PEM_read_...: document that garbage and other PEM data is skipped

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15949)

load_key_certs_crls: Avoid reporting any spurious errors

When there is other PEM data in between certs the OSSL_STORE_load
returns NULL and reports error. Avoid printing that error unless
there was nothing read at all.

Fixes #15945

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15949)

test_pem_reading: Test loading a key from a file with multiple PEM data

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15949)

load_pkey_pem: Check for spurious errors when loading

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15949)

pem_read_bio_key: Add passphrase caching to avoid asking for password twice

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15949)

pem_read_bio_key_decoder: Avoid spurious error on unknown PEM data

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15949)

Fix compile errors when building with --api=1.1.0 no-deprecated.

Fixes #15963

INSTALL.md uses these exact options as an example so it should work.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15965)

update pyca-cryptography regression test suite

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15962)

Defined out MUTEX attributes not available on NonStop SPT Threads.

Standard Posix Threads (SPT) Threads are an older separate branch of
pthreads that do not support some of the capabilities in the current
Posix User Threads (PUT).

The change also includes a rename of the close field of OSSL_STORE_LOADER
which was causing preprocessor conflicts.

Fixes #15885

Signed-off-by: Randall S. Becker <rsbecker@nexbridge.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15886)

doc: Mention the update of der data pointers in d2i/i2d

Fixes #15958

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/15959)

DOC: clarify OPENSSL_API_COMPAT

Fixes #15928

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15956)

ci: add a memory sanitiser test run

This omission noted in #15950

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/15952)

afalg: add some memory initialisation calls to pacify memory sanitisation.

The engine is modifying memory without the sanitiser realising.  By pre-
initialising this memory, the sanitiser now thinks that read accesses are okay.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/15952)

Document cross-compile considerations for NonStop x86 builds.

Fixes #15919

Signed-off-by: Randall S. Becker <rsbecker@nexbridge.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15948)

Fix use of uninitialized memory in test_rsa_oaep

48f1739600f33c92387debce2002acec6e365f1d did not convert the RSA OAEP
tests correctly. The corrupted ciphertext and truncation tests were
really decrypting uninitialized memory, rather than the sample
ciphertext. This results in an error in tools like MSan.

The test is somewhat roundabout. In the original version, before the
conversion, ctext_ex was an OAEP test vector from key1(), etc.,
functions. The test would:

1. Encrypt ptext_ex as ctext.
2. Decrypt ctext and check it gives ptext_ex.
3. Decrypt ctext_ex and check it gives ptext_ex.
4. Try corrupted and truncated versions of ctext.

48f1739600f33c92387debce2002acec6e365f1d then moved steps 1 and 2 into
test_rsa_simple, which meant ctext is no longer available for step 4. It
then mistakenly left the variable around, but uninitialized, so the test
wasn't testing anything. (Confusingly, test_rsa_simple outputs ctext_ex
to the caller, but doesn't do anything with it. The ctext_ex output is
also only usable for OAEP, not PKCS#1 v1.5.)

It doesn't really matter whether we use ctext or ctext_ex for step 4, so
this PR fixes it by using ctext_ex instead.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15950)

OSSL_CRMF_{CERTTEMPLATE,CERTID}_get0_serialNumber(): Make result const for consistency

Also make doc/man3/OSSL_CRMF_MSG_get0_tmpl.pod consistent with crmf.h.in regarding const results

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15790)

cmp_mock_srv.c: Add missing OldCertID check for 'kur' cert update requests

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15790)

CMP: Add missing getter functions to CRMF API and CMP API

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15790)

test: fix test ordering in threads test

Fixes #15953

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15954)

evp: fix coverity 1473380 Copy into fixed size buffer (STRING_OVERFLOW)

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15943)

dh_test: fix coverity 1473239 Argument cannot be negative (NEGATIVE_RETURNS)

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15943)

s_time: avoid unlikely division by zero

Fixing coverity 966560 Division or modulo by zero (DIVIDE_BY_ZERO)

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15943)

bio: check for valid socket when closing

Fixes coverity 271258 Improper use of negative value (NEGATIVE_RETURNS)

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15943)

test: fix coverity 1469427 Improper use of negative value (NEGATIVE_RETURNS)

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15943)

x509: improve error reporting

Distinguish between not being able to extract a public key versus not knowing
the key's type.

