Add the extension parser in the s_client, ocsp and x509 apps.

This can be used to speed up SRTP with libsrtp, e.g. on TI omap/sitara based devices.

Add the RFC6962 OIDs to the objects table.

Improve CBC decrypt and CTR by ~13/16%, which adds up to ~25/33%
improvement over "pre-Silvermont" version. [Add performance table to
aesni-x86.pl].

When a chain is complete and ends in a trusted root checks are also
performed on the TA and the callback notified with ok==1. For
consistency do the same for chains where the TA is not self signed.

(other processors unaffected).

Atom Silvermont. On other CPUs one can observe 1% loss on some
algorithms.

+5% on Atom Silvermont, up to +8% improvement of legacy code.
Harmonize sha1-586.pl and aesni-sha1-x86_86.p with sha1-x86_64.pl.

Always add a dynamically loaded ENGINE to list. Otherwise it can cause
problems when multiply loaded, especially if it adds new public key methods.
For all current engines we only want a single implementation anyway.

If available rdrand is used as an additional entropy source for the
PRNG and for additional input in FIPS mode.

(and shave off cycle even from integer-only code)

(and update performance data, and fix typo)

but keep it disabled, too little gain... Add some Atom-specific
optimization.

PR: 3202

If content is detached and not binary mode translate the input to
CRLF format. Before this change the input was verified verbatim
which lead to a discrepancy between sign and verify.

(cherry picked from commit a6c62f0c25a756c263a80ce52afbae888028e986)

(and ensure stack alignment in the process)

It worked because it was never called.

SHA512_Transform was initially added rather as tribute to tradition
than for practucal reasons. But use was recently found in ssl/s3_cbc.c
and it turned to be problematic on platforms that don't tolerate
misasligned references to memory and lack assembly subroutine.