docs: update man3 and man7 with cipher pipeline APIs

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/24636)

test: add evp_extra_test case for cipher pipeline API with fake pipeline provider

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/24636)

feat: add EVP APIs for cipher pipelining

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/24636)

docs: EVP pipeline API design document

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/24636)

Update HISTORY sections of libssl functions > 3.0.0

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26158)

Fix breakage on arm64_32 platforms

We just avoid the special handling needed for Apple M1.

Fixes #26135

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Saša Nedvědický <sashan@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
(Merged from https://github.com/openssl/openssl/pull/26145)

openssl-dgst: Document that xoflen is required for shake

With b911fef216d1386210ec24e201d54d709528abb4, there is no longer a
default xoflen for shake algorithms. Update the manual to reflect this.

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26129)

Extension of OPENSSL_ia32cap to accommodate additional CPUID bits

bits 128 - 191 CPUID.(EAX=07H,ECX=0H).EDX and CPUID.(EAX=07H,ECX=1H).EAX
bits 192 - 255 CPUID.(EAX=07H,ECX=1H).EDX and CPUID.(EAX=07H,ECX=1H).EBX
bits 256 - 319 CPUID.(EAX=07H,ECX=1H).ECX and CPUID.(EAX=24H,ECX=0H).EBX

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25709)

Fix typo in SSL_get_shared_sigalgs docs

psighash -> psignhash

CLA: trivial

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26151)

Limit tests with non-openssl clients

Several quic interop implementations have a server implementation, but
not a client implementation.  Don't bother trying to run those

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26130)

Add HISTORY to BIO_s_mem documentation

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26146)

Fix leak of a RAND_POOL in error condition

Fixes Coverity 1636676

Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26154)

X509: document non-standard behavior checking EKU extensions in CA and TA certs

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/26142)

Fix the intermittent lhash_test failure on Windows

We must set pending_delete before the actual deletion as another inserting
or deleting thread can pick up the delete callback before the
ossl_ht_write_unlock() call.

This can happen only if no read locks are pending and only on Windows where
we do not use the write mutex to get the callback list.

Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/26152)

To verify MAC, we need a MAC

Fixes #26106

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26140)

test: the timeSpecification X.509v3 extension

Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25476)

doc: the timeSpecification X.509v3 extension

Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25476)

feat: support the timeSpecification X.509v3 extension

Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25476)

dh_cms_set_peerkey(): Fix the incorrect condition

Only absent parameters allowed in RFC 3370.

Fixes #25824

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26058)

Add function to mix in an additional input into a RAND_POOL

It will be just xor-ed over the existing entropy
in the pool.

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Saša Nedvědický <sashan@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26128)

Fix potential use-after-free in REF_PRINT_COUNT

We use REF_PRINT_COUNT to dump out the value of various reference
counters in our code

However, we commonly use this macro after an increment or decrement.  On
increment its fine, but on decrement its not, because the macro
dereferences the object holding the counter value, which may be freed by
another thread, as we've given up our ref count to it prior to using the
macro.

The rule is that we can't reference memory for an object once we've
released our reference, so lets fix this by altering REF_PRINT_COUNT to
accept the value returned by CRYPTO_[UP|DOWN]_REF instead.  The
eliminates the need to dereference the memory the object points to an
allows us to use the call after we release our reference count

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25664)

Fix memory ordering guarantees and TSAN errors

If we had refcounted object allowing lockless writes
the relaxed semantics on DOWN_REF would allow scheduling
these writes after simultaneous release of the object by
another thread.

We do not have any such objects yet, but better to make
the refcount correct just in case we will have them
in future.

TSAN doesn't properly understand this so we use
even stronger acq_rel semantics if building with TSAN.

