Added support for ESSCertIDv2

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/771)

Update igetest to use the test framework.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3210)

Convert uses of snprintf to BIO_snprintf

Fixes #2360
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3366)

Fix some error path logic in i2v_AUTHORITY_INFO_ACCESS and i2v_GENERAL_NAME

Fixes #1653 reported by Guido Vranken

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3362)

Fix comment around safari fingerprint check

Fixes #2442

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3367)

Fix URL links in comment

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3365)

Add some man page cross-references

The old/deprecated servername callback should refer back to the
new/preferred early callback mechanism, as well as indicate that
it is superseded by the early callback.

The early callback should also mention the API for turning the
raw cipherlist octets from the client into usable data structures.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3338)

Convert danetest, ssl_test_ctx_test

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3256)

TLS1.3 Padding

Add padding callback for application control
Standard block_size callback
Documentation and tests included
Configuration file/s_client/s_srver option

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3130)

Fix time offset calculation.

ASN1_GENERALIZEDTIME and ASN1_UTCTIME may be specified using offsets,
even though that's not supported within certificates.

To convert the offset time back to GMT, the offsets are supposed to be
subtracted, not added. e.g. 1759-0500 == 2359+0100 == 2259Z.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2654)

Fix a stack smash

It occurs when memory compares are made that are larger
than the on stack temporary buffers (either malloced or supplied).

Rework the test test so it doesn't use a macro with a branch.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3155)

Remove duplicates from clang_devteam_warnings

Since the clang_devteam_warnings are appended to the gcc_devteam_warnings
when strict-warnings are requested, any items present in both the gcc
and clang variables will be duplicated in the cflags used for clang builds.
Remove the extra copy from the clang-specific flags in favor of the
gcc_devteam_warnings that are used for all strict-warnings builds.

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3239)

Address some -Wold-style-declaration warnings

gcc's -Wextra pulls in -Wold-style-declaration, which triggers when a
declaration has a storage-class specifier as a non-initial qualifier.
The ISO C formal grammar requires the storage-class to be the first
component of the declaration, if present.

Seeint as the register storage-class specifier does not really have any effect
anymore with modern compilers, remove it entirely while we're here, instead of
fixing up the order.

Interestingly, the gcc devteam warnings do not pull in -Wextra, though
the clang ones do.

[extended tests]

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3239)

Add -Wextra to gcc devteam warnings

clang already has it; let's flip the switch and deal with the fallout.
Exclude -Wunused-parameter, as we have many places where we keep unused
parameters to conform to a uniform vtable-like interface.
Also exclude -Wmissing-field-initializers; it's okay to rely on
the standard-mandated behavior of filling out with 0/NULL.

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3239)

test/asn1_encode_test.c: test "next negative minimum" corner case.

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

asn1/a_int.c: fix "next negative minimum" corner case in c2i_ibuf.

"Next" refers to negative minimum "next" to one presentable by given
number of bytes. For example, -128 is negative minimum presentable by
one byte, and -256 is "next" one.

Thanks to Kazuki Yamaguchi for report, GH#3339

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

Check fflush on BIO_ctrl call

Bug found and fix suggested by Julian Rüth.
Push error if fflush fails

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3266)

Update the pyca-cryptography submodule to version 1.8.1

It was released a couple of days after our latest update

[extended tests]

Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3346)

Ensure blank lines between tests.

Also add a comment describing the file format.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3337)

Refactor crltest.c to separate the test cases into individual functions.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3327)

testutil: Remove test_puts_std{out,err}, they are superfluous

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3345)

testutil: Add OpenSSL error stack printing wrapper TEST_openssl_errors

Also added a internal error printing callback to be used both with
ERR_print_errors_cb() and with CRYPTO_mem_leaks_cb

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3345)

testutil: Add commodity printing functions test_printf_std{out,err}

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3345)

testutil: make subtest_level() internal

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3345)

testutil: Move printing function declarations to "internal" header

These functions aren't meant to be used directly by the test programs,
reflect that by making the declarations a little harder to reach, but
still available enough if there's a need to override them.

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3345)

Add checks on return code when applying some settings.
Remove hardcoded bound checkings.

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3141)

Output prog name within error message

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3141)

Add a 'max_send_frag' option to configure maximum size of send fragments

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3141)

Fix s_client when no-dtls

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3344)

Fix a pedantic gcc-7 warning.

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3328)

TLSProxy: When in debug mode, show the exact subprocess commands

When you want to debug a test that goes wrong, it's useful to know
exactly what subprocess commands are run.

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3342)

Remove (broken) diagnostic print

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3336)

fuzz/{client,server}.c: omit _time64 "overload method".

Approach was opportunistic in Windows context from its inception
and on top of that it was proven to be error-prone at link stage.
Correct answer is to introduce library-specific time function that
we can control in platform-neutral manner.  Meanwhile we just let
be attempts to override time on Windows.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3320)

Ensure s_client sends an SNI extension by default

Enforcement of an SNI extension in the initial ClientHello is becoming
increasingly common (e.g. see GitHub issue #2580). This commit changes
s_client so that it adds SNI be default, unless explicitly told not to via
the new "-noservername" option.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2614)

Add parentheses on public macros where appropriate.
Fixes #3063.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3100)

Remove unnecessary loop in pkey_rsa_decrypt.

