improper access of sp_int_minimal using sp_int

related thread: https://bugzilla.redhat.com/show_bug.cgi?id=2047439

also found sp_uint8 that was not updated to sp_sign_t

Author: sebastian-carpenter

wolfcrypt/src/sp_int.c

@@ -5102,7 +5102,7 @@ static void _sp_init_size(sp_int* a, unsigned int size)
 #endif
     _sp_zero((sp_int*)am);
 
-    a->size = (sp_size_t)size;
+    am->size = (sp_size_t)size;
 }
 
 /* Initialize the multi-precision number to be zero with a given max size.

wolfssl/wolfcrypt/sp_int.h

@@ -932,7 +932,7 @@ typedef struct sp_int_minimal {
     sp_size_t    size;
 #ifdef WOLFSSL_SP_INT_NEGATIVE
     /** Indicates whether number is 0/positive or negative.  */
-    sp_uint8     sign;
+    sp_sign_t    sign;
 #endif
 #ifdef HAVE_WOLF_BIGINT
     /** Unsigned binary (big endian) representation of number. */