diff --git a/library/chacha20.c b/library/chacha20.c
index 78467d3..9862ea5 100644
--- a/library/chacha20.c
+++ b/library/chacha20.c
@@ -54,13 +54,6 @@
 #define CHACHA20_VALIDATE( cond )                                           \
     MBEDTLS_INTERNAL_VALIDATE( cond )
 
-#define BYTES_TO_U32_LE( data, offset )                           \
-    ( (uint32_t) (data)[offset]                                   \
-      | (uint32_t) ( (uint32_t) (data)[( offset ) + 1] << 8 )     \
-      | (uint32_t) ( (uint32_t) (data)[( offset ) + 2] << 16 )    \
-      | (uint32_t) ( (uint32_t) (data)[( offset ) + 3] << 24 )    \
-    )
-
 #define ROTL32( value, amount ) \
     ( (uint32_t) ( (value) << (amount) ) | ( (value) >> ( 32 - (amount) ) ) )
 
diff --git a/library/common.h b/library/common.h
index 6a9f259..bf867ae 100644
--- a/library/common.h
+++ b/library/common.h
@@ -62,9 +62,19 @@
  * To tidy up code and save horizontal and vertical space, use byte 
  * reading macros to cast
  */
-#define BYTE_0( x ) ( (uint8_t) ( ( x ) & 0xff )  )
-#define BYTE_1( x ) ( (uint8_t) ( ( ( x ) >> 8 ) & 0xff )  )
+#define BYTE_0( x ) ( (uint8_t) ( ( x ) & 0xff ) )
+#define BYTE_1( x ) ( (uint8_t) ( ( ( x ) >> 8  ) & 0xff ) )
 #define BYTE_2( x ) ( (uint8_t) ( ( ( x ) >> 16 ) & 0xff ) )
 #define BYTE_3( x ) ( (uint8_t) ( ( ( x ) >> 24 ) & 0xff ) )
 
+/** 
+ * 
+ */
+#define BYTES_TO_U32_LE( data, offset )                         \
+    ( (uint32_t) (data)[offset]                                 \
+      | (uint32_t) ( (uint32_t) (data)[( offset ) + 1] << 8 )   \
+      | (uint32_t) ( (uint32_t) (data)[( offset ) + 2] << 16 )  \
+      | (uint32_t) ( (uint32_t) (data)[( offset ) + 3] << 24 )  \
+    )
+
 #endif /* MBEDTLS_LIBRARY_COMMON_H */
diff --git a/library/poly1305.c b/library/poly1305.c
index 492d145..a30b170 100644
--- a/library/poly1305.c
+++ b/library/poly1305.c
@@ -52,13 +52,6 @@
 
 #define POLY1305_BLOCK_SIZE_BYTES ( 16U )
 
-#define BYTES_TO_U32_LE( data, offset )                           \
-    ( (uint32_t) (data)[offset]                                     \
-          | (uint32_t) ( (uint32_t) (data)[( offset ) + 1] << 8 )   \
-          | (uint32_t) ( (uint32_t) (data)[( offset ) + 2] << 16 )  \
-          | (uint32_t) ( (uint32_t) (data)[( offset ) + 3] << 24 )  \
-    )
-
 /*
  * Our implementation is tuned for 32-bit platforms with a 64-bit multiplier.
  * However we provided an alternative for platforms without such a multiplier.