Make manuals with TYPE conform with man-pages(7)

Details from man-pages(7) that are used:

    Formatting conventions for manual pages describing functions

        ...
        Variable names should, like argument names, be specified in italics.
        ...

    Formatting conventions (general)

        ...
        Special macros, which are usually in uppercase, are in bold.
        Exception: don't boldface NULL.
        ...

Furthermore, for TYPE used as a placeholder for types and correponding
part of function names, we extrapolate that it's both a type and a
variable, and should therefore be bold (typical for types and function
names) and italic (typical for variables).  POD processors don'e know
this, so we have to help them along.  Therefore:

   SPARSE_ARRAY_OF(TYPE)        => B<SPARSE_ARRAY_OF>(B<I<TYPE>>)
   ossl_sa_TYPE_num()           => B<ossl_sa_I<TYPE>_num>()
   TYPE                         => B<I<TYPE>>

There are some other less typical uses where one simply has to give
formatting some extra though.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10041)

Make ASN1 manuals conform with man-pages(7)

Details from man-pages(7) that are used:

    Formatting conventions for manual pages describing functions

        ...
        Variable names should, like argument names, be specified in italics.
        ...

    Formatting conventions (general)

        ...
        Special macros, which are usually in uppercase, are in bold.
        Exception: don't boldface NULL.
        ...

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10042)

Consistent formatting of flags with args

For documentation of all commands with "-flag arg" format them
consistently: "B<-flag> I<arg>", except when arg is literal
(for example "B<-inform> B<PEM>|B<DER>")
Update find-doc-nits to complain if badly formatted strings are found.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10022)

Add defines for __NR_getrandom for all Linux architectures

Fixes: #10015

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
GH: #10044

 Make ASYNC manuals conform with man-pages(7)

 Details from man-pages(7) that are used:

     Formatting conventions for manual pages describing functions

         ...
         Variable names should, like argument names, be specified in italics.
         ...

     Formatting conventions (general)

         ...
         Special macros, which are usually in uppercase, are in bold.
         Exception: don't boldface NULL.
         ...

Additionally, expanded some lists to make better use of POD formatting.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10043)

Correct the function names in SSL_CTX_set_stateless_cookie_generate_cb.pod

Although the synopsis used the correct function names, the description did
not. Also the description of the equivalent DTLSv1_listen() callbacks was
missing, so these have been added.

Fixes #10030

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10033)

Fix a bundle of mischecks of return values

Several EVP_PKEY_xxxx functions return 0 and a negative value for
indicating errors. Some places call these functions with a zero return
value check only, which misses the check for the negative scenarios.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10055)

Make default values by ERR_get_error_all() and friends more consistent

Unset data defaults to the empty string ("") or 0.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9948)

Fix a return value bug in apps/speed.c

Those functions returns less than and equal to 0 to indicate an error
occured.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10054)

Consistent naming for context gettable param queries .

All instances of EVP_*_CTX_gettable_params functions have been renamed
to EVP_*_gettable_ctx_params.  Except for the EVP_MD ones which were changed
already.

These functions do not take EVP_*_CTX arguments so their prior naming was
misleading.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10052)

Make relevant tests more sensitive to 'no-fips'

This applies to test/recipes/30-test_evp.t and
test/recipes/30-test_evp_fetch_prov.t.

Additionally, we make test/recipes/30-test_evp_fetch_prov.t data
driven, to make test number planning more automated, and to separate
what is unique from what is common to all the test cases.

[extended tests]

Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/10047)

Fix a document description in apps/req

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9958)

Fix a double free issue when signing SM2 cert

If the SM2 ID value has not been passed correctly when signing an SM2
certificate/certificate request, a double free occurs. For instance:

  openssl req -x509 ... -sm2-id 1234567812345678

The '-sm2-id' should not be used in this scenario, while the '-sigopt' is
the correct one to use. Documentation has also been updated to make the
options more clear.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9958)

Add util/fix-includes script

This script contains all adjustments to header files which were made
during the reorganization of the header files. It is meant as an aid
for other contributors which encounter preprocessor #include errors
after rebasing over this pull request. Simply running

  util/fix-includes

from the root of the source directory should hopefully fix the problem.