It is not necessary to remove leading zeros here because
RSA_padding_check_PKCS1_OAEP_mgf1 appends them again. As this was not done
in constant time, this might have leaked timing information.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3313)

Fix ISO C function/object pointer issue

Showed up on GCC with strict warnings.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3325)

Convert sslapitest to test framework

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3252)

Convert bntest to TEST_ framework

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3265)

Return success in custom_ext_parse_old_cb_wrap if parse_cb is NULL

This fixes a segfault if a NULL parse_cb is passed to
SSL_CTX_add_{client,server}_custom_ext, which was supported in the
pre-1.1.1 implementation.

This behaviour is consistent with the other custom_ext_*_old_cb_wrap
functions, and with the new SSL_CTX_add_custom_ext function.

CLA: trivial

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3310)

Convert dtls_mtu_test, dtlsv1listentest

Also converted most of ssltestlib but left the packet_dump output
as-is (for now).

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3257)

Fix no-ec

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3321)

Don't treat PACKET_remaining() as boolean

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3244)

Break before && operator

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3244)

Call init and finalization functions per extension message

Previously, init and finalization function for extensions are called
per extension block, rather than per message.  This commit changes
that behaviour, and now they are called per message.  The parse
function is still called per extension block.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3244)

Clarify that SSL_CTX_remove_session() marks a session as non-resumable

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3008)

More SSL_SESSION documentation tweaks based on feedback

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3008)

Tweak SSL_get_session.pod wording

Based on feedback received.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3008)

Always duplicate the session on NewSessionTicket in TLSv1.3

Because NST messages arrive post-handshake, the session may have already
gone into the cache. Once in the cache a session must be immutable -
otherwise you could get multi-thread issues.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3008)

Documentation updates for TLSv1.3 sessions

Add documentation for SSL_SESSION_is_resumable(). Also describe the interaction
of the various session functions and TLSv1.3 post-handshake sessions.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3008)

Add support for SSL_SESSION_is_resumable()

Provide a way to test whether the SSL_SESSION object can be used to resume a
sesion or not.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3008)

Remove TLS1.3 TODO around testing for session id length

TLSv1.3 will do the same thing as TLSv1.2 with tickets with regards to session
ids, i.e. it will create a synthetic session id when the session is established,
so it is reasonable to check the session id length, even in TLSv1.3.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3008)

Fix unit-tests when no-srp configured

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3274)

Convert modular exponentiation tests to new framework

Updated due to test framework changes
Updates after code review
Missed some checks

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3269)

Add tests for version/ciphersuite sanity checks

The previous commits added sanity checks for where the max enabled protocol
version does not have any configured ciphersuites. We should check that we
fail in those circumstances.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3316)

Add a ciphersuite config sanity check for servers

Ensure that there are ciphersuites enabled for the maximum supported
version we will accept in a ClientHello.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3316)

Add a ciphersuite config sanity check for clients

Ensure that there are ciphersuites enabled for the maximum supported
version we are claiming in the ClientHello.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3316)

Don't overwrite the alert value if there is no alert to send

The function tls_early_post_process_client_hello() was overwriting the
passed "al" parameter even if it was successful. The caller of that
function, tls_post_process_client_hello(), sets "al" to a sensible default
(HANDSHAKE_FAILURE), but this was being overwritten to be INTERNAL_ERROR.
The result is a "no shared cipher" error (and probably other similar errors)
were being reported back to the client with an incorrect INTERNAL_ERROR
alert.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3314)

evp_test.c: Add PrivPubKeyPair tests

Reviewed-by: Stephen Henson <steve@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3245)

test: don't make it more complicated than necessary.

Original rationale behind using write in testutil was to accommodate
no-stdio builds. But is there evidence that no-stdio users would have
write or pre-defined meaning for file descriptors 1 and 2? Correct
answer is to provide way for no-stdio users who want to exercise
tests to plug in own BIO, not to make assumption that they have write.
And since we don't have to make such assumption, we can as well go
for simplest that works with standard library as specified by C
language standard.

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

make update

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3301)

Update documentation

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3301)

Add PSS certificate signature tests

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3301)

Add certificates with PSS signatures

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3301)

Add custom sig_info setting for RSA-PSS

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3301)

Use X509_get_signature_info() when checking security levels.

Make signature security level checking more flexible by using
X509_get_signaure_info(): some signature methods (e.g. PSS, ED25519)
do not indicate the signing digest (if any) in the signature OID.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3301)

Add support for custom signature parameters

Many signature types define the digest and public key type by a single OID
such as ecdsa_with_sha256.

Some types (RSA-PSS for example) use a single OID to indicate the signature
scheme and additional parameters are encoded in the AlgorithmIdentifier.

