const uint64_t rho8[4], const uint64_t rho56[4])
+  ABI:
+    Architecture: x86_64
+    CallingConvention: SysV
+    Features: [AVX2]
+    rdi:
+      type: buffer
+      size_bytes: 800
+      permissions: read/write
+      c_parameter: uint64_t states[100]
+      description: Four sequential Keccak states (4 x 25 x uint64_t)
+    rsi:
+      type: buffer
+      size_bytes: 192
+      permissions: read-only
+      c_parameter: const uint64_t rc[24]
+      description: Round constants (24 x uint64_t)
+    rdx:
+      type: buffer
+      size_bytes: 32
+      permissions: read-only
+      c_parameter: const uint64_t rho8[4]
+      description: Rotation constant rho8 (4 x uint64_t)
+    rcx:
+      type: buffer
+      size_bytes: 32
+      permissions: read-only
+      c_parameter: const uint64_t rho56[4]
+      description: Rotation constant rho56 (4 x uint64_t)
+*/
+
 
 /*
  * WARNING: This file is auto-derived from the mlkem-native source file
@@ -119,7 +153,7 @@ MLK_ASM_FN_SYMBOL(keccak_f1600_x4_avx2_asm)
         vinserti128 $0x1, %xmm0, %ymm1, %ymm2
         movq $0x0, %r10
 
-LLkeccak_f1600_x4_avx2:
+LLmlk_keccak_f1600_x4_avx2:
         vmovdqu 0xa0(%rsp), %ymm4
         vpxor 0x1c0(%rsp), %ymm9, %ymm0
         vmovdqu %ymm9, 0x200(%rsp)
@@ -343,7 +377,7 @@ LLkeccak_f1600_x4_avx2:
         addq $0x8, %rsi
         addq $0x1, %r10
         cmpq $0x18, %r10
-        jne LLkeccak_f1600_x4_avx2
+        jne LLmlk_keccak_f1600_x4_avx2
         vmovdqu (%rsp), %ymm4
         vmovdqu 0x40(%rsp), %ymm5
         vmovdqu 0x20(%rsp), %ymm0

data/ext/pqcrypto/vendor/mlkem-native/mlkem/src/indcpa.c

@@ -55,6 +55,7 @@
  *                  [0,..,MLKEM_Q-1].
  * @param[in]  seed Input public seed.
  */
+#if !defined(MLK_CONFIG_NO_KEYPAIR_API)
 static void mlk_pack_pk(uint8_t r[MLKEM_INDCPA_PUBLICKEYBYTES],
                         const mlk_polyvec *pk,
                         const uint8_t seed[MLKEM_SYMBYTES])
@@ -63,6 +64,7 @@ static void mlk_pack_pk(uint8_t r[MLKEM_INDCPA_PUBLICKEYBYTES],
   mlk_polyvec_tobytes(r, pk);
   mlk_memcpy(r + MLKEM_POLYVECBYTES, seed, MLKEM_SYMBYTES);
 }
+#endif /* !MLK_CONFIG_NO_KEYPAIR_API */
 
 /**
  * De-serialize public key from a byte array; approximate inverse of
@@ -75,6 +77,7 @@ static void mlk_pack_pk(uint8_t r[MLKEM_INDCPA_PUBLICKEYBYTES],
  * @param[out] seed     Output seed to generate matrix A.
  * @param[in]  packedpk Input serialized public key.
  */
+#if !defined(MLK_CONFIG_NO_ENCAPS_API) || !defined(MLK_CONFIG_NO_DECAPS_API)
 static void mlk_unpack_pk(mlk_polyvec *pk, uint8_t seed[MLKEM_SYMBYTES],
                           const uint8_t packedpk[MLKEM_INDCPA_PUBLICKEYBYTES])
 {
@@ -86,6 +89,7 @@ static void mlk_unpack_pk(mlk_polyvec *pk, uint8_t seed[MLKEM_SYMBYTES],
    * specifications and proofs, however, do _not_ assume this, and instead
    * work with the easily provable bound by MLKEM_UINT12_LIMIT. */
 }
+#endif /* !MLK_CONFIG_NO_ENCAPS_API || !MLK_CONFIG_NO_DECAPS_API */
 
 /**
  * Serialize the secret key.
@@ -95,12 +99,14 @@ static void mlk_unpack_pk(mlk_polyvec *pk, uint8_t seed[MLKEM_SYMBYTES],
  * @param[out] r  Output serialized secret key.
  * @param[in]  sk Input vector of polynomials (secret key).
  */
+#if !defined(MLK_CONFIG_NO_KEYPAIR_API)
 static void mlk_pack_sk(uint8_t r[MLKEM_INDCPA_SECRETKEYBYTES],
                         const mlk_polyvec *sk)
 {
   mlk_assert_bound_2d(sk->vec, MLKEM_K, MLKEM_N, 0, MLKEM_Q);
   mlk_polyvec_tobytes(r, sk);
 }
+#endif /* !MLK_CONFIG_NO_KEYPAIR_API */
 
