ARM assembler modules.

- introduce common modes_lcl.h;
- ctr128.c: implement additional CRYPTO_ctr128_encrypt_ctr32 interface;
- gcm128.c: add omitted ARM initialization, remove ctx.ctr;

rely on indirect call to block functions, they are not as fast as
non-consolidated routines. However, performance loss(*) is within
measurement error and consolidation advantages are considered to
outweigh it.

(*) actually one can observe performance *improvement* on e.g.
    CBC benchmarks thanks to optimization, which also becomes
    shared among ciphers.

have a uniform representation for those over all architectures, so a
little bit of hackery is needed.

Contributed by nagendra modadugu <nagendra@cs.stanford.edu>

[most importantly] put back dependencies accidentaly eliminated in
check-in #13342.

during "make errors" and thus during "make update".

Fix lots of bugs that util/ck_errf.pl can detect automatically.
Various others of these are still left to fix; that's why
"make update" will complain loudly when run now.

if $(EXHEADER) is empty.

Notified by many, solution suggested by Carson Gaspar <carson@taltos.org>

This is part of a large change submitted by Markus Friedl <markus@openbsd.org>

give it.
For 0.9.7 and up, that means util/domd needs to remove those double
dashes from the argument list when gcc is used to find the
dependencies.

handled properly.
Part of PR 75

errors can be tolerated, hide the error from 'make'.
This gives shorter output both if ranlib fails and if
it works.