ec: 56-bit Limb Solinas' Strategy for secp384r1

ec: 56-bit Limb Solinas' Strategy for secp384r1

Adopt a 56-bit redundant-limb Solinas' reduction approach for efficient
modular multiplication in P384. This has the affect of accelerating
digital signing by 446% and verification by 106%. The implementation
strategy and names of methods are the same as that provided in
ecp_nistp224 and ecp_nistp521.

As in Commit 1036749883cc ("ec: Add run time code selection for p521
field operations"), allow for run time selection of implementation for
felem_{square,mul}, where an assembly implementation is proclaimed to
be present when ECP_NISTP384_ASM is present.

Signed-off-by: Rohan McLure <rohanmclure@linux.ibm.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/21471)