Add a test for a server changing the ciphersuite

Test that if a server selects a differenct ciphersuite with the same hash
in TLSv1.3 then this is accepted by the client.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3623)

Allow the server to change the ciphersuite on resume

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3623)

Add the target 'build_all_generated'

This new target is used to build all generated files and only that.
This can be used to prepare everything that requires things like perl
for a system that lacks perl and then move everything to that system
and do the rest of the build there.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3695)

Fix no-ec

Also remove nested OPENSSL_NO_EC conditional; it was properly indented,
but a no-op.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3693)

Add apps/progs.h to gitignore

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3692)

Add sha/asm/keccak1600-armv8.pl.

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

Standardize Levitte's dual-license

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3689)

Build apps/progs.h dynamically

Because apps/progs.h isn't configuration agnostic, it's not at all
suited for 'make update' or being versioned, so change it to be
dynamically generated.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3688)

Move bn and evp test programs input data to their respective data dir

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3472)

.travis.yml: Detect if 'make update' updated something

If it did, it really is something that should be checked in, and should
therefore make a CI build fail.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3686)

Correct Oracle copyrights & clarify.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3684)

Remove OLD_STR_TO_KEY compile option

This flag was added in 1992 and only documented in the CHANGES file.

Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3681)

Fix possible crash in X931 code.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3675)

Fix another possible crash in rsa_ossl_mod_exp.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3675)

Fix a possible crash in dsa_builtin_paramgen2.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3675)

Fix crash in ecdh_simple_compute_key.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3671)

Fix a possible crash in the error handling.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3672)

Update copyright on progs.h

The generating script got updated, but the generated file did not.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3668)

Fix a memleak in ec_copy_parameters.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3666)

perlasm/ppc-xlate.pl: add PowerISA 3.0B instructions.

[As well as few extra instructions from earlier spec.]

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

Make SNI behavior more clear in s_client doc & help

Update s_client -help and pod file.

Signed-off-by: Paul Yang <paulyang.inf@gmail.com>
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3654)

Rework writing crypto/err/openssl.txt

Reading the prologue of this file conserved the "# Function codes"
line, and then duplicated it when rewriting this file, adding a new
"# Function codes" line everytime there's an update.

Better then to skip over all comment lines and have the prologue
defined in mkerr.pl, just the same as we do with the other affected
files.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3664)

Add -module option to util/mkerr.pl

Sometimes, one might only want to rework a subset of all the internal
error codes.  -module allows the caller to specify exactly which
library modules to rewrite.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3662)

Clean up a bundle of codingstyle stuff in apps directory

Mostly braces and NULL pointer check and also copyright year bump

Signed-off-by: Paul Yang <paulyang.inf@gmail.com>
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3657)

Fix a memleak in tls13_generate_secret.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3660)

Put message strings in state files

Add "*" as indicator meaning the function/reason is removed, so put an
empty string in the function/reason string table; this preserves backward
compatibility by keeping the #define's.

In state files, trailing backslash means text is on the next line.

Add copyright to state files

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3640)

Fix memleak in EVP_DigestSignFinal/VerifyFinal.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3658)

Refactor functions in testdsa.h

To reduce duplicate code

Signed-off-by: Paul Yang <paulyang.inf@gmail.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3656)

Add sslapitest for SSL_early_get1_extensions_present()

Call it from the early callback used for testing these functions, and verify
the expected contents of the ClientHello

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2976)

Add SSL_early_get1_extensions_present()

It is an API to be used from the early callback that indicates what
extensions were present in the ClientHello, and in what order.
This can be used to eliminate unneeded calls to SSL_early_get0_ext()
(which itself scales linearly in the number of extensions supported
by the library).

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2976)

TLS 1.3 client sigalgs test no longer needs TLS 1.2

Per the TODO comment, we now have proper certificate selection for
TLS 1.3 client certificates, so this test can move into its own
block.  (It cannot merge with the previous block, as it requires EC.)

