+#if !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM) \
+ && !defined(PEDANTIC) && !defined(BN_DIV3W)
+# if defined(__GNUC__) && __GNUC__>=2
+# if defined(__i386) || defined (__i386__)
+ /*
+ * There were two reasons for implementing this template:
+ * - GNU C generates a call to a function (__udivdi3 to be exact)
+ * in reply to ((((BN_ULLONG)n0)<<BN_BITS2)|n1)/d0 (I fail to
+ * understand why...);
+ * - divl doesn't only calculate quotient, but also leaves
+ * remainder in %edx which we can definitely use here:-)
+ *
+ * <appro@fy.chalmers.se>
+ */
+# define bn_div_words(n0,n1,d0) \
+ ({ asm volatile ( \
+ "divl %4" \
+ : "=a"(q), "=d"(rem) \
+ : "a"(n1), "d"(n0), "g"(d0) \
+ : "cc"); \
+ q; \
+ })
+# define REMAINDER_IS_ALREADY_CALCULATED
+# elif defined(__x86_64) && defined(SIXTY_FOUR_BIT_LONG)
+ /*
+ * Same story here, but it's 128-bit by 64-bit division. Wow!
+ * <appro@fy.chalmers.se>
+ */
+# define bn_div_words(n0,n1,d0) \
+ ({ asm volatile ( \
+ "divq %4" \
+ : "=a"(q), "=d"(rem) \
+ : "a"(n1), "d"(n0), "g"(d0) \
+ : "cc"); \
+ q; \
+ })
+# define REMAINDER_IS_ALREADY_CALCULATED
+# endif /* __<cpu> */
+# endif /* __GNUC__ */
+#endif /* OPENSSL_NO_ASM */
+
+
+/* BN_div computes dv := num / divisor, rounding towards zero, and sets up
+ * rm such that dv*divisor + rm = num holds.
+ * Thus:
+ * dv->neg == num->neg ^ divisor->neg (unless the result is zero)
+ * rm->neg == num->neg (unless the remainder is zero)
+ * If 'dv' or 'rm' is NULL, the respective value is not returned.
+ */