Clarify the usage of EVP_PKEY_get_raw_[private|public]_key()

EVP_PKEY_get_raw_private_key() and EVP_PKEY_get_raw_public_key() expect
the size of the key buffer to be populated in the |*len| parameter on
entry - but the docs made no mention of this.

Fixes #11245

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11254)

Use .cnf for config files, not .conf

The default is openssl.cnf  The project seems to prefer xxx.conf these
days, but we should use the default convention.

Rename all foo.conf (except for Configurations) to foo.cnf

Fixes #11174

Reviewed-by: Paul Yang <kaishen.yy@antfin.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11176)

Remove unused files

Reviewed-by: Paul Yang <kaishen.yy@antfin.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11176)

fix a minor bug of s_client

CLA: trivial

-CAstore's option should be OPT_CASTORE, instead of OPT_CAFILE
correct also -no-CAstore option from OPT_NOCAPATH to OPT_NOCASTORE

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/11215)

Add CIFuzz action

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11140)

crypto/ec/curve448/eddsa.c: fix EBCDIC platforms

Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11229)

aes-s390x.pl: fix stg offset caused by typo in perlasm

Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11234)

.github/PULL_REQUEST_TEMPLATE.md: Fix link to contributors guide

The file was converted to Markdown and renamed appropriately in
2e07506a12e126894cd820304465162bc0e732b4.

CLA: trivial

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11223)

Add DSA Key validation to default provider

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10912)

Add DH key validation to default provider

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10911)

config, Configure: move the check of removed crypto/ sub-systems

The 'config' script checked for a bunch of crypto/ sub-system
directories, and added 'no-' options if they weren't there.

We move it to 'Configure' in an effort to simplify 'config' for
further work.

Note: this is pretty much a historical thing.  In modern OpenSSL, it's
much simpler to edit the SUBDIRS statement in crypto/build.info.
However, it's been claimed the there are those who still remove some
of these sub-system sources.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11217)

.travis.yml: where it matters, have build and source nesting levels differ

Where we build out of source, the source directory was _srcdist and
the build directory was _build.  That gives the same nesting level for
both, which doesn't quite exercise all aspects of relative back
references from build to source tree.

Changing the build tree to be in _build/tree will challenge back
references a bit more, and ensure a bit more that we got it right.

Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/11186)

Add some missing env var documentation

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11189)

.gitignore: Add /apps/progs.{c,h}

These files were removed from the source tree in
fe909ee4aeb6eb64f6f31a1544c5d3c81c5fe1f1.

CLA: trivial

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11222)

Add Serializers for EC

Provide EC serializers for text, pem and der.

EC parameters use ANS1 'CHOICE' - which means they are more embedded than other parameters used by
other KEY types (which normally have a SEQUENCE at the top level).
For this reason the ANS1_STRING type that was being passed around has been changed to a void so that the
code can still be shared with EC.

The EC serializer only supports named curves currently.

NOTE the serializer code assumes PKCS8 format - if the older encode methods are needed they will need to be
added in another PR. (Probably when deserialization is considered).

EVP_PKEY_key_fromdata_init was changed from using a keypair selection to all bits of a key. A side effect of this was
that the very restrictive checks in the ecx code needed to be relaxed as it was assuming all selection flags were non
optional. As this is not the case for any other key the code has been modified.

Fixed a bug in legacy_ctrl_str_to_params() - "ecdh_cofactor_mode" was being incorrectly converted to the wrong keyname.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11107)

Correct two small documentation issues

The find-doc-nits complains about non-zero word and about missing
line before =head1 which causes build failure.

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11231)

Documenting newly added CMS modification

Documented CMS-related API functions.
Documented flags added to openssl-cms command

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10904)

Implementation of Russian GOST CMS

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10904)

doc: document that 'openssl rand' is cryptographically secure

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11213)

Fix build with clang assembler

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11219)

Add pairwise consistency self tests to asym keygenerators

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10952)

.travis.yml: Remove NOUPDATE support

It was a temporary measure to deal with the fact that util/progs.pl
didn't work right at all times, but that has now been fixed.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11185)

Remove apps/progs.c and apps/progs.h

Since they are generated in build time, there's not need to keep them
in the source tree.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11185)

Configure: Diverse cleanups

There were some remaining old code and comments that don't serve a
purpose any longer.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11185)

Build: Generate apps/progs.c and apps/progs.h in build time

util/progs.pl depends on the build tree (on configdata.pm,
specifically), so it needs to be run from the build tree.  But why
stop there?  We might as well generate apps/progs.c and apps/progs.h
when building.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11185)

build.info: Implement simply substitutions in variable values

Use case: having a variable with multiple source files in its value,
and wanting to refer to the corresponding object file.

