Coverity 1528496: remove assignment of unused value

ctx is used only within the loop and always assigned at start

Coverity 1529992: Check return value of sscanf()

Also moving the call to setup_tests() where it
fits better.

Coverity 1531836: Check return value of CRYPTO_atomic_add()

Coverity 1531872: j is not used anywhere later, remove the assignment

Include poll.h instead of incorrect sys/poll.h

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Anton Arapov <anton@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21131)

Fix failures of OS Zoo CI

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Anton Arapov <anton@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21131)

Set RC4 defines on libcrypto/liblegacy

Also add missing prototype for rc4_md5_enc.

Fixes #21150

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21153)

Cast the argument to unsigned char when calling isspace()

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21151)

util/find-doc-nits: extend regex to match new OPT_INFORM A

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/7320)

apps/asn1parse: improve RFC7462 compliance

The asn1parse command now supports three different input formats:

     openssl asn1parse -inform PEM|DER|B64

       PEM: base64 encoded data enclosed by PEM markers (RFC7462)
       DER: der encoded binary data
       B64: raw base64 encoded data

The PEM input format is the default format. It is equivalent
to the former `-strictpem` option which is now marked obsolete
and kept for backward compatibility only.

The B64 is equivalent to the former default input format of the
asn1parse command (without `-strictpem`)

Fixes #7317

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/7320)

apps/opt: refactor input format parsing

- split OPT_FMT_PEMDER flag into OPT_FMT_PEM and OPT_FMT_DER
- add OPT_FMT_B64 option (`-inform b64`)

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/7320)

Optimize CMAC_Update for better performance.

    Reduce the number of EVP_Cipher function calls in CMAC_Update,
    to improve performance of CMAC.
    Below are command and result of performance improvement.

    COMMAND: openssl speed -cmac ALGORITHM

    IMPROVEMENT(%):
    A72   stands for Cortex A72
    N1    stands for Neoverse N1
    N2    stands for Neoverse N2
                        A72 N1 N2 x86
    aes-128-cbc@256 65.4 54.6 37.9 86.6
    aes-128-cbc@1024 156.0 105.6 65.8 197.1
    aes-128-cbc@8192 237.7 139.2 80.5 285.8
    aes-128-cbc@16384 249.1 143.5 82.2 294.1
    aes-192-cbc@256 65.6 46.5 30.9 77.8
    aes-192-cbc@1024 154.2 87.5 50.8 167.4
    aes-192-cbc@8192 226.5 117.0 60.5 231.7
    aes-192-cbc@16384 236.3 120.1 61.7 238.4
    aes-256-cbc@256 66.0 40.3 22.2 69.5
    aes-256-cbc@1024 136.8 74.6 35.7 142.2
    aes-256-cbc@8192 189.7 93.5 41.5 191.7
    aes-256-cbc@16384 196.6 95.8 42.2 195.9
    des-ede3-cbc@64 6.9 4.4 2.9 7.2
    des-ede3-cbc@256 9.3 6.1 4.3 13.1
    des-ede3-cbc@1024 10.0 6.4 4.8 14.9
    des-ede3-cbc@8192 10.3 6.5 5.1 15.5
    des-ede3-cbc@16384 10.3 6.4 5.1 15.5
    sm4-cbc@256 9.5 3.0 - 18.0
    sm4-cbc@1024 12.3 3.6 - 24.6
    sm4-cbc@8192 13.2 3.8 - 27.0
    sm4-cbc@16384 13.5 3.8 - 27.2

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21102)

Update CMAC cipher algorithm list and test cases.

    1. Update manual, add SM4-CBC to CMAC cipher algorithm list.
    2. Add test case for SM4-CBC CMAC, add "data length is greater
       than 4 block-length" cases for aes-128-cbc, aes-192-cbc,
       aes-256-cbc and des-ede3-cbc.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21102)

Add 3.0.9 to list of FIPS releases

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21140)

Update versions tested to include 3.1.1

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21140)

kdf test: restrict the version of the FIPS provider

Concatenation tests are provider version specific, limit them to supporting
versions.