Fixes #25660

Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25664)

Add test for releasing a shared EVP_PKEY across threads

Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25664)

ossl_cms_get1_crls_ex(): Avoid doublefree if CRL up ref fails

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/26100)

(cherry picked from commit ef0be53f90045d0a9f0d085a7a4289335f49ee41)

Revert the behavior change of CMS_get1_certs() and CMS_get1_crls()

Fixes #26079

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/26100)

(cherry picked from commit afd36cbef8b3b7b00bd4bcdc33802d4cb39fdffa)

Remove references to git.openssl.org from README.md

Reviewed-by: Saša Nedvědický <sashan@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26122)

speed.c: Changed the declaration of aead_ivlen to a #define

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26109)

Exchange jitter and no-ct jobs between daily and on push CI

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Saša Nedvědický <sashan@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26112)

jitter_generate(): Properly mix in the additional input

By adding the additional input directly to the pool
we were using just the additional input.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Saša Nedvědický <sashan@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26112)

A typo fix in a comment

CLA: trivial

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26095)

Add spaces to make macros more clear

CLA: trivial

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26095)

Code & comments changes to make them in consistent

CLA: trivial

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26095)

one more empty line for code style consistency

CLA: trivial

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26095)

Change all one's complement to ones' complement for consistency

CLA: trivial

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26095)

BN_secure_new function indentation correction

CLA: trivial

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26095)

Delete redundant parentheses for code consistency

CLA: trivial

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26095)

Delete a redundant break

CLA: trivial

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26095)

Run interop tests on feature/quic-server branch

We currently run interop tests as a client only from the master branch.
While we are developing quic-server it would be beneficial to also get
interop test results from the quic-server branch run as both a client
and a server, until such time as the feature branch is merged.  Add
building and running of a container in the test harness to our CI set

Reviewed-by: Saša Nedvědický <sashan@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26090)

Add handling for additional input in jitter rng

Fixes #25917

Reviewed-by: Saša Nedvědický <sashan@openssl.org>
Reviewed-by: Hugo Landau <hlandau@devever.net>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25966)

fips: zeroization of ECX public keys

Commit fa338aa7cd added zeroization of public security parameters as
required by ISO 19790:2012/Cor.1:2015 7.9. However, that commit
overlooked ECX keys, which are used for EdDSA and X25519/X448.

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25807)

providers: stop probing for getentropy(3) on recent FreeBSD

FreeBSD has supported both getrandom(2) and getentropy(3) since 12.0.
The last version which did *not* have these went EoL in September 2021.
Use getrandom(2) unconditionally and fallback to sysctl kern.arandom if
we do happen to have a FreeBSD that old.

This is generally a necessary step for FreeBSD's _FORTIFY_SOURCE
implementation, which needs to do some symbol renaming tricks with the
getentropy declaration that would otherwise add some platform-specific
hacks here to accommodate.  getentropy(3) uses getrandom(2) internally
on FreeBSD, so we just cut out the middleman.

While we're here, it doesn't seem to make sense to ever prefer the
sysctl on FreeBSD or NetBSD.  For both platforms, it's limited to 256
bytes in a single request while getrandom(2) will generally use the same
backend but service the entire request in one shot, even for larger
amounts of entropy, modulo the EINTR possibility that presents itself
with larger requests.

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/24903)

fips-label.yml: Fix ABI change label removal

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/26080)

Avoid NULL dereference with PKCS7_OP_SET_DETACHED_SIGNATURE

We would dereference p7->d.sign pointer which can be NULL.

Reported by Han Zheng.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26078)

Document version-specific utility of EVP_PKEY_Q_keygen

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26040)

Minor nit fix to EVP_CipherFinal_ex() return code documentation.

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25849)

Fix memleak in dsa_gen()

Free the stack return value `dsa` on each early exit.

Fixes #25905

Signed-off-by: Holger Dengler <dengler@linux.ibm.com>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Saša Nedvědický <sashan@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25948)

Fix EVP_PKEY_print_private() so that it works with non default providers.

At some point in time it was decided that the EC keymanagers ec_export()
function would only allow the selection to be both the public + private
parts. If just the private element is selected it returns an error.
Many openssl commandline apps use EVP_PKEY_print_private() which passes
EVP_PKEY_PRIVATE_KEY to the encoder. This selection propagates to
encoder_construct_pkey(). For external providers (such as the fips
provider this will call the keymanagers export() with the selection set
to just the private part.

So we either need to
1) change the selection in EVP_PKEY_print_private() or
2) modify the selection used in the export used in
   encoder_construct_pkey
3) Change the ec_export to allow this.

I have chosen 2) but I am not sure if this is the correct thing to do
or whether it should conditionally do this when the output_type ==
'text'.

Issue was reported by Ilia Okomin (Oracle).

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26004)

die() in .tmpl file should not be silently ignored.

call to die() in perl templates is currently ignored.
any error printed by die() commad appears in template
output.

