Re-implement EVP_MD_name() and EVP_CIPHER_name() as functions

They will do the same as usual for non-provider algorithms
implementations, but can handle provider implementations as well.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9356)

Add a mechnism to save the name of fetched methods

This will be useful for information display, as well as for code that
want to check the name of an algorithm.  This can eventually replace
all NID checks.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9356)

provider-keymgmt(7): Document the KEYMGMT interface

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9429)

Remove superfluous call to OPENSSL_cpuid_setup

Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9417)

Change DH parameters to generate the order q subgroup instead of 2q

This avoids leaking bit 0 of the private key.

Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/9363)

Documentation: add provider-base(7), describing the base functions

The base functions are the first tables of function pointers that
libcrypto and the provider pass to each other, thereby providing a
baseline with which they can communicate further with each other.

This also contains an example for a ficticious provider, providing an
implement of a fictitious algorithm for a fictitious operation.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9409)

Documentation: Move the description of the fetching functions

Now that the general descriptions have moved from
doc/man3/EVP_MD_fetch.pod to doc/man7/provider.pod, the description of
the fetching functions themselves can be moved to other pages where
related functions are already described.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9409)

Documentation: Add provider(7), for general description of providers

This includes an enumeration of the providers supplied with OpenSSL,
and what implementations they offer.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9409)

Include deprecated SYS_F_xxx codes

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9072)

Deprecate SYSerr, add new FUNCerr macro

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9072)

Add ERR_put_func_error, and use it.

Change SYSerr to have the function name; remove SYS_F_xxx defines
Add a test and documentation.
Use get_last_socket_err, which removes some ifdef's in OpenSSL code.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9072)

Allocate DRBG additional data pool from non-secure memory

The additional data allocates 12K per DRBG instance in the
secure memory, which is not necessary. Also nonces are not
considered secret.

[extended tests]

Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9423)

x509: publish X509_PUBKEY_dup

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9353)

x509: add missing X509 dup functions

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9353)

x509: sort X509 dup functions alphabetically

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9353)

Add an internal API to access the KEYMGMT provider functions

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9312)

Add evp_keymgmt_clear_pkey_cache() and use it

This function clears the cache of provider key references, and is used
in evp_keymgmt_export_to_provider() when the internal key is dirty, as
well as by EVP_PKEY_free_it().

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9312)

Adapt int_ctx_new() to use with providers

This affects all its callers: EVP_PKEY_CTX_new(), EVP_PKEY_CTX_new_id().
They are now possible to called with "zero" values, i.e.:

    EVP_PKEY_CTX *pctx = EVP_PKEY_CTX_new(NULL, NULL);

or

    EVP_PKEY_CTX *pctx = EVP_PKEY_CTX_new_id(0, NULL);

This is suitable for provider use, as the key functionality is tied
with its keys, and the operation time is determined by the init
functions the EVP_PKEY_CTX is used with.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9312)

Add evp_keymgmt_export_to_provider(), for key transfer between providers

This function is used to transport private key materia from whatever
is already attached to the EVP_PKEY to the new provider, using key
data export and import functionality.

If a legacy lower level key has been assigned to the EVP_PKEY, we use
its data to create a provider side key, and thereby have a bridge
between old style public key types and the EVP_PKEY on providers.

If successful, this function returns a reference to the appropriate
provider side data for the key.

This can be used by any operation that wants to use this key.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9312)

Add basic EVP_KEYMGMT API and libcrypto <-> provider interface

The idea with the key management "operation" is to support the
following set of functionality:

- Key domain parameter generation
- Key domain parameter import
- Key domain parameter export

- Key generation
- Key import
- Key export
- Key loading (HSM / hidden key support)

With that set of function, we can support handling domain parameters
on one provider, key handling on another, and key usage on a third,
with transparent export / import of applicable data.  Of course, if a
provider doesn't offer export / import functionality, then all
operations surrounding a key must be performed with the same
provider.

This method also avoids having to do anything special with legacy
assignment of libcrypto key structures, i.e. EVP_PKEY_assign_RSA().
They will simply be used as keys to be exported from whenever they are
used with provider based operations.

This change only adds the EVP_KEYMGMT API and the libcrypto <->
provider interface.  Further changes will integrate them into existing
libcrypto functionality.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9312)

Add param builder free function.

This means include deallocation information in the return from
the ossl_param_bld_to_param function.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9404)

Cygwin: enable the use of Dl_info and dladdr()

These weren't available in Cygwin at the time our DSO code was
written, but things have changed since.