Fixes #21134

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21140)

Workaround false positive warning of MSAN in eng_rdrand.c

Reviewed-by: Paul Dale <pauli@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/21136)

Bump coverallsapp/github-action from 2.1.2 to 2.2.0

Bumps [coverallsapp/github-action](https://github.com/coverallsapp/github-action) from 2.1.2 to 2.2.0.
- [Release notes](https://github.com/coverallsapp/github-action/releases)
- [Commits](https://github.com/coverallsapp/github-action/compare/v2.1.2...v2.2.0)

---
updated-dependencies:
- dependency-name: coverallsapp/github-action
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>
CLA: trivial

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21147)

build.info: Introduce special syntax for dependencies on script modules

The DEPEND statement, when applied on files generated with GENERATE, may
be used to specify script modules that the template to be generated from
depends on.  In short, this sort of depend:

    DEPEND[generated]=util/perl/OpenSSL/something.pm

... would generate a perl run that has the inclusion directory
'util/perl/OpenSSL' and 'something' as the module to be loaded.  However,
the package name for this module is 'OpenSSL::something', so to load it the
way it's expected, the inclusion directory should be 'util/perl', and the
module to be loaded should be specified as 'OpenSSL/something' (to be
massaged into a proper module name by the build file template).

To allow this, we introduce a file syntax, where a single '|' is used as a
directory separator, to delineate what part should be used as the inclustion
directory, and which part the module name to be loaded should be derived
from:

    DEPEND[generated]=util/perl|OpenSSL/something.pm

Fixes #21112

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21117)

Modify ENGINE_pkey_asn1_find_str() to use a read lock instead of a write

ENGINE_pkey_asn1_find_str() does not make any modifications to fields
controlled by the global_engine_lock. The only change made is the struct_ref
field which is controlled separately. Therefore we can afford to only take
a read lock. This also impacts EVP_PKEY_asn1_find_str().

This lock ends up being obtained indirectly from numerous public API
functions including EVP_PKEY_key_gen(), EVP_PKEY_new_raw_public_key_ex(),
EVP_PKEY_copy_parameters() etc. This occurs even if no engines are actually
in use.

Some tests showed this lock being obtained 6 times after a "warmed up"
s_server instance with default configuration processed a handshake from a
default s_client. When processing a resumption handshake from s_client it
was obtained 8 times.

Partially fixes #20286

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20950)

Convert the ENGINE struct_ref field to be an atomic

We use atomic primitives to up ref and down the struct_ref field rather
than relying on the global lock for this.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20950)

Add SSL_get0_group_name() to  get name of the group used for KEX

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20866)

Cast the argument to unsigned char when calling isdigit()

Fixes #21123

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21127)

Make link to RFC 1578 in CHANGES.md be a proper link

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21130)

Restrict the size of OBJECT IDENTIFIERs that OBJ_obj2txt will translate

OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
numeric text form.  For gigantic sub-identifiers, this would take a very
long time, the time complexity being O(n^2) where n is the size of that
sub-identifier.

To mitigate this, a restriction on the size that OBJ_obj2txt() will
translate to canonical numeric text form is added, based on RFC 2578
(STD 58), which says this:

> 3.5. OBJECT IDENTIFIER values
>
> An OBJECT IDENTIFIER value is an ordered list of non-negative numbers.
> For the SMIv2, each number in the list is referred to as a sub-identifier,
> there are at most 128 sub-identifiers in a value, and each sub-identifier
> has a maximum value of 2^32-1 (4294967295 decimal).

Fixes otc/security#96
Fixes CVE-2023-2650

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

allow to disable http

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21108)

doc: note that out ChaCha20 isn't standard compliant.

Fixes #21095

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/21098)

Don't take a write lock to retrieve a value from a stack

ossl_x509_store_ctx_get_by_subject() was taking a write lock for the
store, but was only (usually) retrieving a value from the stack of
objects. We take a read lock instead.

