Use --debug with no-caching build as sanitizers need it

The memleak test otherwise fails.

Also disable async, dtls, and old tls versions to test some
different combination of disableables and speed up tests.

Fixes #14337

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14536)

Add a CHANGES entry for EVP_PKEY_public_check() and EVP_KEY_param_check()

These functions now work for more key types than they did in 1.1.1

Fixes #14477

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14485)

Ensure that ECX keys pass EVP_PKEY_param_check()

RSA keys have no parameters and pass EVP_PKEY_param_check(). Previously,
ECX keys had no parammeters and failed EVP_PKEY_param_check(). We should
be consistent. It makes more sense to always pass, and therefore this
commit implements that behaviour.

Fixes #14482

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14485)

Add a CHANGES entry for the cosmetic differences in textual output

Numerous functions have had their textual output amended. We add
a CHANGES entry for this.

Fixes #14476

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14485)

apps: Make load_key_certs_crls to read only what is expected

The load_key_certs_crls tried to read the whole input stream
instead of returning once expected data is obtained.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14449)

apps: Add maybe_stdin argument to load_certs and set it in pkcs12

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14449)

Tiny clarification of comment for RSA_sign

CLA: trivial

On line 136, a period is added. I think this is what was intended.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14540)

Fix DSA EVP_PKEY_param_check() when defaults are used for param generation.

Fixes #14480

An internal flag that is set during param gen was not being tested, so
the wrong type was used to select the dsa domain param validation method.

In the default provider - if no gen_type is set then by default the fips186_4 gentype
will be selected when pbits >=2048 otherwise it selects fips186_2.
The fips provider ignores the gen_type and always uses fips186_4.

Before this change dsa used fips186_2 by default in the default
provider.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14508)

keymgmt_meth: remove two TODO 3.0

The first TODO 3.0 is not really a TODO, just a comment.

The second one is something that is needed for compatibility
with existing applications. There is no major reason in
trying to change this behavior right now.

Fixes #14400

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14534)

Fix option description for PKCS#12 export

Refs: https://github.com/openssl/openssl/pull/4930

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14520)

Convert a TODO(3.0) in OPENSSL_thread_stop_ex to a comment

The TODO is describing something that would be nice to fix. In fact the
problem exists even in 1.1.1. It would be nice to fix it, but it does
not need to be done in the 3.0 timeframe.

Fixes #14376

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14533)

Remove a TODO from async_delete_thread_state()

There is nothing to be done here for the time being. If at some point
we make the async code libctx aware then we might need to make a change
but there are no plans to do that at the moment.

Fixes #14402

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14532)

Remove TODO in rsa_ameth.c

Fixes #14390

The only caller of this function tests EVP_KEYMGMT_is_a() beforehand
which will fail if the RSA key types do not match. So the test is not
necessary. The assert has been removed when it does the test.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14524)

Remove TODO in test/acvp_test.c related to setting AES-GCM iv.

Fixes #14330

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14525)

Always check CRYPTO_LOCK_{read,write}_lock

Some functions that lock things are void, so we just return early.

Also make ossl_namemap_empty return 0 on error.  Updated the docs, and added
some code to ossl_namemap_stored() to handle the failure, and updated the
tests to allow for failure.

Fixes: #14230
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14238)

apps/ts.c: Allow -untrusted arg to refer to multiple sources

This requires moving generally useful functions from apps/cmp.c to apps/lib/apps.c

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14504)

TS ESS: Let TS_RESP_verify_signature() make use of untrusted certs also from token response

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14504)

ssl: fix format specifier for size_t argument to BIO_printf

Fixes #14519

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14521)

acvp_test: Do not expect exact number of self tests

There might be more because internal instances of the DRBG
might be initialized for the first time and thus
self-tested as well.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14497)

Remove the RAND_get0_public() from fips provider initialization

It is not needed anymore and it causes leaks because
it is called when the FIPS provider libctx is not yet
properly set up.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14497)

Use OPENSSL_init_crypto(OPENSSL_INIT_BASE_ONLY, NULL) in libcrypto

Calling OPENSSL_init_crypto(0, NULL) is a no-op and will
not properly initialize thread local handling.