Fixes #9385

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9402)

test/enginetest.c: Make sure no config file is loaded

If a config file gets loaded, the tests get disturbed.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9393)

doc: fix some links

Some links are aged and need an adjustment.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9407)

util/find-doc-nits: fixups

- Treat .pod.in files as well, and parse out the base name for those
  too.
- Correct the detection of the description part in the NAME section
  (the separating dash MUST be preceeded with a space)
- Allow slahes in names of the NAME section (convert them to dashes
  for file name comparison).  This allows manual pages for some of our
  header files, such as openssl/core.h.
- Properly detect repeated names in the NAME section.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9407)

Fix no-dh

The recent move of the DH code into the default provider broke no-dh. This
adds back in various missing guards.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9399)

Correct some OSSL_PARAM documentation

The documentation wasn't quite in sync with the implementation.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9408)

Add value_barriers in constant time select functions

The barriers prevent the compiler from narrowing down the
possible value range of the mask and ~mask in the select
statements, which avoids the recognition of the select
and turning it into a conditional load or branch.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/9217)

Fix S390X compile error due to missing defines

Add the missing S390X_aes_XXX_gcm_CAPABLE() macros into aes_platform.h.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9403)

Cleanup use of X509 STORE locks

Cosmetic changes to use the X509_STORE_lock/unlock functions.
Renamed some ctx variables to store.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9366)

Fix init_get_thread_local()

Previously, init_get_thread_local() pushed the thread event handler
list onto the global register before calling CRYPTO_THREAD_set_local(),
and when the latter failed, forgot to pop the list from the stack again.

Instead of cleaning the stack on error, this commit avoids the situation
entirely by postponing the push operation until all other operations
succeeded. This reordering also significantly reduces the scope of the
critical section.

Another simplification of the code is achieved by moving the push operation
onto the register (which is disabled in FIPS mode) into a separate function.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9401)

Deprecated {OPENSSL,CRYPTO}_debug_mem_{push,pop}

They were only used for recursive ASN1 parsing.
Even if the internal memory-debugging facility remains,
this simplification seems worthwhile.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9342)

Fix the return value for SSL_get0_chain_certs()

This function was always returning 0. It should return 1 on success.

Fixes #9374

Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/9395)

Parameter building utilities.

A fuller implementation of PARAMS_TEMPLATE as per #9266 but renamed.
This introduces a statis data type which can be used to constructor a
description of a parameter array.  It can then be converted into a OSSL_PARAM
array and the allocated storage freed by a single call to OPENSSL_free.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9305)

remove end of line spaces

Reviewed-by: Paul Yang <yang.yang@baishancloud.com>
(Merged from https://github.com/openssl/openssl/pull/9397)

Remove tab characters from C source files.

Some have been creeping into the source code.

Reviewed-by: Paul Yang <yang.yang@baishancloud.com>
(Merged from https://github.com/openssl/openssl/pull/9397)

Auto add a label depending on the type of issue they report.

Reviewed-by: Richard Levitte <levitte@openssl.org>
GH: #9319

Provider config module: allow providers to already be loaded

This allows 'default' to be configured in the config file, if needed.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9371)

Fix SSL_CTX_set_session_id_context() docs

Also, use define rather than sizeof

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9377)

s390x assembly pack: fix restoring of SIGILL action

Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9381)

Actually silently ignore GET / OCSP requests

Reviewed-by: Matt Caswell <matt@openssl.org>

Document the new EVP_KEYEXCH type and related functions

Previous commits added the EVP_KEYEXCH type for representing key exchange
algorithms. They also added various functions for fetching and using them,
so we document all of those functions.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9266)

Add the ability to set PKCS#3 DH padding in providers

This also adds the ability to set arbitrary parameters on key exchange
algorithms. The ability to pad the output is one such parameter for DH.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9266)

Enable PKCS#3 DH in the providers

The default provider now has support for PKCS#3 Diffie-Hellman so we
switch libcrypto to using providers for that algorithm.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9266)

Implement PKCS#3 DH Key Exchange in the default provider

We add the capability for the default provider to perform PKCS#3
Diffie-Hellman key exchange. At this point the implementation is not used
because libcrypto still uses legacy handling for Diffie-Hellman.

Note X9.42 DH is not touched by this commit.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9266)

Make the EVP Key Exchange code provider aware

We introduce a new EVP_KEYEXCH type to represent key exchange algorithms
and refactor the existing code to use it where available.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9266)

Regenerate mkerr files

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9058)

Remove function name from errors

Deprecate all xxx_F_ defines.
Removed some places that tested for a specific function.
Use empty field for the function names in output.
Update documentation.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9058)

Remove DRBG from SSL structure.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9390)

Add Common shared code needed to move aes ciphers to providers

Custom aes ciphers will be placed into multiple new files
(instead of the monolithic setup used in the e_aes.c legacy code)
so it makes sense to have a header for the platform specific
code that needs to be shared between files.
modes_lcl.h has also moved to modes_int.h to allow sharing with the
provider source.
Code that will be common to AEAD ciphers has also been added. These
will be used by seperate PR's for GCM, CCM & OCB.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9301)

remove end of line whitespace

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9111)

Convert asn1_dsa.c to use the WPACKET API instead

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9111)

Give WPACKET the ability to have a NULL buffer underneath it

This means the WPACKET API can be used for calculating the number of
bytes that would have been written if a non-NULL buffer had been used.
This enables us to calculate the number of length bytes required when
encoding ASN.1

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9111)

Convert asn1_dsa.c to use the PACKET API instead

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9111)

Make the PACKET/WPACKET code available to both libcrypto and libssl

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9111)

Add simple ASN.1 utils for DSA signature DER.