Alternative to #15921

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15944)

Missing link to fips_config documentation

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15940)

coverity #1486532: fix potential NULL dereference in test_mk_file_path()

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15942)

coverity #1486531: return error properly from x509_pubkey_ex_new_ex()

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15942)

EVP: Have EVP_PKCS82PKEY_ex() pass a correct selection to OSSL_DECODER

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15934)

DECODER & ENCODER: Make sure to pass around the original selection bits

When decoding a key and asking the keymgmt to import the key data, it
was told that the key data includes everything.  This may not be true,
since the user may have specified a different selection, and some
keymgmts may want to be informed.

Our key decoders' export function, on the other hand, didn't care
either, and simply export anything they could, regardless.

In both cases, the selection that was specified by the user is now
passed all the way.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15934)

ossl_cmp_error_new(): Fix Coverity issue 1486534, and consequently also issues 1486536 and 1486533

The issues are due to an integer overflow that may happen on '(ERR_SYSTEM_FLAG << 1)'.

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15938)

doc: make error checking in ticket handling code explicit

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15918)

darwin platform replacing getentropy usage by platform api instead.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15924)

OSSL_STORE: Fix crash when tracing STORE

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15931)

ENCODER & DECODER: Make a tighter coupling between en/decoders and keymgmt

If there are keymgmts and en/decoders from the same provider, try to
combine them first.

This avoids unnecessary export/import dances, and also tries to avoid
issues where the keymgmt doesn't fully support exporting and importing,
which we can assume will be the case for HSM protected keys.

Fixes #15932

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15933)

PROV: Have our PEM->DER decoder only recognise our PEM names

This is to avoid creating confusion where other PEM decoder
implementations may know better what PEM names that are unknown to us
actually mean.

Fixes #15929

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15930)

Ensure ordinals are created during release process

We introduce a new makefile target "make release-update" that forces
ordinal file renumbering, and also does the fips checksum updates. We
then call that  from the release script.

Fixes #15806

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15901)

Update config.pm

Missing '(' added into a PowerPC-specific command

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15911)

evp_test: use correct size in memory clear

Fixes #15917

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15920)

DOC: Clarify the role of EKUs including defaults for TLS client and server use

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14199)

Fix 'openssl req' to correctly use the algorithm from '-newkey algo:nnnn'

We used the original string, which meant fetching for, for example,
'rsa:2048'.  That was, of course, doomed to fail.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15912)

ENCODER & DECODER: Allow en/decoders to have multiple names

We had prepared for this a little bit, but apparently not completed it.

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15904)

doc: update up call documentation

Some of the BIO functions weren't included in the provider-base documentation.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15909)

test/recipes/90-test_includes_data/vms-includes.cnf: correct the directory

... to mimic includes.cnf

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15823)

apps/CA.pl.in: restore the quotes around -CAfile, they were there for a reason

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15823)

test/recipes/80-test_ca.t: Don't force quotes around the config file in $cnf

However, when passing it through the OPENSSL_CONFIG environment
variable, we still need the quotes, just to make sure.

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15823)

test/recipes/66-test_ossl_store.t: ensure native paths

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15823)

testutil: teach test_mk_file_path() how to merge VMS file specs

This isn't a full solution, it only handles current use cases.

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15823)

test/ossl_store_test.c: Adapt the use of datadir for VMS paths

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15823)

UTF-8 not easily supported on VMS command line yet

Some tests are designed to test UTF-8 on the command line.
We simply disable those on VMS.