Only the calls that are needed to initialize thread locals
are kept, the rest of the no-op calls are removed.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14497)

Update CHANGES with info about AuthEnvelopedData addition

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14456)

rename ossl_provider_forall_loaded to ossl_provider_doall_activated

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14489)

doc: describe the return from ossl_provider_forall_loaded()

Also correct an incorrect statement about non-activated providers.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14489)

core: modify ossl_provider_forall_loaded() to avoid locking for the callbacks

To avoid recursive lock issues, a copy is taken of the provider list and
the callbacks are made without holding the store lock.

Fixes #14251

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14489)

Don't crash if the pkeyopt doesn't have a value

All pkeyopt's must have a ":" and a value for the option. Not supplying
one can cause a crash

Fixes #14494

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14496)

update set_ctx_param store management calls to return 1 for a NULL params

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

update set_ctx_param DRBG calls to return 1 for a NULL params

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

update set_ctx_param MAC calls to return 1 for a NULL params

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

prov: add extra params argument to KDF implementations

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

doc: note that get_params and set_params calls should return true if the param array is null

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

doc: document the additional params argument to the various init() calls

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

support params argument to AES cipher init calls

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

doc: update cipher documentation to include the new init functions with params

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

prov: support params argument to common cipher init calls

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

prov: support param argument to DES cipher init calls

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

prov: support param argument to null cipher init calls

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

prov: support params argument to CHACHA20 ciphers

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

prov: support params argument to RCx ciphers

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

prov: support params arguments to signature init calls

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

prov: update digests to support modified ctx params

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

doc: update digest documentation to include the new init functions with params

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

prov: support param argument to digest init calls

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

doc: document param argument to RSA calls

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

doc: document param argument to cipher init calls

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

test: support params arguments to init functions

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

ssl: support params arguments to init functions

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

apps: support param argument to init functions

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

prov: update KEM to support params on init()

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

prov: update exchange algorithms to support params on the init call

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

misc: other init function param additions

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

doc: update PKEY documentation to include the new init functions with params

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

evp: add params arguments to init functions

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

core: add params arguments to init calls

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

prov: asym ciphers take an extra init() params argument

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

doc: add params argument to key manager's gen_init call

Fixes #14286

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

core: add params argument to key manager's gen_init call

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

provider: add params argument to key manager's gen_init call

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

evp: add params argument to key manager's gen_init call

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

test: add params argument to key manager's gen_init call

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14383)

Fix error-checking compiles for mutex

Fixes: #14229
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14264)

Increase the upper limit on group name length

While all the standardized groups would fit within the old limit,
with the addition of providers, some might want to experiment with
new and unstandardized groups. As such, their names might not fit
within the old limit.

Define it as GROUP_NAME_BUFFER_LENGTH with value 64.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14502)

TEST: Stop the cleanup in test/recipes/20-test_mac.t

Let the files remain to make test forensics easy

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14484)

PROV: use EVP_CIPHER_CTX_set_params() rather than EVP_CIPHER_CTX_ctrl()

This is in gmac_final(), where the cipher is known to be fetched.
It's more suitable to use OSSL_PARAMs than _ctrl functions, as the
latter are expected to become obsolete.

Fixes #14359

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14484)

openssl-cmp.pod.in and apps/cmp.c: Various minor do improvements

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14493)

Prepare for 3.0 alpha 14

Reviewed-by: Richard Levitte <levitte@openssl.org>

Prepare for release of 3.0 alpha 13

Reviewed-by: Richard Levitte <levitte@openssl.org>

Update copyright year

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14512)

Use read/write locking on Windows

Fixes #13914

The "SRWLock" synchronization primitive is available in Windows Vista
and later.  CRYPTO_THREAD functions now use SRWLock functions when the
target operating system supports them.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14381)

Check SSL_set1_chain error in set_cert_cb

CLA: trivial

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14469)

Fix reason code: EVP_R_OPERATON_NOT_INITIALIZED

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14429)

test: convert store test to use relative paths

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14419)

core: add up_ref callback for OSSL_CORE_BIO

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14419)

Add a real type for OSSL_CORE_BIO which is distinct from and not castable to BIO

Providers (particularly the FIPS provider) needs access to BIOs from libcrypto.
Libcrypto is allowed to change the internal format of the BIO structure and it
is still expected to work with providers that were already built.  This means
that the libcrypto BIO must be distinct from and not castable to the provider
side OSSL_CORE_BIO.

Unfortunately, this requirement was broken in both directions.  This fixes
things by forcing the two to be different and any casts break loudly.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14419)

Use BIO_f_readbuffer() in the decoder to support stdin.