Adds simple utility functions to allow both the default and fips providers to
encode and decode DSA-Sig-Value and ECDSA-Sig-Value (DSA_SIG and ECDSA_SIG
structures) to/from ASN.1 DER without requiring those providers to have a
dependency on the asn1 module.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9111)

Coverity #1451596: check dirlen for being negative

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/9362)

Coverity #1451595: use correct free function.

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/9362)

util/opensslwrap.sh: adjust to define OPENSSL_MODULES as well

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9379)

In documentation, consistently refer to OpenSSL 3.0

3.0.0 is a habit from pre-3.0 OpenSSL, which doesn't make sense with
the new version scheme.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9376)

Make sure all BIGNUM operations work within the FIPS provider

The FIPS provider does not have a default OPENSSL_CTX so, where
necessary, we need to ensure we can always access an explicit
OPENSSL_CTX. We remove functions from the FIPS provider that use
the default OPENSSL_CTX, and fixup some places which were using
those removed functions.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9310)

Move the code for 'openssl list' to its own translation unit.

That makes it easier to work with than going through apps/openssl.c
This also moves the implementation of calculate_columns() and makes it
generally accessible.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9340)

Refactor apps/progs.* to be generate with 'make update'

This makes for a cleaner apps/progs.h as well as as cleaner
apps/build.info.

We also break out the type declarations to apps/include/function.h

apps/progs.c and apps/progs.h are NOT regenerated when 'apps' is
disabled.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9340)

Fix wrong lock claimed in x509 dir lookup.

x509 store's objects cache can get corrupted when using dir lookup
method in multithreaded application. Claim x509 store's lock when
accessing objects cache.

CLA: trivial

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9326)

issue-9316: Update return documentation for RAND_set_rand_engine

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9351)

Make allocation/free/clean available to providers

Also make OPENSSL_hexstr2buf available to providers.
EVP control functions need hexstring conversion, so move any
memory-allocating functions in o_str.c into new file mem_str.c

Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8886)

Adapt the provider AES for more use of OSSL_PARAM

The cipher context IV was a bit interesting.  EVP_CIPHER_CTX_iv()
returns a pointer to the live IV, while EVP_CIPHER_CTX_ctrl() with the
type EVP_CTRL_GET_IV gets a copy of the live IV.  To support both, we
support getting it with both the OSSL_PARAM_OCTET_STRING and
OSSL_PARAM_OCTET_PTR datatypes.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9328)

Adapt diverse EVP_CIPHER functions to use get_params and set_params interfaces

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9328)

Make more use of OSSL_PARAM for ciphers

A lot of the different numbers associated with ciphers are really
algorithm parameters.  Key length, block size, IV length, that sort of
thing.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9328)

test/evp_test.c: [ciphers] Test that we get back the same IV we gave

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9328)

Replace long dash characters with normal ascii minus

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9321)

Add X9.42 KDF.

Move the KDF code for CMS DH key agreement into an EVP_KDF object.
There are 2 specifications for X9.42 KDF. This implementation uses DER for
otherinfo which embeds the KDF loop counter inside the DER object.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8898)

Avoid NULL pointer dereference. Fixes #9043.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9059)

A very brief explanation of how to add custom functions to OpenSSL.

Inspired by Rich Salz's letter to openssl-users@

[edited to remove non-ASCII characters and end of line white space]

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9267)

iOS build: Replace %20 with space in config script

CLA: trivial

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9297)

Fix build error for aarch64 big endian.

Modified rev to rev64, because rev only takes integer registers.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90827
Otherwise, the following error will occur.

Error: operand 1 must be an integer register -- `rev v31.16b,v31.16b'

CLA: trivial

Signed-off-by: Lei Maohui <leimaohui@cn.fujitsu.com>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9151)

man: fix typo in OPENSSL_fork_prepare.pod

Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/9318)

Fix an endless loop in BN_generate_prime_ex

Happens when trying to generate 4 or 5 bit safe primes.

