ec/ecp_nistp{224,256,521}.c: harmonize usage of __uint128_t.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5449)

{ec/curve25519,poly1305/poly1305}.c: relax pedantic constraint.

As it turns out gcc -pedantic doesn't seem to consider __uint128_t
as non-standard, unlike __int128 that is.

Fix even MSVC warnings in curve25519.c.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5449)

ec/curve448: portability fixups.

SPARC condition in __SIZEOF_INT128__==16 is rather performance thing
than portability. Even though compiler advertises int128 capability,
corresponding operations are inefficient, because they are not
directly backed by instruction set.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5449)

ec/asm/x25519-x86_64.pl: fix up ADCX/ADOX fallback.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5447)

Remove OSSLzu, and fix the one place that used it.

Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5439)

Fix some bugs with the cfb1 bitsize handling

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5426)

Remove executable bit from test/recipes/03-test_internal_sm4.t

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5434)

Add a note in CHANGES

[extended tests]

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2688)

Test the storeutl searching options

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2688)

STORE: Add documentation on search criteria

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2688)

Adapt storeutl to allow search for specific objects

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2688)

STORE 'file' scheme loader: Add search capibility

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2688)

STORE: Add the possibility to search for specific information

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2688)

Test the storeutl expectation options

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2688)

STORE: Add documentation on expecting specific infos

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2688)

Adapt storeutl to allow looking for a specific info type

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2688)

STORE 'file' scheme loader: Add info type expectation

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2688)

STORE: Add the possibility to specify an expected info type

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2688)

STORE: In preparation for coming work, mark when loading is started

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2688)

Remove unreachable statement.

The return at the end isn't reachable.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5442)

Build files: when using $(CPP), use the C flags alongside the CPP flags

The reason for this is that some of the C flags affect built in macros
that we may depend on.

Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5436)

Remove unused num.pl,segregnam scripts

Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5433)

initialise dc variable to satisfy old compilers.

When compiling with -Wall on a machine with an old compiler it gives a false
positive that the dc variable which is a structure of type DISPLAY_COLUMNS
could be used uninitialised. In fact the dc variable's members will always get
set in the case it is used, otherwise it is left uninitialised.
This fix just causes the dc variable's members to always get initialised to 0
at declaration, so the false positive will not get flagged.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5337)

Update EC_POINT_new.pod

CLA: trivial
fix typo:
EC_point2buf => EC_POINT_point2buf

Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5367)

Build file templates: be less verbose when reconfiguring

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5247)

Replace the message about configdata.pm as a script

The new message is geared toward issue reports

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5247)

Add a comment in configdata.pm regarding script-only variables

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5247)

Make CROSS_COMPILE another supported "make variable" in Configure

No more special casing for that one, and this means it gets displayed
by 'perl configdata.pm --make-variables' among all the others.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5247)

Unix Makefile: Rework the assignment of CXX and AS

If the configured value is the empty string, give them a sane default.
Otherwise, give them the configured value prefix with $(CROSS_COMPILE)

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5247)

Refactor the ranlib attribute

It was inconsistent to see this specific command have
'$(CROSS_COMPILE)' in its value when no other command did.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5247)

Have configdata.pm display information on perl

Specifically, the specific perl that was used to run Configure

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5247)

Check return value of time() when getting additional data for the DRBG

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
GH: #5400

Switch the DRBGs from AES-128-CTR to AES-256-CTR

Reviewed-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
GH: #5401

Use malloc to avoid alignment problems.

Thanks to Norm Green for reporting this issue.

Signed-off-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5423)

Add support for PBKDF2 for enc command

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2083)

do_body: fix heap-use-after-free.

The memory pointed to by the 'push' is freed by the
X509_NAME_ENTRY_free() in do_body(). The second time
it is referenced to (indirectly) in certify_cert:X509_REQ_free().

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4698)

X509V3_EXT_add_nconf_sk, X509v3_add_ext: fix errors handling

X509v3_add_ext: free 'sk' if the memory pointed to by it
was malloc-ed inside this function.
X509V3_EXT_add_nconf_sk: return an error if X509v3_add_ext() fails.
This prevents use of a freed memory in do_body:sk_X509_EXTENSION_num().

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4698)

Replaced variable-time GCD with consttime inversion to avoid side-channel attacks on RSA key generation

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/5161)

Sanity check the ticket length before using key name/IV

This could in theory result in an overread - but due to the over allocation
of the underlying buffer does not represent a security issue.

Thanks to Fedor Indutny for reporting this issue.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/5414)

Fix no-ec build

[extended tests]

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5418)

Fix some undefined behaviour in the Curve448 code

We can't add NULL data into a hash

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5418)

Remove a spurious TLSProxy byte in TLSv1.3

When the proxy re-encrypted a TLSv1.3 record it was adding a spurious
byte onto the end. This commit removes that.

The "extra" byte was intended to be the inner content type of the record.
However, TLSProxy was actually adding the original encrypted data into the
record (which already has the inner content type in it) and then adding
the spurious additional content type byte on the end (and adjusting the
record length accordingly).