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15823)

Fix test_errstr for VMS

Occasionally, we get an error code on VMS that doesn't translate
into POSIX, and the error string reflects that

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15823)

Fix segfault in openssl x509 -modulus

The command ``openssl x509 -noout -modulus -in cert.pem`` used to segfaults
sometimes because an uninitialized variable was passed to
``BN_lebin2bn``. The bug triggered an assertion in bn_expand_internal().

Fixes: https://github.com/openssl/openssl/issues/15899
Signed-off-by: Christian Heimes <christian@python.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15900)

Add wrap.pl to .gitignore

This file is now auto-generated from a template (wrap.pl.in). Therefore
it should be ignored by git.

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15895)

Fix two typos in OSSL_trace_enabled.pod

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15894)

test: check for NULL returns better

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/15910)

test: avoid memory leaks on errors

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/15910)

evp_test: address NULL pointer dereference and return failure better

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/15910)

ui: address potential memory leak

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/15910)

apps: address potential memory leaks

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/15910)

x509: address NULL dereference and memory leaks

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/15910)

apps: properly initialise arguments to EVP_PKEY_get_bn_param()

This avoids use of uninitialised memory.

Follow on to #15900

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15908)

Update fips sources and checksums

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15903)

Only the fips module dependencies are relevant for fips.module.sources

Fixes #15639

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15903)

Fix file_name_check() in storemgmt/file_store.c and e_loader_attic.c

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15892)

enable getauxval on android 10

Fixes #9498

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15870)

(cherry picked from commit b2dea4d5f22ec146373324c282fb1bcecd5a7d90)

ppccap.c: Split out algorithm-specific functions

Fixes #13336

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15828)

CMP: Improve reporting of error codes and related strings via 'error' msg

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15879)

ossl_sk_ASN1_UTF8STRING2text(): Minor generalization and refactoring for readability

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15879)

CMP: Clean up internal message creation API and its documentation

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15879)

test: add EVP_Q_mac tests to evp_test

Fixes #15837

Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/15888)

test: add EVP_Q_digest tests to evp_test

Fixes #15837

Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/15888)

OpenSSL::Util::fixup_cmd_elements(): Include '!' among the VMS chars to process

! is the DCL character that starts a comment, and therefore acts as a
cut-off if not quoted.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15889)

BIO_new_from_core_bio: Fix heap-use-after-free after attach

The providers have to call up_ref to keep the cbio pointer, just like
the internal bio_prov.c does.

OSSL_STORE_attach passes a cbio pointer to the provider and then calls
ossl_core_bio_free(cbio). If up_ref is not called, the cbio gets
freed way too early.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15163)

trace: Do not produce dead code calling BIO_printf if disabled

Fixes #15880

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15882)

Fix a race in ossl_provider_add_to_store()

If two threads both attempt to load the same provider at the same time,
they will first both check to see if the provider already exists. If it
doesn't then they will both then create new provider objects and call the
init function. However only one of the threads will be successful in adding
the provider to the store. For the "losing" thread we should still return
"success", but we should deinitialise and free the no longer required
provider object, and return the object that exists in the store.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

Move OPENSSL_add_builtin back into provider.c

An earlier stage of the refactor in the last few commits moved this
function out of provider.c because it needed access to the provider
structure internals. The final version however no longer needs this so
it is moved back again.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

Update documentation following updates to the provider code

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

make struct provider_info_st a full type

Create the OSSL_PROVIDER_INFO to replace struct provider_info_st.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

Don't skip the current provider in ossl_provider_register_child_cb

This restriction was in place to avoid problems with recursive attempts
to aquire the flag lock/store lock from within a provider's init function.
Since those locks are no longer held when calling the init function there
is no reason for the restriction any more.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

Add a test to check that RAND_bytes_ex() works with a child lib ctx

Previously, when locks were held while calling a provider init function,
then RAND_bytes_ex() would fail if called from the init function and
used in conjunction with a child lib ctx. We add an explicit test of that.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

Don't hold any locks while calling the provider init function

Previously providers were added to the store first, and then subsequently
initialised. This meant that during initialisation the provider object
could be shared between multiple threads and hence the locks needed to be
held. However this causes problems because the provider init function is
essentially a user callback and could do virtually anything. There are
many API calls that could be invoked that could subsequently attempt to
acquire the locks. This will fail because the locks are already held.