In order to make sure die() terminates processing we
must ensure we emite `undef` value. This is ensured
by adding a `BROKEN` callback to `fill_in()` Template
method. The callback must return undef to stop processing.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26064)

apps/passwd.c: Convert a redundant check to assert

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Paul Yang <kaishen.yy@antfin.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26053)

Fix solaris build in CRYPTO_atomic_store api

Misnamed variable, just correct it to dst

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Saša Nedvědický <sashan@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26075)

Use static array (length 256) for copy of OPENSSL_MALLOC_FAILURES

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26039)

Deprecate all BIO_meth_get_*() functions

Their use by applications is inherently unsafe.
Fixes #26047

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
(Merged from https://github.com/openssl/openssl/pull/26056)

Use sk_X509_ATTRIBUTE_deep_copy() to copy attribute stacks in pk7_doit.c

Clean up the code by using the dedicated stack copy function.

Reviewed-by: Paul Yang <kaishen.yy@antfin.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25713)

ci: add daily runcheckers to exercise the -DOPENSSL_PEDANTIC_ZEROIZATION option

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26068)

fips: change integrity check zeroization to use the OPENSSL_PEDANTIC_ZEROIZATION define

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26068)

pbkdf2: change FIPS zeroization to use the OPENSSL_PEDANTIC_ZEROIZATION define

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26068)

hkdf: change FIPS zeroization to use the OPENSSL_PEDANTIC_ZEROIZATION define

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26068)

rsa: change FIPS zeroization to use the OPENSSL_PEDANTIC_ZEROIZATION define

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26068)

ffc: change FIPS zeroization to use the OPENSSL_PEDANTIC_ZEROIZATION define

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26068)

ec: change FIPS zeroization to use the OPENSSL_PEDANTIC_ZEROIZATION define

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26068)

A typo fix in a comment

CLA: trivial

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26069)

Document expected BIO operations for libssl

If your custom BIO does not implement BIO_CTRL_FLUSH, it won't work, but
this is not document anywhere.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Saša Nedvědický <sashan@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26060)

Upgrade action/{upload,download}-artifact to v4

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
(Merged from https://github.com/openssl/openssl/pull/25920)

fips-jitter: set provider into error state upon CRNG permanent failures

With fips-jitter build time option, jitter can be inside FIPS
boundary.

Calls to jent_read_entropy() can return permanent failures for
Repetitive Count Test (RTC), Adaptive Proportion Test (APT), LAG
prediction test.

Ensure the module enters error state upon permanent jitter failures.

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Saša Nedvědický <sashan@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25957)

Fix potential memory leak on failure of ecx_gen_init()

When ecx_gen_set_params() returns 0, it could have duplicated the memory
for the parameter OSSL_KDF_PARAM_PROPERTIES already in gctx->propq,
leading to a memory leak.

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26015)

Fix potential memory leak on failure of dsa_gen_init()

When dsa_gen_set_params()  returns 0, it could have duplicated the memory for the parameter
OSSL_PKEY_PARAM_FFC_DIGEST already in gctx->mdname, leading to a memory leak.

Allocated here: https://github.com/openssl/openssl/blob/47a80fd2034cd4314d3b4958539dcd3106087109/providers/implementations/keymgmt/dsa_kmgmt.c#L524
Can return 0 here: https://github.com/openssl/openssl/blob/47a80fd2034cd4314d3b4958539dcd3106087109/providers/implementations/keymgmt/dsa_kmgmt.c#L529-L536

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/26016)

Enable AES and SHA3 optimisations on Apple Silicon M4-based macOS systems

AES gets a performance enhancement of 7-33%.

Tested on an M4 Pro, but the CPU cores are the same on M4 and M4 Max.

Change-Id: I634c03f1d2b50fa5f8ca97dd65975e49d970c72b

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/25940)

fips: remove redundant RSA encrypt/decrypt KAT

FIPS 140-2 IG D.9 has become FIPS 140-3 D.G (see "Mapping FIPS 140-2
IGs to FIPS 140-3" in the FIPS 140-3 IG).

The requirements w.r.t. RSA KATs have now been relaxed, meaning that
existing full-message RSA signature verification (which is performed
separately) is sufficient to meet KAT requirements for all RSA
usecases (KEM/Encrypt/Decrypt/Sign/Verify).