Fixes #13185
Fixes #13352

Removed the existing code in file_store that was trying to figure out the
input type.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14407)

Add new filter BIO BIO_f_readbuffer()

This allows BIO_tell() and BIO_seek() to work for BIO's that do
not support these methods. The main use case for this is file/fd BIO's
that use stdin.

This works for stdin taken from input redirection (command < file),
and stdin via pipe (cat file | command).

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14407)

Fix formatting error of HISTORY section in some manual pages.

Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/14450)

Change default algorithms in PKCS12_create() and PKCS12_set_mac()

Use the modern defaults as now set in the pkcs12 app. This also
allows modifying the application to not override the default values
when calling the API.

Fixes #14034

Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/14450)

Mention the change of licence in NEWS.md

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14465)

Expand the CHANGES entry for SHA1 and libssl

As well as SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 not working at
security level 1 we also document that TLS 1.2 connection will fail
if the ClientHello does not have a signature algorithms extension.

Fixes #14447

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14465)

Add a CHANGES for OSSL_STORE_INFO_get_type()

The function OSSL_STORE_INFO_get_type() may now return a new object
type. Applications may have to be amended accordingly.

Fixes #14446

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14465)

Add a missing CHANGES.md entry for the legacy provider

Numerous ciphers and digests have been moved to the legacy provider.
There should be a CHANGES.md entry pointing this out.

Fixes #14441

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14465)

Non-const accessor to legacy keys

Fixes #14466.

Reverting the changes of the EVP_PKEY_get0 function.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14468)

EVP_KDF-KB man page: Fix typo in the example code

CLA: trivial
Signed-off-by: Arthur Gautier <baloo@superbaloo.net>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14455)

Fixup support for io_pgetevents_time64 syscall

This is a fixup for the original commit 5b5e2985f355c8e99c196d9ce5d02c15bebadfbc
"Add support for io_pgetevents_time64 syscall" that didn't correctly
work for 32-bit architecutres with a 64-bit time_t that aren't RISC-V.

For a full discussion of the issue see:
https://github.com/openssl/openssl/commit/5b5e2985f355c8e99c196d9ce5d02c15bebadfbc

Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14432)

cmp_hdr.c: Fix minor Coverity issue CID 1473605

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14460)

http_test.c: Fix minor Coverity issue CID 1473608

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14460)

Reword repeated words.

A trivial PR to remove some commonly repeated words. It looks like this is
not the first PR to do this.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14420)

apps/pkcs12: Allow continuing on absent mac

Just print a warning in that case.

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14445)

apps/pkcs12: Detect missing PKCS12KDF support on import

Report error message with hint to use -nomacver if
MAC verification is not required.

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14445)

apps/pkcs12: Properly detect MAC setup failure

The MAC requires PKCS12KDF support which is not present
in FIPS provider as it is not an approved KDF algorithm.
Suggest using -nomac if MAC is not required.

Fixes #14057

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14445)

fake_rand_finish should be called if "OPENSSL_NO_SM2" is NOT defined

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14471)

Fix the check for suitable groups and TLSv1.3

If we have TLSv1.3 enabled then we must have at least one TLSv1.3 capable
group available. This check was not always working

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/14430)

Make the EVP_PKEY_get0* functions have a const return type

OTC have decided that the EVP_PKEY_get0* functions should have a const
return type. This is a breaking change to emphasise that these values
should be considered as immutable.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14319)

Document the change in behaviour of the the low level key getters/setters

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14319)

Ensure the various legacy key EVP_PKEY getters/setters are deprecated

Most of these were already deprecated but a few have been missed. This
commit corrects that.

Fixes #14303
Fixes #14317

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14319)

Cache legacy keys instead of downgrading them

If someone calls an EVP_PKEY_get0*() function then we create a legacy
key and cache it in the EVP_PKEY - but it doesn't become an "origin" and
it doesn't ever get updated. This will be documented as a restriction of
the EVP_PKEY_get0*() function with provided keys.

Fixes #14020

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14319)

Avoid a null pointer deref on a malloc failure

Make sure we were sucessful in creating an EVP_PKEY

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14319)

Add a multi thread test for downgrading keys

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14319)

Restore GOST macros compatibility with 1.1.1

Fixes #14440

Before IANA assigned the official codes for the GOST signature
algorithms in TLS, the values from the Reserved for Private Use range
were in use in Russia. The old values were renamed.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14448)