Partially fixes #20286

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20952)

Avoid an unneccessary lock if we didn't add anything to the store

Partially fixes #20286

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20952)

Remove pointless warning on pkcs12 import

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21107)

Update fuzz/corpora submodule to latest data

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21118)

Fix documentation where openssl-genrsa is listed as
deprecated since OpenSSL 3.0

openssl-genrsa is not deprecated however the OpenSSL documentation
states that it is the case from OpenSSL 3.0. This has been fixed in the
documentation, specifically in manpage 1.

Fixes #21055

CLA: trivial

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21059)

possible workaround

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20935)

Update .gitignore

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20935)

gcm: use the new faster param location mechanism.

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20935)

params: provide a faster TRIE based param lookup.

The separate file is a Perl script that generates the appropriate define
directives for inclusion in core_names.h.  By having this separation it
will be possible to prebuild data structures to give faster access when
looking up parameters by name.

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20935)

[feat] SSL RTT in both client and server statem. SSL_get_handshake_rtt makes it available

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/20248)

Avoid taking a write lock in ossl_provider_doall_activated()

We refactor ossl_provider_doall_activated() so that we only need to take
a read lock instead of a write lock for the flag_lock. This should improve
performance by avoiding the lock contention. We achieve this by protecting
the activatecnt via atomics rather than via a lock and by avoiding the full
provider activation/deactivation procedure where it is not needed.

Partial fix for #20286

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20927)

Enable obtaining certain DRBG params without a lock

Even if a DRBG has locking enabled on it, there are certain parameters
which are still safe to obtain even without a lock. The max_request
value is constant for all our DRBGs. The reseed_counter does not matter
if we get it wrong - so it is safe to avoid the lock. So if all we are
reading are those parameters then we take no lock at all.

Partially fixes #20286

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20970)

Refactor the DRBG implementations to manage locking themselves

Previously the EVP layer would call lock and unlock functions on the
underlying DRBG implementation to say when a lock should be acquired and
released. This gives the DRBG implementation no say as to what kind of
lock should obtained (e.g. read/write) or even whether a lock is actually
needed or not.

In reality we know whether a DRBG is supposed to be in locking mode or
not because the "enable_locking()" function will have been called if
locks should be used. Therefore we re-interpret the lock and unlock
functions as "hints" from the EVP layer which we ignore. Instead we
acquire locks only when we need them. By knowing the context we can obtain
either a read or a write lock as appropriate.

This may mean that in some rare cases we acquire the locks more than once
for a single EVP call, if the EVP call makes several calls to the underlying
DRBG. But in practice almost all EVP calls only make one such call.
EVP_RAND_generate() is an example of a call where multiple DRBG calls may
be made. One of these gets the "max_request" parameter (which is constant
for all of our own DRBGs) and it may make several calls to the DRBG generate
call - but only if the requested size is very large which will rarely be
the case.

Partially fixes #20286

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20970)

openssl-cmp.pod.in: tweak doc of -subject, -issuer, -keep_alive, and -untrusted

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/21086)

cmp_mock_srv.c: improve comment on cert to be produced from request template

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/21086)

cmp_client.c: add comment on certConf and add 'ossl_unused' to two functions

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/21086)

apps.c: add comment to do_X509_sign() referring to question #19805

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/21086)

apps/cmp.c: make management of http_cb_arg pointer more robust

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/21086)

apps/cmp.c: improve warnings on option use

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/21086)

CMS_ContentInfo_free(): fix mem leak on encrypted content key

Fixes #21026

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/21058)

SMIME_crlf_copy(): check for NULL pointer arguments

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/21058)

Ignore the fetch error when a legacy algorithm is found

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/21035)

(cherry picked from commit cb1f87f59d0b290eca9207ec4ef36d8073e6feec)

CMP: add support for genm/genp messages with id-it-caCerts

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/19231)

QUIC: CID conformance

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21078)

doc: update FIPS provider version information

With 3.0.8 validated, we need to note this in the documentation.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/21049)

tags: fix `make tags` target for out of tree builds

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/21079)

Add PEM fuzzer

This fuzzer can find CVE-2022-4450

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20242)

Compute RSA-PSS algorithm params in libcrypto for legacy

Fixes regression of RSA signatures for legacy keys caused
by quering the provider for the algorithm id with parameters.