Dropping this KAT is very useful, because it is large/expensive on
module startup, but also because it enables in the future to block RSA
Encrypt/Decrypt operations with paddings other than OAEP, which are
legacy or deprecated by either current or draft algorithm transition
SP.

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25988)

fips: zeroize temporary self-check out MD variable

At least this is done on module startup only.

To satisfy ISO/IEC 19790:2012/Cor.1:2015(E) Section 7.5 [05.10]
requirement.

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25945)

.gitignore: Ignoring demo executables

Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26009)

Remove the negative return values of EVP_KDF_* functions in documents

None of the EVP_KDF_* functions will ever return a negative value.

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25811)

Fix wrong return value checks for some functions

- in particular in use of X509_LOOKUP_load_file, EVP_PKEY_print_params,
EVP_PKEY_keygen, X509_CRL_add1_ext_i2d, EVP_PKEY_keygen_init

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25811)

Fix multiple wrong use of BN_check_prime

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25811)

Fix multiple wrong checks of EVP_PKEY_set1_encoded_public_key

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25811)

apps/lib/apps.c: fix the wrong check in check_cert_attributes

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25811)

Mark OPENSSL_armcap_P .hidden in arm asm

Fixes #25601
Fixes #22414

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22181)

Fix SSL_write_[ex|ex2] on blocking quic streams

When writing to a blocking quic stream, we sometimes get duplicate
transmitted data.  This occurs when a call to quic_write_blocking has to
wait for space to become available in the ring buffer.  When we do a
wait, the call sets *written to the value returned in args.total_written
as filled out by the calls to block_until_pred->quic_write_again.
However, the value there is based on the amount we requested, which is
only the remaining data that we didn't append in xso_sstream_write.  So
if we call quic_write_blocking with a buffer of length X, and initially
append Y bytes, and write the remainig X-Y bytes via a block_until_pred
call, then *written will return with the value X-Y, even though we wrote
the full X bytes to the ring buffer.

Fix it by recording the initial amount appended into *written, and then
add the args.total_written value if we have to wait on more space

Fixes openssl/project#924

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26023)

sm2_sig_verify(): Do not call BN_CTX_end() without BN_CTX_start()

In case of memory allocation failure this
could happen.

Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/25994)

add new keytype test for EVP_PKEY_Q_keygen

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25972)

pkeyutl.c: Avoid freeing pkey at multiple places

Also fixes a leak of pkey in error case for -verifyrecover.

Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/25987)

README.md: Fix typo. Change 'the are' to 'there are'

This pull request fixes a typo in the documentation.
The phrase "the are" has been corrected to "there are".

CLA: trivial

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25977)

s_cb.c: Move the negotiated group output

It needs to be always displayed not just with -brief.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25959)

Add CTX copy function for EVP_MD to optimize the performance of EVP_MD_CTX_copy_ex.

1. Add OSSL_FUNC_digest_copyctx_fn function for EVP_MD, which is used to copy algctx from the old EVP_MD_CTX to the new one.

2. Add implementation of OSSL_FUNC_digest_copyctx_fn function for default providers.

3. Modify EVP_MD_CTX_copy_ex: When the fetched digest is the same in in and out contexts, use the copy function to copy the members in EVP_MD_CTX if the OSSL_FUNC_digest_copyctx_fn function exists. Otherwise, use the previous method to copy.

4. Add documentation for OSSL_FUNC_digest_copyctx function in doc/man7/provider-digest.pod.

5. Add testcase.

Fixes #25703

Signed-off-by: wangcheng <bangwangnj@163.com>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25726)

x509_vfy.c and x509_lu.c: refactor find_issuer(), X509_STORE_CTX_get1_issuer(), etc.

Reviewed-by: Hugo Landau <hlandau@devever.net>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/18762)

os-zoo CI: Replace macos-12 run with macos-15

macos-12 runners will be removed in December.