Note: such #include errors are expected only for pull requests which
add a lot of new code, in particular new compilation modules.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9333)

Add legacy include guards to public header files

For every public header file, the old include guard definition needs
to be added in addition to the new one

include/openssl/file.h:

    #ifndef OPENSSL_FILE_H
    # define OPENSSL_FILE_H
    # pragma once

    # include <openssl/macros.h>
    # if !OPENSSL_API_3
    #  define HEADER_FILE_H
    # endif
    ...

This is going to ensure that applications which use the old include guards
externally, for example like this

    #ifndef HEADER_FILE_H
    # include <openssl/file.h>
    #endif

will not fail to compile.

In addition to the legacy guard, the public header files also receive a
'# pragma once' directive.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9333)

Fix header file include guard names

Make the include guards consistent by renaming them systematically according
to the naming conventions below

For the public header files (in the 'include/openssl' directory), the guard
names try to match the path specified in the include directives, with
all letters converted to upper case and '/' and '.' replaced by '_'. For the
private header files files, an extra 'OSSL_' is added as prefix.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9333)

Reorganize public header files (part 2)

Add an <openssl/ossl_typ.h> compatibility header.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9333)

Reorganize public header files (part 1)

Rename <openssl/ossl_typ.h> to <openssl/types.h>.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9333)

Reorganize local header files

Apart from public and internal header files, there is a third type called
local header files, which are located next to source files in the source
directory. Currently, they have different suffixes like

  '*_lcl.h', '*_local.h', or '*_int.h'

This commit changes the different suffixes to '*_local.h' uniformly.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9333)

Reorganize private crypto header files

Currently, there are two different directories which contain internal
header files of libcrypto which are meant to be shared internally:

While header files in 'include/internal' are intended to be shared
between libcrypto and libssl, the files in 'crypto/include/internal'
are intended to be shared inside libcrypto only.

To make things complicated, the include search path is set up in such
a way that the directive #include "internal/file.h" could refer to
a file in either of these two directoroes. This makes it necessary
in some cases to add a '_int.h' suffix to some files to resolve this
ambiguity:

  #include "internal/file.h"      # located in 'include/internal'
  #include "internal/file_int.h"  # located in 'crypto/include/internal'

This commit moves the private crypto headers from

  'crypto/include/internal'  to  'include/crypto'

As a result, the include directives become unambiguous

  #include "internal/file.h"       # located in 'include/internal'
  #include "crypto/file.h"         # located in 'include/crypto'

hence the superfluous '_int.h' suffixes can be stripped.

The files 'store_int.h' and 'store.h' need to be treated specially;
they are joined into a single file.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9333)

Add legacy include guard manually to opensslconf.h.in

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9333)

Funtion name with variable part in doc/man7/ and doc/internal/man3/

We have a few pages where part of function names can be considered
variable.  There are no normative guidelines for such a case, but if
we draw from the formatting convention of variable and argument names,
we can draw the conclusion that this variable part should be italized,
within already given conventions.  In other words, we need to help the
POD processor along in cases like these:

    SPARSE_ARRAY_OF(TYPE)
    ossl_sa_TYPE_num()

These need explicit formatting:

    B<SPARSE_ARRAY_OF>(I<TYPE>)
    B<ossl_sa_I<TYPE>_num>()

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10034)

Make doc/man7/ and doc/internal/man3/ conform with man-pages(7)

It's all in the details, from man-pages(7):

    Formatting conventions for manual pages describing functions

        ...
        Variable names should, like argument names, be specified in italics.
        ...

    Formatting conventions (general)

        ...
        Special macros, which are usually in uppercase, are in bold.
        Exception: don't boldface NULL.
        ...

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10034)

[KDF] Add KBKDF implementation for counter-mode HMAC

Implement SP800-108 section 5.1 with HMAC intended for use in Kerberos.
Add test vectors from RFC 8009.

Adds error codes PROV_R_INVALID_MAC and PROV_R_MISSING_MAC.

Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9924)

OSSL_PARAM functions: change to allow the data field to be NULL

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10025)

OSSL_PARAM.pod: document the mechanism to figure out buffer sizes

When requesting parameters, it's acceptable to make a first pass with
the |data| field of some parameters being NULL.  That can be used to
help the requestor to figure out dynamically what buffer size is
needed.  For variable size parameters, there's no other way to find
out.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10025)

Certificate Management Protocol (CMP, RFC 4210) extension to OpenSSL
    Also includes CRMF (RFC 4211) and HTTP transfer (RFC 6712)

    CMP and CRMF API is added to libcrypto, and the "cmp" app to the openssl CLI.
        Adds extensive man pages and tests.  Integration into build scripts.