    $SRCS=foo.c bar.c
    SOURCE[program]=$SRCS
    DEPEND[${SRCS/.c/.o}]=prog.h

    GENERATE[prog.h]=...

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11185)

build.info: Make it possible to have more than one item in KEYWORD[]

So far, the "index" part of KEYWORD[whatever] could only handle one
item.  There are cases, however, where we want to add the exact same
value to multiple items.  This is especially helpful if a variable
that may have multi-item values are used in the "index" part.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11185)

PROV: Add a OP_keymgmt_match() function to our DH, DSA, RSA and EC_KEY impl

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11158)

EVP: Add support for copying provided EVP_PKEYs

This adds evp_keymgmt_util_copy() and affects EVP_PKEY_copy_parameters()

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11158)

KEYMGMT: Add a keydata copy function

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11158)

EVP: Add support for comparing provided EVP_PKEYs

This adds evp_keymgmt_util_match() and affects EVP_PKEY_cmp() and
EVP_PKEY_cmp_parameters().

The word 'match' was used for the new routines because many associate
'cmp' with comparison functions that allows sorting, i.e. return -1, 0
or 1 depending on the order in which the two compared elements should
be sorted.  EVP_PKEY_cmp() and EVP_PKEY_cmp_parameters() don't quite
do that.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11158)

KEYMGMT: Add a keydata matching function

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11158)

EVP: Adapt EVP_PKEY_missing_parameters() for provider keys

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11158)

crypto/perlasm/x86_64-xlate.pl: detect GNU as to deal with quirks

It turns out that GNU as and Solaris as don't have compatible ideas on
the .section syntax, so we need to check if we're using GNU as or
another assembler and adapt this .section syntax accordingly.

Fixes #11132

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11191)

Fix drop of const qualifier

The parameter got "const" in 9fdcc21fdc9 but that was not added
to cast. So this throws a -Wcast-qual in user code.

error: cast from 'const DUMMY *' to 'ASN1_VALUE_st *' drops const qualifier [-Werror,-Wcast-qual]

CLA: trivial

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11210)

DOCS: Add and modify docs for internal EVP_KEYMGMT utility functions

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11148)

DOCS: Add internal docs for EVP_PKEY and the export functions

Functions covered:

- evp_pkey_export_to_provider()
- evp_pkey_upgrade_to_provider()

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11148)

EVP: Add evp_pkey_upgrade_to_provider(), for EVP_PKEY upgrades

This function "upgrades" a key from a legacy key container to a
provider side key container.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11148)

Rethink the EVP_PKEY cache of provider side keys

The role of this cache was two-fold:

1.  It was a cache of key copies exported to providers with which an
    operation was initiated.
2.  If the EVP_PKEY didn't have a legacy key, item 0 of the cache was
    the corresponding provider side origin, while the rest was the
    actual cache.

This dual role for item 0 made the code a bit confusing, so we now
make a separate keymgmt / keydata pair outside of that cache, which is
the provider side "origin" key.

A hard rule is that an EVP_PKEY cannot hold a legacy "origin" and a
provider side "origin" at the same time.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11148)

man: openssl-ocsp: separate client and server options

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11033)

x509v3 subjectSignTool extention support

Subject Sign Tool (1.2.643.100.111) The name of the tool used to signs the subject (UTF8String)
This extention is required to obtain the status of a qualified certificate at Russian Federation.
RFC-style description is available here: https://tools.ietf.org/html/draft-deremin-rfc4491-bis-04#section-5
Russian Federal Law 63 "Digital Sign" is available here:  http://www.consultant.ru/document/cons_doc_LAW_112701/

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/11093)

Fix util/mktar.sh to use the new VERSION information

Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/11190)

Fix comment placement in ecp_nistp256.ci

CLA: trivial

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/11175)

Deprecate ASN1_sign(), ASN1_verify() and ASN1_digest()

These are old functions that fell out of use with OpenSL 0.9.7.
It's more than time to deprecate them.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11161)

Implement the ECX Serializers

Provide serializers for X25519 and X448 for text, pem and der. There are
no parameter serializers because there are no parameters for these
algorithms.