It is interesting to look at why this didn't cause a failure:

The receiving peer first attempts to decrypt the data. Because this is
TLSProxy we always use a GCM based ciphersuite with a 16 byte tag. When
we decrypt this it actually gets diverted to the ossltest engine. All this
does is go through the motions of encrypting/decrypting but just passes
back the original data. Crucially it will never fail because of a bad tag!
The receiving party thinks the spurious additional byte is part of the
tag and the ossltest engine ignores it.

This means the data that gets passed back to the record layer still has
an additional spurious byte on it - but because the 16 byte tag has been
removed, this is actually the first byte of the original tag. Again
because we are using ossltest engine we aren't actually creating "real"
tags - we only ever emit 16, 0 bytes for the tag. So the spurious
additional byte always has the value 0. The TLSv1.3 spec says that records
can have additional 0 bytes on the end of them - this is "padding". So the
record layer interprets this 0 byte as padding and strips it off to end up
with the originally transmitted record data - which it can now process
successfully.

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5370)

VMS: Fix curve448 internal test program

The internals test programs access header files that aren't guarded by
the public __DECC_INCLUDE_PROLOGUE.H and __DECC_INCLUDE_EPILOGUE.H
files, and therefore have no idea what the naming convention is.
Therefore, we need to specify that explicitely in the internals test
programs, since they aren't built with the same naming convention as
the library they belong with.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5425)

test/recipes/80-test_pkcs12.t: handle lack of Win32::API.

So far check for availability of Win32::API served as implicit check
for $^O being MSWin32. Reportedly it's not safe assumption, and check
for MSWin32 has to be explicit.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5416)

Configure: engage x25519 assembly support.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/5408)

ec/curve25519.c: facilitate assembly implementations.

Currently it's limited to 64-bit platforms only as minimum radix
expected in assembly is 2^51.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/5408)

Add x25519-x86_64.pl module, mod 2^255-19 primitives.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/5408)

crypto/ec/curve25519.c: remove redundant fe[51]_cswap.

3 least significant bits of the input scalar are explicitly cleared,
hence swap variable has fixed value [of zero] upon exit from the loop.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/5408)

Add tests for newly added constant time functions

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Some more cleanups of curve448 code

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

fixup! More style fixes for the curve448 code

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

fixup! Improve readability of f_impl.c and f_impl.h

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Improve readability of f_impl.c and f_impl.h

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Remove unrolled loops

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Yet more style updates to the curve448 code

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Further style changes to curve448 code

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Simplify some code

The original curve448 code was templated to allow for a 25519
implementation. We've just imported the 448 stuff - but a remnant of
the original templated approach remained. This just simplifies that.

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Remove some unneccessary use of constant time code in curve448

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Fix a travis failure in the curve448 code

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Remove the curve448 vector code

We removed various platform specific optimisation files in an earlier
commit. The vector code was related to that and therefore is no longer
required. It may be resurrected at a later point if we reintroduce the
opimtisations.

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

More style fixes for the curve448 code

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Remove a strict aliasing issue with pre-computed curve448 constants

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

More style fixes to Curve448 code based on review feedback

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Some style fixes

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Formatting tweak based on review feedback

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Use NLIMBS where appropriate to simplify the code

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Remove cplusplus guards in internal headers

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Remove the curve448 specific constant time implementation

Instead we should use the standard OpenSSL constant time routines.

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Add some new constant time functions needed by curve448

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Use the NLIMBS macro rather than try and calculate the number of limbs

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Rename a function to avoid a clash

We already have a constant_time_select() function so, to avoid
confusion/clashing we shouldn't have a second one.

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Remove curve448 architecture specific files

Remove all architecture specific files except for the reference arch_32
version. These files provide archicture specific performance optimisation.
However they have not been integrated yet. In order to avoid review issues
they are removed for now. They may be reintroduced at a later time.

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Fix AppVeyor failure in eddsa.c

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Fix travis failure in f_impl.c

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Fix a typo in a comment

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Update Curve448 copyright for 2018

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Code tidy up

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Fixes for compilation using clang

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Update the curve448 internal test to use testutil.h

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Move curve448_test.c to be a full internal test

This ensures that this test is run as part of the test suite

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Fix build errors for Curve448 code on Windows (VC-WIN32 and VC-WIN64A)

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Remove some gcc/clang specific attributes we don't support

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Remove duplicated 448 in the names of various things

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Merge f_field.h into field.h

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Merge f_arithmetic.c into f_generic.c

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Remove references to libdecaf

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Remove some unneeded code

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Manual formatting tweaks to Curve448 code

Following running openssl-format-source there were a lot of manual tweaks
that were requried.

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Run util/openssl-format-source on the Curve448 code

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Update the imported curve448 code to use OpenSSL copyright headers

Some files talk about the MIT license. This code was contributed under
CLA and was relicensed to the OpenSSL licence when imported.

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Fix the ED448 key lengths

Unlike X448 the key lengths for ED448 are 57 bytes (as opposed to 56)

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Integrate Curve448 into the build system

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Convert to C90 from C99

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Rename the decaf files to curve448 files

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Remove portable_endian.h

It is no longer used

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Remove the old shake256 implementation

We have fully converted curve448 to use the OpenSSL shake256 implementation
so we can now remove the old one.

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Convert Curve448 internals to use OpenSSL shake256

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Use OpenSSL shake256

Convert the curve448 test to use the OpenSSL implementation of shake256.

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Add tests for Ed448ph

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)

Add Ed448 tests

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5105)