 /**
  * De-serialize the secret key; inverse of mlk_pack_sk.
@@ -110,11 +116,13 @@ static void mlk_pack_sk(uint8_t r[MLKEM_INDCPA_SECRETKEYBYTES],
  * @param[out] sk       Output vector of polynomials (secret key).
  * @param[in]  packedsk Input serialized secret key.
  */
+#if !defined(MLK_CONFIG_NO_DECAPS_API)
 static void mlk_unpack_sk(mlk_polyvec *sk,
                           const uint8_t packedsk[MLKEM_INDCPA_SECRETKEYBYTES])
 {
   mlk_polyvec_frombytes(sk, packedsk);
 }
+#endif /* !MLK_CONFIG_NO_DECAPS_API */
 
 /**
  * Serialize the ciphertext as the concatenation of the compressed and
@@ -127,12 +135,14 @@ static void mlk_unpack_sk(mlk_polyvec *sk,
  * @param[in]  b Input vector of polynomials b.
  * @param[in]  v Input polynomial v.
  */
+#if !defined(MLK_CONFIG_NO_ENCAPS_API) || !defined(MLK_CONFIG_NO_DECAPS_API)
 static void mlk_pack_ciphertext(uint8_t r[MLKEM_INDCPA_BYTES],
                                 const mlk_polyvec *b, mlk_poly *v)
 {
   mlk_polyvec_compress_du(r, b);
   mlk_poly_compress_dv(r + MLKEM_POLYVECCOMPRESSEDBYTES_DU, v);
 }
+#endif /* !MLK_CONFIG_NO_ENCAPS_API || !MLK_CONFIG_NO_DECAPS_API */
 
 /**
  * De-serialize and decompress ciphertext from a byte array; approximate
@@ -144,12 +154,14 @@ static void mlk_pack_ciphertext(uint8_t r[MLKEM_INDCPA_BYTES],
  * @param[out] v Output polynomial v.
  * @param[in]  c Input serialized ciphertext.
  */
+#if !defined(MLK_CONFIG_NO_DECAPS_API)
 static void mlk_unpack_ciphertext(mlk_polyvec *b, mlk_poly *v,
                                   const uint8_t c[MLKEM_INDCPA_BYTES])
 {
   mlk_polyvec_decompress_du(b, c);
   mlk_poly_decompress_dv(v, c + MLKEM_POLYVECCOMPRESSEDBYTES_DU);
 }
+#endif /* !MLK_CONFIG_NO_DECAPS_API */
 
 /* Helper function to ensure that the polynomial entries in the output
  * of gen_matrix use the standard (bitreversed) ordering of coefficients.
@@ -345,6 +357,7 @@ __contract__(
  * @param[out] e    Output polynomial vector.
  * @param[in]  seed Seed bytes for sampling.
  */
+#if !defined(MLK_CONFIG_NO_KEYPAIR_API)
 static void mlk_keypair_getnoise_eta1(mlk_polyvec *pv, mlk_polyvec *e,
                                       const uint8_t seed[MLKEM_SYMBYTES])
 __contract__(
@@ -378,6 +391,7 @@ __contract__(
                             seed, 4, 5, 6, 7);
 #endif /* MLKEM_K == 4 */
 }
+#endif /* !MLK_CONFIG_NO_KEYPAIR_API */
 
 /**
  * Compute and fill the sp, ep, and epp polynomial structures needed by
@@ -391,6 +405,7 @@ __contract__(
  * @param[out] epp   Output polynomial.
  * @param[in]  coins Seed bytes for sampling.
  */
+#if !defined(MLK_CONFIG_NO_ENCAPS_API) || !defined(MLK_CONFIG_NO_DECAPS_API)
 static void mlk_enc_getnoise_eta1_eta2(mlk_polyvec *sp, mlk_polyvec *ep,
                                        mlk_poly *epp,
                                        const uint8_t coins[MLKEM_SYMBYTES])
@@ -429,6 +444,7 @@ __contract__(
   mlk_poly_getnoise_eta2(epp, coins, 8);
 #endif /* MLKEM_K == 4 */
 }
+#endif /* !MLK_CONFIG_NO_ENCAPS_API || !MLK_CONFIG_NO_DECAPS_API */
 