    Incremental pull request based on OpenSSL commit 8869ad4a39f of 2019-04-02

    4th chunk: CMP context/parameters and utilities
    in crypto/cmp/cmp_ctx.c, crypto/cmp/cmp_util.c, and related files

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9107)

Reformat param description lines

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10024)

Make OSSL_PARAM descriptions uniformly formatted.

  =item B<MACRO> ("name") <type>

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10024)

When building of modules is disabled, build the legacy provider into libcrypto

This makes the legacy provider available regardless of building conditions.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9637)

Configure, build.info: make it possible to use variables in indexes

That will make it possible to assign different goals for translation
units depending on need.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9637)

Ensure man1 POD files start with openssl-

Commit b6b66573 (PR #9679) renamed most POD files. This change causes
find-doc-nits to flag misnamed files.
Also fix the two misnamed files that it found.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10000)

Missing else in cipher_hw_aes_ocb_generic_initkey

This came from commit 3837c202 "Add aes_ocb cipher to providers". It
causes the default non-hardware accelerated AES implementation to be
used even if HWAES_CAPABLE is set. Affects all platforms except X86 and
SPARC.

Patch by: Nick Gasson <Nick.Gasson@arm.com>
Change-Id: I26001a3a922ff23f6090fdcefefaecf68e92e2a6

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10012)

Use OSSL_PARAM types.  Limits are explained in the description where appropriate.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10011)

Use OSSL_PARAM types.  Limits are explained in the description where appropriate.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10011)

KDF section 3 clean up

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10011)

Use OSSL_PARAM types for MAC documentation

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10011)

Use OSSL_PARAM types for KDF documentation

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10011)

s390x assembly pack: accelerate X25519, X448, Ed25519 and Ed448

using PCC and KDSA instructions.

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/10004)

s390x assembly pack: cleanse only sensitive fields

of instruction parameter blocks.

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/10004)

s390x assembly pack: fix OPENSSL_s390xcap z15 cpu mask

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/10004)

s390x assembly pack: fix msa3 stfle bit detection

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/10004)

Make EVP_MD_CTX_[gettable|settable]_params() take an EVP_MD_CTX

EVP_MD_CTX_gettable_params() and EVP_MD_CTX_settable_params() were
confusingly named because they did not take an EVP_MD_CTX parameter.

In addition we add the functions EVP_MD_gettable_ctx_params() and
EVP_MD_settable_ctx_params() which do the same thing but are passed
an EVP_MD object instead.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9998)

Remove the engine parameter from the provider MAC documentations

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9971)

Remove the engine parameter from the individual MAC documentation

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9971)

Remove engine param from EVP_KDF and EVP_MAC documentation

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9971)

Remove engine param from the settable list

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9971)

Remove engine param macros from wrapper APIs

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9971)

Make the "engine" parameter to some provider algorithms (KDF/PRF) hidden.

This parameter will disappear once engines are wrapped by a provider so
it shouldn't ever be visible to the public.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9971)

Add rc4 cipher to default provider

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9992)

DOC: fix documentation of som EVP_MD_CTX functions

They were documented to take an EVP_MD pointer, when they really take
an EVP_MD_CTX pointer.

Fixes #9993

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9996)

Fix bugs in "info" commands flags

Remove -c/-e/-m aliases, OpenSSL commands don't do that.
Fix typo's in the documentation.
Fix -module flag to print the right thing.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9986)

fix for dsa key size feature request issue: pkey -text or -text_pub should show dsa key size

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/9983)

Support printing out some otherName variants

The supported variants are
- SmtpUTF8Name
- xmppAddr
- MS UPN
- SRVName

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9965)

Add SSKDF test vectors from RFC 8636

RFC 8636 defines the Pkinit Agility KDF, which turns out to be just a
standard SSKDF with the Info built out of the ASN.1 option of SP 800 56A
(See 5.8.2.1.2 of NIST SP 800-56A Rev. 3)

RFC 8636 Also defines test vectors, so let's add them in addition to the
tests from "non-official" test vectors.