Add some documentation about the various import/export types available
Add additional testing for the serializers

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11095)

Replace util/shlib_wrap.sh with util/wrap.pl in diverse docs

Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/11110)

Build file templates: don't set OPENSSL_{ENGINES,MODULES}

Since we've now switched to use util/wrap.pl to wrap uninstalled
programs everywhere, there's no need to set the environment variables
OPENSSL_ENGINES and OPENSSL_MODULES globally for the tests.

Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/11110)

TEST: add util/wrap.pl and use it

util/wrap.pl is a script that defines the environment variables
OPENSSL_ENGINES and OPENSSL_MODULES, then calls the command line
that's given as its arguments.

On a POSIX platform, the command line call is done via
util/shlib_wrap.sh to ensure that the shared library paths are
correct.  For other platforms, util/wrap.pl currently assumes that
similar things are already in place through other means.

Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/11110)

VMS: mitigate for the C++ compiler that doesn't understand certain pragmas

This only affects __DECC_INCLUDE_EPILOGUE.H and __DECC_INCLUDE_PROLOGUE.H,
which are used automatically by HP and VSI C/C++ compilers.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/11159)

(cherry picked from commit 605a0c709f4d50497a1c49ee117a0ec4bb956d58)

doc: add a fancy CHANGES entry to celebrate the new Markdown format

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10545)

doc: revamp the SUPPORT file

Too be continued...

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10545)

doc: revamp the INSTALL file

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10545)

doc: add missing CHANGES entries for all versions >= 1.0.0

Up to now, CHANGES entries for older releases where only added to the
corresponding stable branches, so they were missing in the master
branch. This commit adds the missing entries, taking them from the
respective stable branches.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10545)

doc: add missing NEWS entries for all versions >= 1.0.0

Up to now, NEWS entries for older releases where only added to the
corresponding stable branches, so they were missing in the master
branch. This commit adds the missing entries, taking them from the
respective stable branches.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10545)

doc: revamp the README file

 * Add an OpenSSL logo and CI badges
 * Add a table of contents
 * Add a lot of links

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10545)

doc: add OpenSSL logo

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10545)

doc: introduce some minimalistic markdown without essential changes

The goal is to transform the standard documents

    README, INSTALL, SUPPORT, CONTRIBUTING, ...

from a pure text format into markdown format, but in such a way
that the documentation remains nicely formatted an easy readable
when viewed with an normal text editor.

To achieve this goal, we use a special form of 'minimalistic' markdown
which interferes as little as possible with the reading flow.

 * avoid [ATX headings][] and use [setext headings][] instead
   (works for `<h1>` and `<h2>` headings only).
 * avoid [inline links][] and use [reference links][] instead.
 * avoid [fenced code blocks][], use [indented-code-blocks][] instead.

The transformation will take place in several steps. This commit
introduces mostly changes the formatting and does not chang the
content significantly.

[ATX headings]:         https://github.github.com/gfm/#atx-headings
[setext headings]:      https://github.github.com/gfm/#setext-headings
[inline links]:         https://github.github.com/gfm/#inline-link
[reference links]:      https://github.github.com/gfm/#reference-link
[fenced code blocks]:   https://github.github.com/gfm/#fenced-code-blocks
[indented code blocks]: https://github.github.com/gfm/#indented-code-blocks

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10545)

doc: convert standard project docs to markdown

In the first step, we just add the .md extension and move some
files around, without changing any content. These changes will
occur in the following commits.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10545)

Use a wrapper for pod2html

Remove unused util/process_docs.pl

Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10856)

apps x509: restrict CAkeyform option to OPT_FMT_PDE

CAkeyform may be set to PEM, DER or ENGINE, but the current options
are not using the proper optionformat 'E' (OPT_FMT_PDE) for this.

Set the valtype for CAkeyform to 'E' and use OPT_FMT_PDE when extracting
the option value.

This amends 0ab6fc79a9a ("Fix regression on x509 keyform argument") which
did the same thing for keyform and changed the manpage synopsis entries
for both keyform and CAkeyform but did not change the option section.
Hence, change the option section for both of them.

CLA: trivial

Co-developed-by: Torben Hohn <torben.hohn@linutronix.de>
Signed-off-by: Torben Hohn <torben.hohn@linutronix.de>
Signed-off-by: Bastian Germann <bage@linutronix.de>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11085)

bugfix in cmac calculation example

The example never executes code inside of the while loop, as read()
returns bigger number than 0. Thus the end result is wrong.