Verified that the test passes when configured with enable-tls1_3
no-tls1 no-tls1_1 no-tls1_2.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3131)

drop some no-longer-relevant TODO(TLS1.3) entries

We prevent compression both when the server is parsing the ClientHello
and when the client is constructing the ClientHello.  A 1.3 ServerHello
has no way to hand us back a compression method, and we already check
that the server does not try to give us back a compression method that
we did not request, so these checks seem sufficient.

Weaken the INSTALL note slightly, as we do now expect to interoperate
with other implementations.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3131)

Remove leading space-before-tab

Reviewed-by: Kurt Roceckx <kurt@openssl.org>

improve comment: use "optimization" for clarity
The previous word was a misspelling of nicety

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3464)

Introduce ASN1_TIME_set_string_X509 API

Make funcs to deal with non-null-term'd string
in both asn1_generalizedtime_to_tm() and asn1_utctime_to_tm().

Fixes issue #3444.

This one is used to enforce strict format (RFC 5280) check and to
convert GeneralizedTime to UTCTime.

apps/ca has been changed to use the new API.

Test cases and documentation are updated/added

Signed-off-by: Paul Yang <paulyang.inf@gmail.com>
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3566)

doc/man3: use the documented coding style in the example code

Adjust brace placement, whitespace after keywords, indentation and empty
lines after variable declarations according to
https://www.openssl.org/policies/codingstyle.html.

Indent literal sections by exactly one space.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3580)

Fix spelling errors in manpages

spelling: algorithm
spelling: anyway
spelling: assigned
spelling: authenticated
spelling: callback
spelling: certificate
spelling: compatibility
spelling: configuration
spelling: digest
spelling: encrypted
spelling: function
spelling: output
spelling: receive
spelling: renegotiation
spelling: signing
spelling: similar
spelling: string

(Merged from https://github.com/openssl/openssl/pull/3580)Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3580)

If-guard to avoid null ptr deref in statem_srvr.c

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
CLA: trivial
(Merged from https://github.com/openssl/openssl/pull/3419)

Remove needless type casting.

CLA: trivial

Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3627)

Fix possible usage of NULL pointers in apps/spkac.c

Check return value of NETSCAPE_SPKI_new() and
NETSCAPE_SPKI_b64_encode(), and also clean up coding style incidentally.

Signed-off-by: Paul Yang <paulyang.inf@gmail.com>
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3647)

Fix memory leaks in CTLOG_new_from_base64

Move the call to ct_base64_decode(), which allocates, until after
the check for NULL output parameter.

Also place a cap on the number of padding characters used to decrement
the output length -- any more than two '='s is not permitted in a
well-formed base64 text.  Prior to this change, ct_base64_decode() would
return a length of -1 along with allocated storage for an input of
"====".

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3379)

mark V_ASN1_PRIMATIVE_TAG as compat

This incorrectly spelled item exists for compatibility purposes

CLA: Trivial

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3460)

fix check of broken implementations of GOST ciphersuites

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3642)

Fix speed command for alternation of ciphers and digests.

CLA: trivial

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3487)

Add support for using engine-backed keys in spkac

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3599)

Fix a bundle of trailing spaces in several files

Signed-off-by: Paul Yang <paulyang.inf@gmail.com>
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3622)

List undocumented macros

The search is approximate; look only for those that look like
functions.

[skip ci]

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3641)

Remove stale note from s_server.pod

Modern browsers are now, well, pretty modern.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3644)

Ignore -named_curve auto value to improve backwards compatibility

Fixes #3490

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3518)

Fix a read off the end of the input buffer

when building with OPENSSL_SMALL_FOOTPRINT defined.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3533)

Use memset to clear SRP_CTX instead of NULL and zero assignments

This uses memset() to clear all of the SRP_CTX when free'ing or
initializing it as well as in error paths instead of having a series
of NULL and zero assignments as it is safer.

It also changes SSL_SRP_CTX_init() to reset all the SRP_CTX to zero
in case or error, previously it could retain pointers to freed
memory, potentially leading to a double free.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3467)

Make SRP_CTX.info ownership and lifetime be the same as SRP_CTX.login.

