Only allow ephemeral RSA keys in export ciphersuites.

OpenSSL clients would tolerate temporary RSA keys in non-export
ciphersuites. It also had an option SSL_OP_EPHEMERAL_RSA which
enabled this server side. Remove both options as they are a
protocol violation.

Thanks to Karthikeyan Bhargavan for reporting this issue.
(CVE-2015-0204)
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>

CHANGES: mention "universal" ARM support.

This is re-commit without unrelated modification.

Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 0548505f4cbd49b9724fab28881e096f9d951e6f)

Revert "CHANGES: mention "universal" ARM support."

This reverts commit caeed719fe3fd619415755f245ab8a904978d99d.

Reviewed-by: Matt Caswell <matt@openssl.org>

CHANGES: mention "universal" ARM support.

Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 4fec91506975f62a2f93be71a46acc7fae7eef45)

Remove inconsistency in ARM support.
This facilitates "universal" builds, ones that target multiple
architectures, e.g. ARMv5 through ARMv7. See commentary in
Configure for details.

Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit c1669e1c205dc8e695fb0c10a655f434e758b9f7)

ECDH downgrade bug fix.

Fix bug where an OpenSSL client would accept a handshake using an
ephemeral ECDH ciphersuites with the server key exchange message omitted.

Thanks to Karthikeyan Bhargavan for reporting this issue.

CVE-2014-3572
Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit b15f8769644b00ef7283521593360b7b2135cb63)

update ordinals

Reviewed-by: Emilia Käsper <emilia@openssl.org>
(cherry picked from commit 31c65a7bc0de7ff1446645d41af388893362f579)

Ensure that the session ID context of an SSL* is updated
when its SSL_CTX is updated.

From BoringSSL commit
https://boringssl.googlesource.com/boringssl/+/a5dc545bbcffd9c24cebe65e9ab5ce72d4535e3a

Reviewed-by: Rich Salz <rsalz@openssl.org>
(cherry picked from commit 61aa44ca99473f9cabdfb2d3b35abd0b473437d1)

Constify ASN1_TYPE_cmp add X509_ALGOR_cmp.

Reviewed-by: Emilia Käsper <emilia@openssl.org>
(cherry picked from commit 4c52816d35681c0533c25fdd3abb4b7c6962302d)

Fix various certificate fingerprint issues.

By using non-DER or invalid encodings outside the signed portion of a
certificate the fingerprint can be changed without breaking the signature.
Although no details of the signed portion of the certificate can be changed
this can cause problems with some applications: e.g. those using the
certificate fingerprint for blacklists.

1. Reject signatures with non zero unused bits.

If the BIT STRING containing the signature has non zero unused bits reject
the signature. All current signature algorithms require zero unused bits.

2. Check certificate algorithm consistency.

Check the AlgorithmIdentifier inside TBS matches the one in the
certificate signature. NB: this will result in signature failure
errors for some broken certificates.

3. Check DSA/ECDSA signatures use DER.

Reencode DSA/ECDSA signatures and compare with the original received
signature. Return an error if there is a mismatch.

This will reject various cases including garbage after signature
(thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
(negative or with leading zeroes).

CVE-2014-8275
Reviewed-by: Emilia Käsper <emilia@openssl.org>
(cherry picked from commit 684400ce192dac51df3d3e92b61830a6ef90be3e)

Additional fix required for no-srtp to work

RT3638

Reviewed-by: Emilia Käsper <emilia@openssl.org>

Fix building with no-srtp

RT3638

Reviewed-by: Emilia Käsper <emilia@openssl.org>
Conflicts:
ssl/t1_lib.c

Add a clang build target for linux-x86_64

This change documents the world as-is, by turning all warnings on,
and then turning warnings that trigger off again.

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

ecp_nistz256-x86_64.pl: fix occasional failures.

RT: 3607
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Emilia Kasper <emilia@openssl.org>
(cherry picked from commit 9e557ab2624d5c5e8d799c123f5e8211664d8845)

RT2914: NULL check missing in X509_name_canon

Check for NULL return from X509_NAME_ENTRY_new()

Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
(cherry picked from commit 2c60925d1ccc0b96287bdc9acb90198e7180d642)

Update SGC flag comment.

Since SGC has been removed from OpenSSL 1.0.2 the
SSL3_FLAGS_SGC_RESTART_DONE is no longer used. However the #define is
retained for compatibility.
Reviewed-by: Matt Caswell <matt@openssl.org>

Remove MS SGC

MS Server gated cryptography is obsolete and dates from the time of export
restrictions on strong encryption and is only used by ancient versions of
MSIE.
Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 63eab8a620944a990ab3985620966ccd9f48d681)

Clear existing extension state.

When parsing ClientHello clear any existing extension state from
SRP login and SRTP profile.

Thanks to Karthikeyan Bhargavan for reporting this issue.
Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 4f605ccb779e32a770093d687e0554e0bbb137d3)

Conflicts:
ssl/t1_lib.c

typo in s_client

Signed-off-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Geoff Thorpe <geoff@openssl.org>

Make "run" volatile

RT#3629

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

Document openssl dgst -hmac option

Signed-off-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Richard Levitte <levitte@openssl.org>

dlfcn: always define _GNU_SOURCE

We need this for the freebsd kernel with glibc as used in the Debian kfreebsd
ports.  There shouldn't be a problem defining this on systems not using glibc.

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

Fix memory leak in the apps

The BIO_free() allocated ex_data again that we already freed.