Signed-off-by: Simo Sorce <simo@redhat.com>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9957)

Update new TLS version options to s_time man page

Reviewed-by: Paul Yang <kaishen.yy@antfin.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9947)

Add TLS version options to s_time

Reviewed-by: Paul Yang <kaishen.yy@antfin.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9947)

Rework cipher / digest fetching for legacy nids with multiple name support

With multiple names, it's no longer viable to just grab the "first" in
the set and use that to find the legacy NID.  Instead, all names for
an algorithm must be checked, and if we encounter more than one NID
asssociated with those names, we consider it an error and make that
method unloadable.

This ensures that all methods that do have an internal NID associated
will get that NID in their structure, thereby ensuring that other
parts of libcrypto that haven't gone away from using NIDs for
comparison will continue to work as expected.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9969)

include/openssl/macros.h: better OPENSSL_FUNC fallback

Make sure OPENSSL_FUNC gets defined to something, no matter what.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9976)

Add des ciphers to default provider

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9954)

Note that the mac command is preferrable to the MAC command line options.

The dgst command allows MACs to be calculated, the mac command is the more
recent interface for doing the same and provides better access to a wider
range of MACs.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/9962)

Configure: add missing dependency to fix parallel builds on Windows

The issue was encountered when testing parallel builds of OpenSSL on
Windows using `jom` instead of `nmake`. The builds persistently failed
with the following error message because the generated file "buildinf.h"
did not exist yet.

crypto\info.c(15): fatal error C1083:
    cannot open include file: "buildinf.h": No such file or directory

Apparently this error does not occur on Linux because `make` parallelizes
the builds differently such that `crypto\cversion.c`, which has an
explicit dependency on `buildinf.h`, gets compiled first. Also, the
include dependency was added only recently in commit 096978f0990.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9960)

Use the correct maximum indent

Found by OSS-Fuzz

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
GH: #9959

include/openssl/macros.h: Rework OPENSSL_FUNC for div C standards

OPENSSL_FUNC was defined as an alias for __FUNCTION__ with new enough
GNU C, regardless of the language standard used.  We change this
slightly, so this won't happen unless __STDC_VERSION is defined.

Fixes #9911

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9913)

Remove name string from PROV_CIPHER and PROV_DIGEST

It was short lived, as it's not necessary any more.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9946)

Refactor SSKDF to create the MAC contexts early

The SSKDF implementation fetched the digest(s) for the underlying MAC,
just to get their names and pass those down to the MAC, which in turn
would fetch those same digests again.

This change circumvents this by fetching the MAC and create the MAC
contexts for them directly when this PRF receives the relevant
parameters, thus only having to pass EVP_MAC_CTX pointers around.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9946)

Refactor TLS-PRF's kdf_tls1_prf_mkmacctx() to a provider utility

ossl_prov_macctx_load_from_params() creates a EVP_MAC_CTX *, or sets
new common parameters for an existing one.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9946)

Add aes_wrap cipher to providers

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9406)

Fix missing bn_mul_mont symbol in solaris fips module

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9943)

Move gcm decryption tag check higher up in the callstack

Code was updated for s390 that accidently removed the check inside the final method.
Moving the check up before the final method is called is a better way of handling this.
The oneshot method also calls the final method but doesnt need to do this check.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9945)

Exit non-zero if find-doc-nits finds nits

Filter all output to a new &err() routine, which sets the global
exit status, $status.
Also, fix all subroutine definitions and references to be consistent:
no prototypes, no & before function calls.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Yang <kaishen.yy@antfin.com>
(Merged from https://github.com/openssl/openssl/pull/9733)

Avoid ?: construct in XXXerr calls

It either makes the flow of control simpler and more obvious, or it is
just a "cleanup" so that the editing scripts will find and fixup things.

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9441)

Add merge-err-lines script

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9441)

Deprecate XXXerr() macros

Actually, for transition, they're not really deprecated.  Remove the
"1 ||" from the ifdef line (in include/openssl/err.h) when ready to
do this in production/"for real"

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9441)

Add script convert XXerr to ERR_raise

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9441)

Modify providers that keep track of underlying algorithms

With some provider implementations, there are underlying ciphers,
digests and macs.  For some of them, the name was retrieved from the
method, but since the methods do not store those any more, we add
different mechanics.

For code that needs to pass on the name of a cipher or diges via
parameters, we simply locally store the name that was used when
fetching said cipher or digest.  This will ensure that any underlying
code that needs to fetch that same cipher or digest does so with the
exact same name instead of any random name from the set of names
associated with the algorithm.

