WIP: Made changes to be able to build gost-engine as static library. Added gost-engine.h to be able to load engine as static engine for openssl.
Made gost_core library independent.
Corrected openssl version in cirrus.yml
Related issue: #340

* backport commit 2dd3a2f2

 from master

update magma cipher ctr_acpkm mode encrypting

Fixed bug when acpkm key meshing didn't apply at appropriate time
during TLS secure exchange.
Unify usage of 'num' variable of EVP_CIPHER_CTX for kuznetchik and
magma.

* correct define variable dependency

add optional compiling setting for EVP_CTRL_TLS1_2_TLSTREE variable

Co-authored-by: Igor Kirillov <i.kirillov@kryptonite.ru>

If at the input of the encryption function in the CFB mode we have
an integer number of blocks, then in the main loop all blocks will be
processed, except for the last one due to an incorrect border check.
The last block will be fully processed as a "partial" remainder, but
the initialization vector will not be updated. And, thus, the value
of IV will always be incorrect in this case.

This breaks stateless protocols due to an invalid initialization vector:
all messages except the first cannot be decrypted. (Naturally, we are
talking about a case with disabled key meshing, which does not allow
context recovery due to an erroneous implementation.)

It is worth noting here that the code for processing partial blocks
(both at the input of the encryption functions and at the output) is
a historically unnecessary artifact, since we do not set the
EVP_CIPH_FLAG_CUSTOM_CIPHER flag and, as a result, OpenSSL processes
partial blocks for us.

This patch corrects the checking of the main loop boundary.

A similar error is present in the code for the CNT mode, but there it
does not manifest itself in any way, because the restoration of the
state in this mode is impossible: even after disabling key meshing, we
still have the state-dependent transformation of the IV.

As an extra result of this fix, the code for processing partial blocks
can be completely removed now.

(cherry picked from commit cf402dd4)

(cherry picked from commit e1afd2a1)

/root/rpmbuild/BUILD/openssl-gost-engine-1.1.1/CMakeFiles/CMakeTmp/src.c:4:14: warning: initialization of 'int *' from incompatible pointer type 'char *' [-Wincompatible-pointer-types]
    4 |     int *p = buf + 1;
      |              ^~~
/root/rpmbuild/BUILD/openssl-gost-engine-1.1.1/CMakeFiles/CMakeTmp/src.c:5:14: warning: initialization of 'int *' from incompatible pointer type 'char *' [-Wincompatible-pointer-types]
    5 |     int *q = buf + 2;
      |              ^~~

Reported-by: Ilya Shipitsin <https://github.com/chipitsine>
Fixes: #288
(cherry picked from commit 6c7addf7)

(cherry picked from commit 47be42da)

(cherry picked from commit 1997dd99)

This is based on AX_CHECK_ALIGNED_ACCESS_REQUIRED from autoconf-archive.

Note, that on some arches unaligned access behavior could be changed at
runtime via prctl(1). Also, unaligned memory access is still slower (and
very slow on some arches) even if it's not strictly required.

(cherry picked from commit d2810d23)

(cherry picked from commit f1f47c6c)

ECCKiila Changes from

https://gitlab.com/nisec/ecckiila/-/issues/2

https://gitlab.com/nisec/ecckiila/-/commit/7445ecabef77965743e0ae8d39d7433b07820be6

* X3 -> X1
* eliminate a temp variable

X3 = X1 as pointers is the reason this was still passing unit tests.
But that might not hold in future versions of ECCKiila, so fix it now.

(cherry picked from commit 259301a5)

char defaults to signed on x86/x64, but unsigned on ARM.

(cherry picked from commit 409a1c2b)

Standalone EC implementations from ECCKiila.

https://gitlab.com/nisec/ecckiila
(cherry picked from commit bc346202)

* GOST key agreement cofactor fix

(cherry picked from commit dbc8f478)

* Tests that gost-engine correctly computes the public key from the
private key. (Twice -- Alice and Bob.)

* Tests that gost-engine correctly computes the derived shared key.
(Twice -- Alice and Bob, should be identical.)

(cherry picked from commit 95dd55b1)

Support of pkeyutl -pkeyopt ukmhex:0102030405060708 syntax

There need to be more tests than 2 constant fixed size blocks.
Final digest values are generated from the first run.

Test description in comments.

Plus some minor chnages
- function defenition doen't match function declaration
- Consecutive break statement is unnecessary.
- Type missmatch in format string

Only stack buffers are considered.

`RAND_priv_bytes' is supposed to be used for private data.

OpenSSL suggests to use (and internally itself uses)
`BN_{CTX_,}secure_new' primitives to work with private keys.

These are using `OPENSSL_secure_malloc' et al. calls, which use
special 'secure heap' memory.

Along, optimize out `hashsum2bn' with `BN_lebin2bn'.

`Test2::V0' is non standard and rarely present on systems, making `make
test' unnecessarily fail.

lcc emulates x86_64 SIMD intrinsics, so SSE2 code will work as is. Even
though in benchmarks it is slightly slower than regular implementation

This is lcc bug. Curiously, it is not triggered in `test_context.c' and
`test_grasshopper.c'.

"lcc: "/usr/src/gost-engine/test_curves.c", строка 32: ошибка:
          вероятное использование "=" вместо "=="
          [-Werror=assign-where-compare-meant]
  #define TEST_ASSERT(e) {if ((test = (e))) \
                              ^
 в раскрытие макроса "TEST_ASSERT" на строке 217
      TEST_ASSERT(0);"

"lcc: "/usr/src/gost-engine/gosthash2012_ref.h", строка 15: ошибка:
          нераспознанная #pragma [-Werror=unknown-pragmas]
  # pragma message "Use regular implementation"
           ^"

Also, cmake does not detect lcc.

There is other architectures besides __x86_64__, so EMMS should be
issued on IA-32 only where we are using MMX.

This probably would reduce size of the library. And remove
some 'not used' warnings (-Wunused-const-variable).

Previously `Ax' is generated dynamically from `A` and `Pi` in
`GOST3411Init', but in 2013 it is moved into `gosthash2012_precalc.h`.
`Tau' is unrolled and eliminated in 2013 too.

`_mm_empty' is not needed on x86_64, because we only using SSE2.

But, I didn't notice that EXTRACT32 (32-bit version of EXTRACT) is
using MMX registers and intrinsics, so complete removing of
`_mm_empty' (EMMS) was mistake.

Make it presence conditional only for IA-32.

Fixes: 211489fa ("gosthash2012: Improve SIMD implementation")