Legacy keys do not have a method that would create the
algorithm id. So we revert to what was done in 3.0.7 and
earlier versions for these keys.

Fixes #21008

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21019)

Fix arm64 asm code back compatible issue with gcc 4.9.4

Fix: #20963

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20967)

Bump actions/setup-python from 4.6.0 to 4.6.1

Bumps [actions/setup-python](https://github.com/actions/setup-python) from 4.6.0 to 4.6.1.
- [Release notes](https://github.com/actions/setup-python/releases)
- [Commits](https://github.com/actions/setup-python/compare/v4.6.0...v4.6.1)

---
updated-dependencies:
- dependency-name: actions/setup-python
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <support@github.com>
CLA: trivial

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21047)

CMS, PKCS7, and CRMF: simplify use of EVP_PKEY_decrypt() by helper function

Also remove needless constant_time_* and ERR_clear_error() calls
from OSSL_CRMF_ENCRYPTEDVALUE_get1_encCert().

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/17354)

ssl/statem_srvr.c: clean up handling of EVP_PKEY_decrypt() outlen

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/17354)

cmp_http.c: Remove obsolete comment w.r.t. ERR_clear_error()

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/17354)

OSSL_sleep(): Calling sleep() function if sleepTime > 1sec

On some systems usleep() func does not support time >1sec.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21010)

Optimise locking in rsa_get_blinding()

We optimise locking in rsa_get_blinding() so that we normally take a
read lock, and only fallback to a write lock if we need to. This will
be very slightly slower in the case of single use RSA objects, but should
be significantly better when an RSA object is reused in a multi-threaded
environment. It's probably worth the trade off.

Partially fixes #20286

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20953)

When we're just reading EX_CALLBACK data just get a read lock

The crypto_ex_data code was always obtaining a write lock in all functions
regardless of whether we were only reading EX_CALLBACK data or actually
changing it. Changes to the EX_CALLBACK data are rare, with many reads so
we should change to a read lock where we can.

We hit this every time we create or free any object that can have ex_data
associated with it (e.g. BIOs, SSL, etc)

Partially fixes #20286

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20943)

Avoid taking a write lock in RAND_get_rand_method()

The function RAND_get_rand_method() is called every time RAND_bytes() or
RAND_priv_bytes() is called. We were obtaining a write lock in order to
find the default random method - even though we rarely write. We change
this to a read lock and only fallback to a write lock if we need to.

Partial fix for #20286

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20929)

Don't take a write lock when freeing an EVP_PKEY

When freeing the last reference to an EVP_PKEY there is no point in
taking the lock for the key. It is the last reference and is being freed
so must only be being used by a single thread.

This should not have been the source of any contention so its unclear to
what extent this will improve performance. But we should not be locking
when we don't need to.

Partially fixes #20286

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20932)

x509: Handle ossl_policy_level_add_node errors

The invocation of ossl_policy_level_add_node in tree_calculate_user_set
did not have any error handling. Add it to prevent a memory leak for the
allocated extra policy data.

Also add error handling to sk_X509_POLICY_NODE_push to ensure that if
a new node was allocated, but could not be added to the stack, it is
freed correctly.

Fix error handling if tree->user_policies cannot be allocated by
returning 0, indicating failure, rather than 1.