For code that needs to check what kind of algorithm was passed, we
provide EVP_{type}_is_a(), that returns true if the given method has
the given name as one of its names.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9897)

In provider implemented methods, save the name number, not the name string

Multiple names per implementation is already supported in the namemap,
but hasn't been used yet.  However, as soon as we have multiple names,
we will get an issue with what name should be saved in the method.

The solution is to not save the name itself, but rather the number
it's associated with.  This number is supposed to be unique for each
set of names, and we assume that algorithm names are globally unique,
i.e. there can be no name overlap between different algorithm types.

Incidently, it was also found that the 'get' function used by
ossl_construct_method() doesn't need all the parameters it was given;
most of what it needs, it can now get through the data structure given
by the caller of ossl_construct_method().  As a consequence,
ossl_construct_method() itself doesn't need all the parameters it was
given either.

There are some added internal functions that are expected to disappear
as soon as legacy code is removed, such as evp_first_name() and
ossl_namemap_num2name().

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9897)

Fix Solaris compile errors in provider ciphers

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9941)

Add option grouping capability to apps

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9920)

Add aes_ocb cipher to providers

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9320)

Add sm4 ciphers to default provider

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9935)

Update fuzzing README for recent clang versions

Recent clang versions ship with libfuzzer, so there's no need to build
libfuzzer yourself. They also have a dedicated -fsanitize=fuzzer-no-link
flag and no longer support the sanitize flags described in the fuzzing
README. Update it to reflect all this.

Fixes #8768.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
GH: #8891

Fix strict-warnings build

..which was broken for s390 due to 1c3ace68.

Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/9937)

Fix aes ofb, cfb and cfb8 for s390x

..which was broken since e1178600.

Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/9939)

Add SEED ciphers to default provider

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9932)

Refactor TLS1-PRF to create the MAC contexts early

The TLS1-PRF implementation fetched the digest(s) for the underlying
MAC, just to get their names and pass those down to the MAC, which in
turn would fetch those same digests again.

This change circumvents this by fetching the MAC (or MACs in the
MD5-SHA1 special case) and create the MAC contexts for them directly
when this PRF receives the relevant parameters, thus only having to
pass EVP_MAC_CTX pointers around.

Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9930)

Add cast5 ciphers to default provider

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9929)

Change PARAMETER NAMES links to PARAMETERS

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9905)

Use PARAMETERS instead of PARAMETER NAMES for the heading for consistency.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9905)

Fix links to functions.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9905)

Cleanup KDF section 1 documentation.

Remove reference to EVP_KDF_ctrl_str and replace it with EVP_KDF_CTX_set_params.

Add missing links, and specify two extra KDFs.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9905)

Clean up KDF documentation in section 7.

The EVP_KDF_ctrl function doesn't exist anymore and have been replaced by
EVP_KDF_CTX_set_params.

The EVP_KDF_new_id function doesn't exist either and EVP_KDF_new should be
used instead.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9905)

Add support for io_pgetevents_time64 syscall

32-bit architectures that are y2038 safe don't include syscalls that use
32-bit time_t. Instead these architectures have suffixed syscalls that
always use a 64-bit time_t. In the case of the io_getevents syscall the
syscall has been replaced with the io_pgetevents_time64 syscall instead.

This patch changes the io_getevents() function to use the correct
syscall based on the avaliable syscalls and the time_t size. We will
only use the new 64-bit time_t syscall if the architecture is using a
64-bit time_t. This is to avoid having to deal with 32/64-bit
conversions and relying on a 64-bit timespec struct on 32-bit time_t
platforms. As of Linux 5.3 there are no 32-bit time_t architectures
without __NR_io_getevents. In the future if a 32-bit time_t architecture
wants to use the 64-bit syscalls we can handle the conversion.

This fixes build failures on 32-bit RISC-V.

Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9819)

Add IDEA ciphers to default provider

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9917)

Fix Compiler error/warning for windows icl build

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9907)

Fix Coverity CID:1453685 'unreachable code' in aes_xts code.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/9902)

crypto/bn/build.info: Correct use of SSE2 definition

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9879)

apps/pkcs12: print multiple PKCS#12 safeBag attribute values if present

Currently the pkcs12 app will only ever print the first value of a multi-value
attribute. This is OK for some attributes (e.g. friendlyName, localKeyId) but
may miss values for other attributes.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/9751)

Fix Solaris aes_hw_t4 compile issue

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9910)

Fix building statically without any dso support

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9895)