 
 /* Reference: `indcpa_keypair_derand()` in the reference implementation @[REF].
@@ -437,6 +453,7 @@ __contract__(
  *            - We use a mulcache to speed up matrix-vector multiplication.
  *            - We include buffer zeroization.
  */
+#if !defined(MLK_CONFIG_NO_KEYPAIR_API)
 MLK_INTERNAL_API
 int mlk_indcpa_keypair_derand(uint8_t pk[MLKEM_INDCPA_PUBLICKEYBYTES],
                               uint8_t sk[MLKEM_INDCPA_SECRETKEYBYTES],
@@ -508,6 +525,7 @@ cleanup:
   MLK_FREE(buf, uint8_t, 2 * MLKEM_SYMBYTES, context);
   return ret;
 }
+#endif /* !MLK_CONFIG_NO_KEYPAIR_API */
 
 /* Reference: `indcpa_enc()` in the reference implementation @[REF].
  *            - We use x4-batched versions of `poly_getnoise` to leverage
@@ -517,6 +535,7 @@ cleanup:
  *            - We use a mulcache to speed up matrix-vector multiplication.
  *            - We include buffer zeroization.
  */
+#if !defined(MLK_CONFIG_NO_ENCAPS_API) || !defined(MLK_CONFIG_NO_DECAPS_API)
 MLK_INTERNAL_API
 int mlk_indcpa_enc(uint8_t c[MLKEM_INDCPA_BYTES],
                    const uint8_t m[MLKEM_INDCPA_MSGBYTES],
@@ -591,10 +610,12 @@ cleanup:
   MLK_FREE(seed, uint8_t, MLKEM_SYMBYTES, context);
   return ret;
 }
+#endif /* !MLK_CONFIG_NO_ENCAPS_API || !MLK_CONFIG_NO_DECAPS_API */
 
 /* Reference: `indcpa_dec()` in the reference implementation @[REF].
  *            - We use a mulcache for the scalar product.
  *            - We include buffer zeroization. */
+#if !defined(MLK_CONFIG_NO_DECAPS_API)
 MLK_INTERNAL_API
 int mlk_indcpa_dec(uint8_t m[MLKEM_INDCPA_MSGBYTES],
                    const uint8_t c[MLKEM_INDCPA_BYTES],
@@ -637,6 +658,7 @@ cleanup:
   MLK_FREE(b, mlk_polyvec, 1, context);
   return ret;
 }
+#endif /* !MLK_CONFIG_NO_DECAPS_API */
 
 /* To facilitate single-compilation-unit (SCU) builds, undefine all macros.
  * Don't modify by hand -- this is auto-generated by scripts/autogen. */

data/ext/pqcrypto/vendor/mlkem-native/mlkem/src/indcpa.h

@@ -45,6 +45,7 @@ __contract__(
   array_bound(a->vec[x].vec[y].coeffs, 0, MLKEM_N, 0, MLKEM_Q))))
 );
 
+#if !defined(MLK_CONFIG_NO_KEYPAIR_API)
 #define mlk_indcpa_keypair_derand \
   MLK_NAMESPACE_K(indcpa_keypair_derand) MLK_CONTEXT_PARAMETERS_3
 /**
@@ -84,7 +85,9 @@ __contract__(
           return_value == MLK_ERR_OUT_OF_MEMORY ||
           return_value == MLK_ERR_RNG_FAIL)
 );
+#endif /* !MLK_CONFIG_NO_KEYPAIR_API */
 
+#if !defined(MLK_CONFIG_NO_ENCAPS_API) || !defined(MLK_CONFIG_NO_DECAPS_API)
 #define mlk_indcpa_enc MLK_NAMESPACE_K(indcpa_enc) MLK_CONTEXT_PARAMETERS_4
 /**
  * Encryption function of the CPA-secure public-key encryption scheme
@@ -123,7 +126,9 @@ __contract__(
   ensures(return_value == 0 || return_value == MLK_ERR_FAIL ||
           return_value == MLK_ERR_OUT_OF_MEMORY)
 );
+#endif /* !MLK_CONFIG_NO_ENCAPS_API || !MLK_CONFIG_NO_DECAPS_API */
 
+#if !defined(MLK_CONFIG_NO_DECAPS_API)
 #define mlk_indcpa_dec MLK_NAMESPACE_K(indcpa_dec) MLK_CONTEXT_PARAMETERS_3
 /**
  * Decryption function of the CPA-secure public-key encryption scheme
@@ -159,5 +164,6 @@ __contract__(
   ensures(return_value == 0 || return_value == MLK_ERR_FAIL ||
           return_value == MLK_ERR_OUT_OF_MEMORY)
 );
+#endif /* !MLK_CONFIG_NO_DECAPS_API */
 
 #endif /* !MLK_INDCPA_H */