Move bn_mul_add_words() and bn_mul_words() from bn_asm.c to bn_mul.c.

These are wrapped with #ifndef HAVE_BN_ADD_MUL_WORDS/HAVE_BN_MUL_WORDS,
which are defined for architectures that provide their own assembly
versions.

Move bn_sqr_words from bn_asm.c to bn_sqr.c.

This is wrapped with #ifndef HAVE_BN_SQR_WORDS, which is then defined for
architectures that provide their own assembly versions.

Move bn_div_words from bn_asm.c to bn_div.c.

This is wrapped with #ifndef HAVE_BN_DIV_WORDS, which are defined for
architectures that provide their own assembly versions.

Move bn_add_words() and bn_sub_words from bn_asm.c to bn_add.c.

These are wrapped with #ifndef HAVE_BN_ADD_WORDS/HAVE_BN_SUB_WORDS, which
are defined for architectures that provide their own assembly versions.

Move bn_{mul,sqr}_comba{4,8}() from bn_asm.c to bn_mul.c/bn_sqr.c.

Wrap these in HAVE_BN_{MUL,SQR}_COMBA{4,8} defines. Add these defines to
bn_arch.h where the architecture currently provides its own version.

ok tb@

Provide a per machine bn_arch.h.

This will provide a location for machine specific defines, prototypes and
inline functions.

ok tb@

Move bn_{mul,sqr}_comba{4,8}() from bn_asm.c to bn_mul.c/bn_sqr.c.

Wrap these in HAVE_BN_{MUL,SQR}_COMBA{4,8} defines. Add these defines to
bn_arch.h where the architecture currently provides its own version.

ok tb@

Provide a per machine bn_arch.h.

This will provide a location for machine specific defines, prototypes and
inline functions.

ok tb@