Thumb-2 ChaCha, Poly1305: implemention in assembly

Implementation of ChaCha algorithm for ARM Thumb-2.
Implementation of Poly1305 algorithm for ARM Thumb-2.

Author: SparkiDev

src/include.am

@@ -922,6 +922,12 @@ if !BUILD_FIPS_RAND
 if BUILD_POLY1305
 if BUILD_ARMASM
 src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/port/arm/armv8-poly1305.c
+src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/port/arm/thumb2-poly1305.c
+if BUILD_ARMASM_INLINE
+src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/port/arm/thumb2-poly1305-asm_c.c
+else
+src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/port/arm/thumb2-poly1305-asm.S
+endif !BUILD_ARMASM_INLINE
 endif
 if BUILD_RISCV_ASM
 src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/port/riscv/riscv-64-poly1305.c

wolfcrypt/src/poly1305.c

@@ -231,7 +231,8 @@ extern void poly1305_final_avx2(Poly1305* ctx, byte* mac);
         p[7] = (byte)(v >> 56);
     }
 #endif/* !WOLFSSL_ARMASM && !WOLFSSL_RISCV_ASM */
-#else /* if not 64 bit then use 32 bit */
+/* if not 64 bit then use 32 bit */
+#elif !defined(WOLFSSL_ARMASM) || !defined(__thumb__)
 
     static word32 U8TO32(const byte *p)
     {
@@ -268,8 +269,8 @@ static WC_INLINE void u32tole64(const word32 inLe32, byte outLe64[8])
 }
 
 
-#if (!defined(WOLFSSL_ARMASM) || !defined(__aarch64__)) && \
-    !defined(WOLFSSL_RISCV_ASM)
+#if (!defined(WOLFSSL_ARMASM) || (!defined(__aarch64__) && \
+    !defined(__thumb__))) && !defined(WOLFSSL_RISCV_ASM)
 /*
 This local function operates on a message with a given number of bytes
 with a given ctx pointer to a Poly1305 structure.
@@ -788,7 +789,8 @@ int wc_Poly1305Final(Poly1305* ctx, byte* mac)
 
     return 0;
 }
-#endif /* (!WOLFSSL_ARMASM || !__aarch64__) && !WOLFSSL_RISCV_ASM */
+#endif /* (!WOLFSSL_ARMASM || (!__aarch64__ && !__thumb__)) &&
+        * !WOLFSSL_RISCV_ASM */
 
 
 int wc_Poly1305Update(Poly1305* ctx, const byte* m, word32 bytes)
@@ -883,8 +885,8 @@ int wc_Poly1305Update(Poly1305* ctx, const byte* m, word32 bytes)
         /* process full blocks */
         if (bytes >= POLY1305_BLOCK_SIZE) {
             size_t want = ((size_t)bytes & ~((size_t)POLY1305_BLOCK_SIZE - 1));
-#if (!defined(WOLFSSL_ARMASM) || !defined(__aarch64__)) && \
-    !defined(WOLFSSL_RISCV_ASM)
+#if (!defined(WOLFSSL_ARMASM) || (!defined(__aarch64__) && \
+    !defined(__thumb__))) && !defined(WOLFSSL_RISCV_ASM)
             int ret;
             ret = poly1305_blocks(ctx, m, want);
             if (ret != 0)

wolfcrypt/src/port/arm/thumb2-chacha.c

@@ -49,8 +49,6 @@
 #endif
 
 
-extern void wc_chacha_setiv(word32* x, const byte* iv, word32 counter);
-
 /* Set the Initialization Vector (IV) and counter into ChaCha context.
  *
  * Set up iv(nonce). Earlier versions used 64 bits instead of 96, this version
@@ -92,8 +90,6 @@ int wc_Chacha_SetIV(ChaCha* ctx, const byte* iv, word32 counter)
     return ret;
 }
 
-extern void wc_chacha_setkey(word32* x, const byte* key, word32 keySz);
-
 /* Set the key into the ChaCha context.
  *
  * Key setup. 8 word iv (nonce)
@@ -141,11 +137,6 @@ int wc_Chacha_SetKey(ChaCha* ctx, const byte* key, word32 keySz)
     return ret;
 }
 
-extern void wc_chacha_use_over(byte* over, byte* output, const byte* input,
-    word32 len);
-extern void wc_chacha_crypt_bytes(ChaCha* ctx, byte* c, const byte* m,
-    word32 len);
-
 /* API to encrypt/decrypt a message of any size.
  *
  * @param [in]  ctx     ChaCha context.