Ownership and lifetime rules of SRP_CTX.info are confusing and different
from those of SRP_CTX.login, making it difficult to use correctly.
This makes the ownership and lifetime be the same as those of SRP_CTX.login,
thet is a copy is made when setting it and is freed when SRP_CTX is freed.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3467)

Clean up s_server documentation

List the options in the same order and in the same style as the output from
"openssl s_server -help"

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3628)

Remove doc of non-existent functions

Fix test for "documenting private functions"
And add -p flag to doc-nits recipe
Mark when things were deprecated, if doc'd as such

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3624)

sha/asm/keccak1600-armv4.pl: switch to more efficient bit interleaving algorithm.

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

sha/keccak1600.c: switch to more efficient bit interleaving algorithm.

[Also bypass sizeof(void *) == 8 check on some platforms.]

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

Add back support for SHA224 based sig algs

This used to work but was inadvertently removed as part of the TLSv1.3
work. This adds it back.

Fixes #3633

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3639)

Windows: rearrange programs cleanup

The list of programs hit nmake's maximum line length, so we split up the
line in smaller chunks.

Fixes #3634

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3636)

Fix possible memory over-read in apps/s_client.c

a buffer returned from BIO_gets is not checked for it's length before
reading its contents.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3630)

Fix #340: Parse ASN1_TIME to struct tm

This works with ASN1_UTCTIME and ASN1_GENERALIZED_TIME

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3378)

doc: use /* ... */ comments in code examples

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1956)

doc/man3: fix SSL_SESSSION typos

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1956)

SSL_CTX_set_verify.pod: move a typedef in front of its first usage

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1956)

doc/man3: unindent a few unintended code blocks

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1956)

doc/man3: reformat the function prototypes in the synopses

I tried hard to keep the lines at 80 characters or less, but in a few
cases I had to punt and just indented the subsequent lines by 4 spaces.

A few well-placed typedefs for callback functions would really help, but
these would be part of the API, so that's probably for later.

I also took the liberty of inserting empty lines in overlong blocks to
provide some visual space.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1956)

ERR_put_error.pod: fix the name of function ERR_add_error_vdata()

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1956)

doc/man3: remove a duplicate BIO_do_accept() call

The SSL server example in BIO_f_ssl.pod contains two copies of the
BIO_do_accept() call. Remove the second one.

Signed-off-by: Beat Bolli <dev@drbeat.li>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1956)

doc/man3: use the documented coding style in the example code

Adjust brace placement, whitespace after keywords, indentation and empty
lines after variable declarations according to
https://www.openssl.org/policies/codingstyle.html.

Indent literal sections by exactly one space.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1956)

make error tables const and separate header file

Run perltidy on util/mkerr
Change some mkerr flags, write some doc comments
Make generated tables "const" when genearting lib-internal ones.
Add "state" file for mkerr
Renerate error tables and headers
Rationalize declaration of ERR_load_XXX_strings
Fix out-of-tree build
Add -static; sort flags/vars for options.
Also tweak code output
Moved engines/afalg to engines (from master)
Use -static flag
Standard engine #include's of errors
Don't linewrap err string tables unless necessary

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3392)

Add a lock around the OBJ_NAME table

Various initialization functions modify this table, which can cause heap
corruption in the absence of external synchronization.

Some stats are modified from OPENSSL_LH_retrieve, where callers aren't
expecting to have to take out an exclusive lock. Switch to using atomic
operations for those stats.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3525)

Fix #2400 Add NO_RENEGOTIATE option

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3432)

Always flush the BIO when we send any alert

At the moment we flush the write BIO if we send a fatal alert, but not a
warning one. This can mean the warning is never sent if we never do another
write and subsequently flush the BIO. Instead we should just always flush
after writing an alert.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3432)

Handle the server refusing to reneg in a reneg_setup

During setup of a reneg test the server can refuse to start reneg.
If that happens we should let the client continue and then fail.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3432)

sha/asm/keccak1600-armv4.pl: add NEON code path.