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Hugo Landau <hlandau@devever.net>
(Merged from https://github.com/openssl/openssl/pull/25715)

Add docs for OSSL_PARAM_print_to_buf

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25630)

Make ossl_trace_param_values an official api function

lots of people may want to print params to a buffer.  Make it part of
our api

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25630)

Add QUERY trace points

Adds trace messages for method store add/remove and fetch operations

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25630)

Add a QUERY trace category

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25630)

Add tracing messages to the provider api

Adds tracing messages to the
init/teardown/gettable_params/get_params/query/unquery operations for a
provider

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25630)

Adding a trace category for PROVIDER calls

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25630)

remove ternary from runs-on, security repo is hosted on github.com

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25939)

man ECDSA_sign(3): fix wrong variable names

the parameters in the function definitions use `siglen` not `sig_len`,
this fixes the doc text.

Signed-off-by: Alicja Kario <hkario@redhat.com>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25927)

Change "a SSL" to "an SSL"

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
(Merged from https://github.com/openssl/openssl/pull/25868)

test_hashtable_multithread(): Stress it more

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Hugo Landau <hlandau@devever.net>
(Merged from https://github.com/openssl/openssl/pull/25669)

Use test_random() in hashtable mt test for reproducibility

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Hugo Landau <hlandau@devever.net>
(Merged from https://github.com/openssl/openssl/pull/25669)

Fix doc and use of_X509v3_add_extensions() in case sk_X509_EXTENSION_num(exts) <= 0

Reviewed-by: Hugo Landau <hlandau@devever.net>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25631)

Add test for SSL_stream_reset

Add a test to the quic_multistream test suite to reset a stream after
all data has been received by a given stream, ensuring that we don't
crash in the reset operation

Fixes #25410

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25910)

Fix SSL_stream_reset for stream objects which have FIN bit set

When calling SSL_stream_reset on a QUIC stream object that has received
all data that is expected to be sent (i.e. when the sender has sent a
STREAM frame with the FIN bit set), we encounter the following segfault:

Program received signal SIGSEGV, Segmentation fault.
0x00007ffff7f0bd28 in ossl_quic_sstream_get_final_size (qss=0x0, final_size=0x0) at ssl/quic/quic_sstream.c:273
273     if (!qss->have_final_size)
(gdb) bt
0)  0x00007ffff7f0bd28 in ossl_quic_sstream_get_final_size (qss=0x0, final_size=0x0) at ssl/quic/quic_sstream.c:273
1)  0x00007ffff7ef65bf in quic_validate_for_write (xso=0x5555555efcb0, err=0x7fffffffd5e0) at ssl/quic/quic_impl.c:2513
2)  0x00007ffff7ef8ae3 in ossl_quic_stream_reset (ssl=0x5555555efcb0, args=0x0, args_len=0) at ssl/quic/quic_impl.c:3657
3)  0x00007ffff7ebdaa6 in SSL_stream_reset (s=0x5555555efcb0, args=0x0, args_len=0) at ssl/ssl_lib.c:7635
4)  0x0000555555557527 in build_request_set (
    req_list=0x55555555ebd0 "neil1.txt neil2.txt neil3.txt neil4.txt neil5.txt neil6.txt neil7.txt neil8.txt neil9.txt neil10.txt neil11.txt neil12.txt neil13.txt neil14.txt neil15.txt neil16.txt neil17.txt neil18.txt neil19.txt "..., ssl=0x5555555b6f80)
    at demos/guide/quic-hq-interop.c:545
5)  0x00005555555587b2 in main (argc=4, argv=0x7fffffffe568) at demos/guide/quic-hq-interop.c:941

This occurs because:
1) When the stream FIN bit is set, the quic stack frees the underlying
   stream structures immediately within the QUIC stack
and
2) when SSL_stream_reset is called, the call stack indicates we call
   quic_validate_for_write, which attempts to access the
   xso->stream->sstream QUIC_SSTREAM object, which was already freed in
   (1)

The fix I think is pretty straightforward.  On receipt of a STREAM frame
with a FIN bit set, the QUIC stack sets the QUIC_STREAM object state to
QUIC_SSTREAM_STATE_DATA_RECVD, which means we can use that state to
simply assert that the stream is valid for write, which allows it to be
reset properly.

Fixes #25410

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25910)

Adjustments for the on schedule workflows

Run them all after 02:00 UTC.
Add possibility to run them on workflow_dispatch.
Add branch 3.4 to the coveralls.yml.
Remove the branches from os-zoo.yml as it is
possible to run on them manually from workflow_dispatch.

Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
(Merged from https://github.com/openssl/openssl/pull/25765)