CLA: trivial

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11143)

Avoid arm64 builds timeout due to silent make taking too long time

Also reuse one of the arm64 builds as a no-deprecated build
Also include a single ppc64le-build

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11181)

secmem: ignore small minsize arguments to CRYPTO_secure_malloc_init().

If the user specifies a minimum allocation size that is smaller than
the free list structure (or zero), calculate the minimum possible size rather
than failing.

Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11149)

sec_mem: add note about the minimum size parameter.

Add a note indicating that the minimum size parameter to
CRYPTO_secure_malloc_init() should be small.

Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11149)

Mem-sec small code adjustment

Conditional code readability improvement.

Remove unused macro

Commit #11042 has introduced a new, unused, CRYPTO_EX_INDEX macro.
Remove before version release.

Reviewed-by: Paul Yang <kaishen.yy@antfin.com>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11103)

x86_64: Replace .asciz "GNU" with .byte

Replace .asciz "GNU" with .byte since .asciz isn't supported on Solaris.
Fixes https://github.com/openssl/openssl/issues/11132

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11137)

Check that the DRBG's internal state has been zeroized after uninstantiation

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11111)

DRBG: delay initialization of DRBG method until instantiation

Previously, the initialization was done immediately in RAND_DRBG_set(),
which is also called in RAND_DRBG_uninstantiate().

This made it difficult for the FIPS DRBG self test to verify that the
internal state had been zeroized, because it had the side effect that
the drbg->data structure was reinitialized immediately.

To solve the problem, RAND_DRBG_set() has been split in two parts

    static int rand_drbg_set(RAND_DRBG *drbg, int type, unsigned int flags);
    static int rand_drbg_init_method(RAND_DRBG *drbg);

and only the first part is called from RAND_DRBG_uninstantiate().

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11111)

config: Drop linux-alpha-gcc+bwx

Its entry in Configuration/10-main.conf was dropped in commit
7ead0c89185c ("Configure: fold related configurations more aggressively
and clean-up.") probably because all but one of its bn_ops were removed
(RC4_CHAR remained). Benchmarks on an Alpha EV7 indicate that RC4_INT is
better than RC4_CHAR so rather than restoring the configuation, remove
it from config.

CLA: trivial
Bug: https://bugs.gentoo.org/697840

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11130)

Fix default provider merge glitch

Property "default" no longer exists, replace "default=yes" with
"provider=default"

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11150)

Refactor evp_pkey_make_provided() to do legacy to provider export

Previously, evp-keymgmt_util_export_to_provider() took care of all
kinds of exports of EVP_PKEYs to provider side keys, be it from its
legacy key or from another provider side key.  This works most of the
times, but there may be cases where the caller wants to be a bit more
in control of what sort of export happens when.

Also, when it's time to remove all legacy stuff, that job will be much
easier if we have a better separation between legacy support and
support of provided stuff, as far as we can take it.

This changes moves the support of legacy key to provider side key
export from evp-keymgmt_util_export_to_provider() to
evp_pkey_make_provided(), and makes sure the latter is called from all
EVP_PKEY functions that handle legacy stuff.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11074)

DSA: More conforming names in crypto/dsa/dsa_aid.c

Made macro names that refer to a known base OID, an commented accordingly.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10557)

doc/man3/EVP_PKEY_CTX_ctrl.pod: cleanup

Clean up a manual we've touched, according to conventions found in
Linux' man-pages(7); function arguments in descriptions should be in
italics, and types, macros and similar should be in bold, with the
exception for NULL.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10557)

PROV: Avoid MDC2 in the RSA signature implementation in the FIPS module

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10557)

test/recipes/30-test_evp_data/evppkey.txt

Tests that go through provider cannot recognise PKEY_CTRL_INVALID from
PKEY_CTRL_ERROR any more, because provided implementations' param
setting functions return 0 or 1.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10557)

test/evp_extra_test.c: adapt for RSA signature tests

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10557)

rsa_get0_all_params(): Allow zero CRT params

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10557)

PROV: add RSA signature implementation

This includes legacy PSS controls to params conversion, and an attempt
to generalise the parameter names when they are suitable for more than
one operation.

Also added crypto/rsa/rsa_aid.c, containing proper AlgorithmIdentifiers
for known RSA+hash function combinations.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10557)

Don't exclude quite so much in a no-sock build

We were excluding more code than we needed to in the OCSP/HTTP code in
the event of no-sock. We should also not assume that a BIO passed to our
API is socket based.