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

sha/asm/keccak1600-armv4.pl: add SHA3_absorb and SHA3_squeeze.

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

sha/asm/keccak1600-armv4.pl: optimization based on profiler feedback.

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

Add sha/asm/keccak1600-armv4.pl.

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

Fix coding style

Signed-off-by: Paul Yang <paulyang.inf@gmail.com>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3614)

Fix doc nits in X509_check_private_key.pod

remove the tailing dot

Signed-off-by: Paul Yang <paulyang.inf@gmail.com>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3614)

Set local variable to 0 to avoid build error

Signed-off-by: Paul Yang <paulyang.inf@gmail.com>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3614)

Add test cases for X509_check_private_key

To test X509_check_private_key and relatives.

Add a CSR and corresponding RSA private key to test
X509_REQ_check_private_key function.

Signed-off-by: Paul Yang <paulyang.inf@gmail.com>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3614)

Document X509_check_private_key and relative

Document two private key check functions:

X509_check_private_key
X509_REQ_check_private_key

Signed-off-by: Paul Yang <paulyang.inf@gmail.com>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3614)

Move and update RSA-PSS documentation.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3621)

Fix define to match documentation

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3621)

Document default client -psk_identity

Document that -psk is required to use PSK cipher
[skip ci]

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3607)

Avoid failing s_server when client's psk_identity is unexpected

s_server has traditionally been very brittle in PSK mode.  If the
client offered any PSK identity other than "Client_identity" s_server
would simply abort.

This is breakage for breakage's sake, and unlike most other parts of
s_server, which tend to allow more flexible connections.

This change accomplishes two things:

 * when the client's psk_identity does *not* match the identity
   expected by the server, just warn, don't fail.

 * allow the server to expect instead a different psk_identity from
   the client besides "Client_identity"

Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3605)

Use common STANZA for bn and exp tests

Standardize file:line messages
Reduce buff size; move to end of STANZA
Add some Title entries (with blank line after)
Add Title to some BN test files.

Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3608)

sha/keccak1600.c: add #ifdef KECCAK1600_ASM.

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

sha/keccak1600.c: reduce temporary storage utilization even futher.

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

sha/keccak1600.c: add another 1x variant.

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

sha/keccak1600.c: add ARM-specific "reference" tweaks.

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

ec/asm/ecp_nistz256-x86_64.pl: minor sqr_montx cleanup.

Drop some redundant instructions in reduction in ecp_nistz256_sqr_montx.

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

Fix ex_data and session_dup issues

Code was added in commit b3c31a65 that overwrote the last ex_data value
using CRYPTO_dup_ex_data() causing a memory leak, and potentially
confusing the ex_data dup() callback.

In ssl_session_dup(), fix error handling (properly reference and up-ref
shared data) and new-up the ex_data before calling CRYPTO_dup_ex_data();
all other structures that dup ex_data have the destination ex_data new'd
before the dup.

Fix up some of the ex_data documentation.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3323)

Add text pointing to full change list.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3606)

Work around Travis "virtual memory exhausted" error

One particular build was running out of memory. By swapping to debug mode
we reduce the optimisation level which should reduce the amount of memory
required.

[extended tests]

Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3601)

Left shift of a negative number is undefined behaviour

Fix undefined behaviour in curve25519.c. Prior to this running with
ubsan produces errors like this:

crypto/ec/curve25519.c:3871:18: runtime error: left shift of negative
value -22867

[extended tests]

Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3600)

Added mysql as starttls protocol.

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3456)

Remove unnecessary if condition from apps.c

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3596)

Add -p (public only) flag to find-doc-nits

Report if any non-public items are documented.
Add util/private.num that lists items that aren't in the public
(lib*.num) files that we do want to document.

Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3603)

Only release thread-local key if we created it.

Thanks to Jan Alexander Steffens for finding the bug and confirming the
fix.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3592)