[extended tests]

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9311)

Prevent the use of RUN_ONCE inside the FIPS module

FIPS module code *always* runs within the scope of an associated
OPENSSL_CTX. When the module is loaded the OPENSSL_CTX gets created, and
when the module is unloaded the OPENSSL_CX gets freed. A module may be
loaded multiple times within the scope of different OPENSSL_CTX objects.
"Global" data should always be stored within the OPENSSL_CTX. In this
way it will always get cleaned up properly when the module is unloaded.

All current code within the FIPS module works this way. To avoid
"accidents" we disabled the RUN_ONCE code inside the FIPS module.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9308)

Clarify documentation of SSL_CTX_set_verify client side behavior

Fixes #9259

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9291)

Fix a typo and a syntax error in opensslconf.h

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9298)

test/recipes/02_test_errstr.t: Make it less fragile

Change it to split the error string returned by `openssl errstr` in a
more robust manner, and ensure it's the reason code we look at.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9304)

Adding Test.pm with workaround for Perl abs2rel bug

If SRCTOP != BLDTOP, and SRCTOP is given in relative form, e.g.
"./config ../openssl", then a bug in Perl's abs2rel may trigger that directory-
rewriting in __cwd results in wrong entries in %directories under certain
circumstances, e.g. when a test executes run(app(["openssl"]) after indir.

There should not be any need to go to a higher directory from BLDDIR or SRCDIR,
so it should be OK to use them in their absolute form, also resolving all
possible symlinks, right from the start.

Following the File::Spec::Functions bug description (reported to perl.org):

When abs2rel gets a path argument with ..s that are crossing over the ..s
trailing the base argument, the result is wrong.

Example
PATH: /home/goal/test/..
BASE: /home/goal/test/../../base
Good result: ../goal
Bad  result: ../..

Bug verified with File::Spec versions
- 3.6301
- 3.74 (latest)

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/7031)

Fix comment; unchecked->checked

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9292)

Add a dummy call to BN_rand_ex() in the FIPS provider

The previous commit made BIGNUM RAND operations available from within
the FIPS provider. We test this out by making a dummy call to check it
completes successfully.

Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/9193)

Make BIGNUM rand functions available within the FIPS module

The BIGNUM rand functions were previously disabled for the FIPS module.
We can now re-enable them.

Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/9193)

Provide rand_bytes_ex and rand_priv_bytes_ex

We provider internal versions of RAND_bytes() and RAND_priv_bytes() which
have the addition of taking an OPENSSL_CTX as a parameter.

Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/9193)

s390x assembly pack: fix various aes modes performance regression

which was introduced with 64adf9aac7.

Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9271)

FIPS module: adapt for the changed error reporting methods

The FIPS module inner provider doesn't need to deal with error reason
strings or error library number, since it uses the outer provider's
error reporting upcalls.  We therefore disable that code in
crypto/provider_core.c when building the FIPS module.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9174)

Replumbing: re-implement error reporting for providers

The idea is that providers should only have to report a reason code.
The library code is considered to be libcrypto internal, and are
allocated dynamically and automatically for providers on creation.

We reserve the upper 8 bits of the reason code for internal OpenSSL
use.  This allows our own providers to report errors in form of a
packed number that includes library number, function number and
reason number.

With this, a provider can potentially use any reason number it wants
from 1 to 16777216, although the current error semantics really only
allow 1 to 4095 (because only the lower 12 bits are currently
considered an actual reason code by the ERR subsystem).

A provider can provide a reason string table in form of an array of
ERR_STRING_DATA, with each item containing just the reason code and
the associated string, with the dispatch function numbered
OSSL_FUNC_PROVIDER_GET_REASON_STRINGS matching the type
OSSL_provider_get_reason_strings_fn.
If available, libcrypto will call that function on provider
activation.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9174)

ossl_provider_upref to ossl_provider_up_ref

Common pattern is that the routines to increment the reference count
are called something_up_ref, not something_upref.  Adapt
ossl_provider_upref() accordingly.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9293)

Check for V_ASN1_BOOLEAN/V_ASN1_NULL in X509_ATTRIBUTE_get0_data

The member value.ptr is undefined for those ASN1 types.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9278)

Fix ASN1_TYPE_get/set with type=V_ASN1_BOOLEAN

BOOLEAN does not have valid data in the value.ptr member,
thus don't use it here.

Fixes #9276

[extended tests]

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9278)

Fix Typos

CLA: trivial

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/9288)

test/p_test.c: silence -Wstringop-overflow

Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9282)

Remove EXPORT_VAR_AS_FUNC

We only export functions, not global, so remove the config option
and some of the #ifdef stuff.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9285)

Remove global-var/function macros

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9284)

Remove DES_check_key global

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9284)

Remove NextStep support

Because of that we can remove OPENSSL_UNISTD and some other
macros from e_os2.h and opensslconf.h

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9204)

Fix a leak in evp_test

If evp_test fails to load the legacy provider then it leaks a reference
to the default provider.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9270)