Signed-off-by: Clemens Lang <cllang@redhat.com>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21040)

x509: Fix possible use-after-free when OOM

ossl_policy_level_add_node() first adds the new node to the level->nodes
stack, and then attempts to add extra data if extra_data is true. If
memory allocation or adding the extra data to tree->extra_data fails,
the allocated node (that has already been added to the level->nodes
stack) is freed using ossl_policy_node_free(), which leads to
a potential use after free.

Additionally, the tree's node count and the parent's child count would
not be updated, despite the new node being added.

Fix this by either performing the function's purpose completely, or not
at all by reverting the changes on error.

Signed-off-by: Clemens Lang <cllang@redhat.com>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21040)

Add missing CHANGES.md entries

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21057)

(cherry picked from commit 3afa246520852ba3bfc16c7c976cb3c2064a3baa)

SSL_handle_events(): Minor fixes to documentation

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20879)

QUIC: Rename SSL_tick etc. in s_client

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20879)

QUIC: Rename SSL_tick etc. in man(7) docs

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20879)

QUIC: Update documentation to reflect compatibility of DTLS APIs

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20879)

QUIC APL: Allow DTLSv1 APIs to be used for compatibility

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20879)

QUIC: Update documentation for SSL_get_event_timeout

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20879)

QUIC APL: Change SSL_get_event_timeout API design

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20879)

QUIC: Rename SSL_tick, SSL_get_tick_timeout

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20879)

QUIC Documentation: Rename SSL_tick, SSL_get_tick_timeout

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20879)

Replace __attribute__((malloc)) with __attribute__((__malloc__)) in macros.h

Fix macro attribute conflict with cmocka
Fixes #20776

CLA: trivial

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20869)

Clarify how to return string data

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21004)

Fix a bug where the result of rehash is unstable

The root cause is that the file entries targeted for rehash are not actually sorted.
Sort was skipped because the compare function was null.
So a compare function has been implemented to allow file entries to be sorted.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21013)

APPS: replace awkward and error-prone pattern by calls to new app_conf_try_number()

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/20971)

APPS: replace awkward and error-prone pattern by calls to new app_conf_try_string()

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/20971)

APPS/ca: remove spurious errors when certain config file entries are not provided

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/20971)

Create internal/ssl.h

We create the internal/ssl.h header file and move the typedef for
ossl_msg_cb. This is needed by both the QUIC code (which generally doesn't
include ssl_local.h) and the rest of libssl.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Rename msg_callback_s to msg_callback_ssl for greater clarity

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Use the %llu format specifier for uint64_t

We change to use %llu when printing uint64_t types for consistency with
what we've done elsewhere.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Create setter functions for the msg_callback and msg_callback_arg

We create setter functions for the msg_callback and msg_callback_arg so
that these values can be properly propagated to the QRX/QTX/TXP even
after the channel has been created.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Add a test for the new QUIC tracing capability

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Fix an SSL_trace bug

Ensure that SSL_trace can print certificate data even with a non-default
libctx.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Update the msg_callback documentation

We provide information about the new QUIC support related to the
msg_callback. We also document SSL_trace() which was previously missing
from the man pages.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Add more trace details for the remaining frame types

Prior to this commit we were just printing the fact that we had received
or sent a frame of a particular type. We now provide more details about
those frames.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Properly handling stream/crypto frames while tracing

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Enable tracing of datagrams we have sent

Extend the tracing capability to also trace when we have a datagram to the
peer.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Enable tracing of packets that have been sent

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Extend tracing of frames to transmitted frames

Previously we were only doing tracing of frames received from the peer.
Now we do that for transmitted frames as well.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Support trace for QUIC Frames

Extend the existing QUIC tracing capability for frames.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Support trace for QUIC Packets

We enable SSL_trace support for when we receive QUIC Packets. This is
called after header protection is removed, but before the packet is
decrypted.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Support trace for QUIC datagrams

Extend SSL_trace so that it knows how to dump information about the
receipt of a QUIC datagram.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Add initial QUIC support for the msg_callback

At this stage we just support msg_callback on receipt of a datagram.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20914)

Resolve a djgpp function name conflict

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20856)