Add an X509_SIG_INFO structure to contain details about the signature type:
specifically the digest algorithm, public key algorithm, security bits and
various flags. This supports both existing algorithms and more complex
types.

Add accessors for the structure and a special case that retrieves signature
information from a certificate.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3301)

Tapify libtestutil a bit better

This includes better signals of skips and subtests according to TAP 12,
and flushing stdout and stderr at the end of every test function to
make sure we get the output in good order.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3309)

Add include path '..' for libtestutil

Since it uses some of the apps/ stuff and some of them include e_os.h...

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3307)

Switch command-line utils to new nameopt API.

The CA names should be printed according to user's decision
print_name instead of set of BIO_printf
dump_cert_text instead of set of BIO_printf
Testing cyrillic output of X509_CRL_print_ex
Write and use X509_CRL_print_ex
Reduce usage of X509_NAME_online
Using X509_REQ_print_ex instead of X509_REQ_print
Fix nameopt processing.
Make dump_cert_text nameopt-friendly
Move nameopt getter/setter to apps/apps.c

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3262)

Ignore all .a files, not just the top ones

Reviewed-by: Rich Salz <rsalz@openssl.org>

Add documentation for the -sctp option in command line apps

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3305)

Correct some badly formated preprocessor lines

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3304)

Add guards around one of use of IPPROTO_SCTP where it was missing

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3304)

TAPify testutil

With the perl test framework comes the output format TAP
(Test Anything Protocol, see http://testanything.org/) with
extra extension for subtests.  This change extends that same
output format to any test program using testutil.

In this implementation, each test program is seen as a full test that
can be used as a subtest.  The perl framework passes on the subtest
level to the test programs with the environment variable
HARNESS_OSSL_LEVEL.  Furthermore, and series of tests added with
ADD_ALL_TESTS is regarded as another subtest level.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3296)

openssl enc: Don't unbuffer stdin

 - unbuffer causes single-byte reads from stdin and poor performance

Fixes #3281
CLA: trivial

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3299)

Fix doc-nits issue

BIO_lookup_ex() should be in the NAME section

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Address review feedback for the SCTP changes

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Fix issue in 18-dtls-renegotiate.conf.in

Don't skip all tests if SCTP is disabled!

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Document BIO_lookup_ex()

We also change the enum type to an int.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Fix some variable references in init_client

We were incorrectly using "res" when we meant "ai"

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Add SCTP testing for 04-client_auth.conf

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Add SCTP testing for 11-dtls_resumption.conf

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Add SCTP testing to 07-dtls-protocol-version.conf

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Add SCTP testing to 18-dtls-renegotiate.conf

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Add SCTP testing to 16-dtls-certstatus.conf

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Add support to test_ssl_new for testing with DTLS over SCTP

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Fix problem with SCTP close_notify alerts

In SCTP the code was only allowing a send of a close_notify alert if the
socket is dry. If the socket isn't dry then it was attempting to save away
the close_notify alert to resend later when it is dry and then it returned
success. However because the application then thinks that the close_notify
alert has been successfully sent it never re-enters the DTLS code to
actually resend the alert. A much simpler solution is to just fail with a
retryable error in the event that the socket isn't dry. That way the
application knows to retry sending the close_notify alert.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Don't attempt to send fragments > max_send_fragment in DTLS

We were allocating the write buffer based on the size of max_send_fragment,
but ignoring it when writing data. We should fragment handshake messages
if they exceed max_send_fragment and reject application data writes that
are too large.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Remove special case code for SCTP reneg handling

There was code existing which attempted to handle the case where application
data is received after a reneg handshake has started in SCTP. In normal DTLS
we just fail the connection if this occurs, so there doesn't seem any reason
to try and work around it for SCTP. In practice it didn't work properly
anyway and is probably a bad idea to start with.

Fixes #3251

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Ask libssl if we should retry not the socket

s_server was asking the underlying socket if it is a retryable error rather
than libssl which has more information.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Add better error logging if SCTP AUTH chunks are not enabled

In order to use SCTP over DTLS we need ACTP AUTH chunks to be enabled in
the kernel.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Add a -sctp option to s_client

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Add a -sctp option to s_server

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Add a BIO_lookup_ex() function

The existing BIO_lookup() wraps a call to getaddrinfo and provides an
abstracted capability to lookup addresses based on socket type and family.
However it provides no ability to lookup based on protocol. Normally,
when dealing with TCP/UDP this is not required. However getaddrinfo (at
least on linux) never returns SCTP addresses unless you specifically ask
for them in the protocol field. Therefore BIO_lookup_ex() is added which
provides the protocol field.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)

Fix typo in OPENSSL_LH_new compat API

CLA: trivial

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3292)

Avoid using BIO streams in bioprinttest.c

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3243)

Adapt all test programs

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3243)

VMS: Make sure to include MAIN from static libraries if needed

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3243)

Refactor the test framework testutil

It's now built as a static library, and greatly simplified for test
programs, which no longer need to include test_main_custom.h or
test_main.h and link with the corresponding object files.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3243)