This fixes the no-sock build

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/11134)

Introduce the provider property

Replace the properties default, fips and legacy with a single property
called "provider". So, for example, instead of writing "default=yes" to
get algorithms from the default provider you would instead write
"provider=default". We also have a new "fips" property to indicate that
an algorithm is compatible with FIPS mode. This applies to all the
algorithms in the FIPS provider, as well as any non-cryptographic
algorithms (currently only serializers).

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11097)

Add DRBG self tests

Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/11010)

pkey: additional EC related options

Add options to change the parameter encoding and point conversions for EC
public and private keys.  These options are present in the deprecated 'ec'
utility.

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/11113)

pkey: update command line tool examples in light of deprecations.

Specifically, refer from the deprecated tools to the pkey equivalents.

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/11113)

apps: distinguish between a parameter error and an unknown parameter.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11049)

pmeth_lib: detect unsupported OSSL_PARAM.

When converting legacy controls to OSSL_PARAMs, return the unsupported -2
value correctly.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11049)

Params: add argument to the _from_text calls to indicate if the param exists.

The extra argument is a integer pointer and is optional.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11049)

Add *.d.tmp files to .gitignore

These are temporary files generated by the build process that should not
be checked in.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11122)

Deprecate the low level Diffie-Hellman functions.

Use of the low level DH functions has been informally discouraged for a
long time.  We now formally deprecate them.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11024)

DH: add CHANGES entry listing the deprecated DH functions.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11024)

DH: fix header file indentation

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11024)

Deprecate the low level RSA functions.

Use of the low level RSA functions has been informally discouraged for a
long time. We now formally deprecate them.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11063)

rsa: document deprecated RSA command line apps

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11063)

rsa: document deprecated low level functions

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11063)

NEWS: DH, DSA, ECDH, ECDSA and RSA public key function deprecation note

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11063)

rsa.h: fix preprocessor indentation

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11063)

Remove unneeded switch statement to fix warning

https://github.com/openssl/openssl/issues/10958

CLA: trivial

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11125)

Handle max_fragment_length overflow for DTLS

Allow for encryption overhead in early DTLS size check
and send overflow if validated record is too long

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11096)

[PROV][EC] Update documentation

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10631)

[BN] harden `BN_copy()` against leaks from memory accesses

`BN_copy()` (and indirectly `BN_dup()`) do not propagate the
`BN_FLG_CONSTTIME` flag: the propagation has been turned on and off a
few times in the past years, because in some conditions it has shown
unintended consequences in some code paths.

Without turning the propagation on once more, we can still improve
`BN_copy()` by avoiding to leak `src->top` in case `src` is flagged with
`BN_FLG_CONSTTIME`.
In this case we can instead use `src->dmax` as the number of words
allocated for `dst` and for the `memcpy` operation.

Barring compiler or runtime optimizations, if the caller provides `src`
flagged as const time and preallocated to a public size, no leak should
happen due to the copy operation.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10631)

[EC] harden EC_KEY against leaks from memory accesses

We should never leak the bit length of the secret scalar in the key,
so we always set the `BN_FLG_CONSTTIME` flag on the internal `BIGNUM`
holding the secret scalar.

This is important also because `BN_dup()` (and `BN_copy()`) do not
propagate the `BN_FLG_CONSTTIME` flag from the source `BIGNUM`, and
this brings an extra risk of inadvertently losing the flag, even when
the called specifically set it.

The propagation has been turned on and off a few times in the past
years because in some conditions has shown unintended consequences in
some code paths, so at the moment we can't fix this in the BN layer.

In `EC_KEY_set_private_key()` we can work around the propagation by
manually setting the flag after `BN_dup()` as we know for sure that
inside the EC module the `BN_FLG_CONSTTIME` is always treated
correctly and should not generate unintended consequences.

Setting the `BN_FLG_CONSTTIME` flag alone is never enough, we also have
to preallocate the `BIGNUM` internal buffer to a fixed public size big
enough that operations performed during the processing never trigger
a realloc which would leak the size of the scalar through memory
accesses.

Fixed Length
------------

The order of the large prime subgroup of the curve is our choice for
a fixed public size, as that is generally the upper bound for
generating a private key in EC cryptosystems and should fit all valid
secret scalars.

For preallocating the `BIGNUM` storage we look at the number of "words"
required for the internal representation of the order, and we
preallocate 2 extra "words" in case any of the subsequent processing
might temporarily overflow the order length.

Future work
-----------

A separate commit addresses further hardening of `BN_copy()` (and
indirectly `BN_dup()`).

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10631)