However, now we have refactored things so that the provider is created and
initialised before being added to the store. Therefore at the point of
initialisation the provider object is not shared with other threads and so
no locks need to be held.

Fixes #15793
Fixes #15712

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

Only associate a provider with a store once it has been added to it

This means we can distinguish providers that have been added to the
store, and those which haven't yet been.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

Merge ossl_provider_activate() and ossl_provider_activate_child()

These 2 functions have become so close to each other that they may as well
be just one function.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

Set use_fallbacks to zero when we add a provider to the store

Update use_fallbacks to zero when we add a provider to the store rather
than when we activate it. Its only at the point that we add it to the store
that it is actually usable and visible to other threads.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

Remove flag_couldbechild

Now that a provider is no longer put into the store until after it has
been activated we don't need flag_couldbechild any more. This flag was
used to indicate whether a provider was eligible for conversion into a
child provider or not. This was only really interesting for predefined
providers that were automatically created.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

Add a new provider to the store only after we activate it

Rather than creating the provider, adding to the store and then activating
it, we do things the other way around, i.e. activate first and then add to
the store. This means that the activation should occur before other threads
are aware of the provider.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

Instantiate configuration supplied providers when we need them

If provider specified in a config file are not "activated" then we defer
instantiating the provider object until it is actually needed.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

Instantiate user-added builtin providers when we need them

Previously we created the provider object for builtin providers at the
point that OPENSSL_add_builtin() was called. Instead we delay that until
the provider is actually loaded.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

Instantiate predefined providers just-in-time

Previously we instantiated all the predefined providers at the point that
we create the provider store. Instead we move them to be instantiated as we
need them.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15854)

OSSL_DECODER_from_bio: Avoid spurious decoder error

If there are any new errors reported we avoid raising the
OSSL_DECODER_from_bio:unsupported error.

Fixes #14566

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15878)

epki2pki_decode: passphrase callback failure is fatal error

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15878)

ossl_pw_get_passphrase: No ui method does not necessarily mean internal error

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15878)

Documentation: SM2 keys can use only the SM2 curve

Fixes #14411

Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15875)

simpledynamic: Add missing include for AIX builds

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15874)

TEST: Modify simpledynamic.[ch] to allow use on VMS as well

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15872)

test/recipes/90-test_shlibload.t: Modify to work with known file names

Using File::Temp::tempfile() is admirable, but isn't necessary for the
sort of thing we use it for.

Furthermore, since tempfile() returns an opened file handle for
reading for the file in question, it may have effect that the file
becomes unwritable.  This is the default on VMS, and since tempfile()
doesn't seem to have any option to affect this, it means that
test/shlibloadtest.c can't write the magic line to that file.

Also, if we consider forensics, to be able to see what a test produced
to determine what went wrong, it's better to use specific and known
file names.

Therefore, this test is modified to use well known file names, and to
open them for reading after the shlibloadtest program has been run
instead of before.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15872)

property: remove spurious incorrect comments

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15871)

property: add locking for the property string database

This previously relied on the caller locking the property store correctly.
This is no longer the case so the string database now requires locking.

Fixes #15866

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15871)

err: add unable to get lock errors

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15871)

doc: Document that the OBJ creation functions don't lock.

Neither OBJ_create() nor OBJ_add_sigid() use locks.  They are not thread safe.
They can and will cause the other OBJ_ query functions to fail in mysterious
ways if called concurrently with them.

There is no problem calling multiple query functions concurrently.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15865)