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

Improves certificates HOWTO

* adds links to various related documents.
* fixes a few typos.
* rewords a few sentences.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(cherry picked from commit 67472bd82bed9d5e481b0d75926aab93618902be)

Small typo

Reviewed-by: Stephen Henson <steve@openssl.org>
(cherry picked from commit 7cfab40f43afac2f46652886e260d1c4de058806)

Fix incorrect OPENSSL_assert() usage.

Return an error code for I/O errors instead of an assertion failure.

PR#3470
Reviewed-by: Stephen Henson <steve@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
(cherry picked from commit 2521fcd8527008ceb3e4748f95b0ed4e2d70cfef)

Fix a problem if CFLAGS is too long cversion.c fails to compile when config
is run with --strict-warnings.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(cherry picked from commit 488f16e31b8f5ec2513410929325d0830d76762d)

Return error when a bit string indicates an invalid amount of bits left

Reviewed-by: Matt Caswell <matt@openssl.org>

Reject invalid constructed encodings.

According to X6.90 null, object identifier, boolean, integer and enumerated
types can only have primitive encodings: return an error if any of
these are received with a constructed encoding.
Reviewed-by: Emilia Käsper <emilia@openssl.org>

Add a comment noting the padding oracle.

Reviewed-by: Andy Polyakov <appro@openssl.org>
(cherry picked from commit 03af843039af758fc9bbb4ae6c09ec2bc715f2c5)

Revert "RT3425: constant-time evp_enc"

Causes more problems than it fixes: even though error codes
are not part of the stable API, several users rely on the
specific error code, and the change breaks them. Conversely,
we don't have any concrete use-cases for constant-time behaviour here.

This reverts commit 738911cde68b2b3706e502cf8daf5b14738f2f42.

Reviewed-by: Andy Polyakov <appro@openssl.org>

Build fixes

Various build fixes, mostly uncovered by clang's unused-const-variable
and unused-function errors.

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

Clear warnings/errors within RL_DEBUG code sections (RL_DEBUG should be renamed)

Reviewed-by: Andy Polyakov <appro@openssl.org>

Clear warnings/errors within TLS_DEBUG code sections

Reviewed-by: Andy Polyakov <appro@openssl.org>

Clear warnings/errors within KSSL_DEBUG code sections

Reviewed-by: Andy Polyakov <appro@openssl.org>

Clear warnings/errors within CIPHER_DEBUG code sections

Reviewed-by: Andy Polyakov <appro@openssl.org>

Clear warnings/errors within CIPHER_DEBUG code sections

Reviewed-by: Andy Polyakov <appro@openssl.org>

Clear warnings/errors within BN_CTX_DEBUG code sections

Reviewed-by: Andy Polyakov <appro@openssl.org>

Premaster secret handling fixes

From BoringSSL
- Send an alert when the client key exchange isn't correctly formatted.
- Reject overly short RSA ciphertexts to avoid a (benign) out-of-bounds memory access.

Reviewed-by: Kurt Roeckx <kurt@openssl.org>
(cherry picked from commit 4aecfd4d9f366c849c9627ab666d1b1addc024e6)

Backport regression test

master branch has a specific regression test for a bug in x86_64-mont5 code,
see commit cdfe0fdde6a966bdb0447de66aa04a85d99a0551.

This code is now in 1.0.2/1.0.1, so also backport the test.

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

Check for invalid divisors in BN_div.

Invalid zero-padding in the divisor could cause a division by 0.

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

Check return value of ssl3_output_cert_chain

(cherry picked from commit 66f96fe2d519147097c118d4bf60704c69ed0635)

Reviewed-by: Tim Hudson <tjh@openssl.org>

The dtls1_output_cert_chain function no longer exists so remove it from
ssl_locl.h

Reviewed-by: Tim Hudson <tjh@openssl.org>

Don't set client_version to the ServerHello version.

The client_version needs to be preserved for the RSA key exchange.

This change also means that renegotiation will, like TLS, repeat the old
client_version rather than advertise only the final version. (Either way,
version change on renego is not allowed.) This is necessary in TLS to work
around an SChannel bug, but it's not strictly necessary in DTLS.

(From BoringSSL)

Reviewed-by: Emilia Käsper <emilia@openssl.org>
(cherry picked from commit ec1af3c4195c1dfecdd9dc7458850ab1b8b951e0)

Add more meaningful OPENSSL_NO_ECDH error message for suite b mode

Reviewed-by: Emilia Käsper <emilia@openssl.org>
(cherry picked from commit db812f2d70f0695fd53b386fe5e870bef8ca3c22)

Add OPENSSL_NO_ECDH guards

Reviewed-by: Emilia Käsper <emilia@openssl.org>
(cherry picked from commit af6e2d51bfeabbae827030d4c9d58a8f7477c4a0)

Remove extraneous white space, and add some braces

Reviewed-by: Emilia Käsper <emilia@openssl.org>
(cherry picked from commit 55e530265a7ea8f264717a4e37338cc04eca2007)

DTLS fixes for signed/unsigned issues

Reviewed-by: Emilia Käsper <emilia@openssl.org>
(cherry picked from commit 1904d21123849a65dafde1705e6dd5b7c2f420eb)

Fix unused variable warning

The temporary variable causes unused variable warnings in opt mode with clang,
because the subsequent assert is compiled out.

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

Fix memory leak in s2_srvr.c if BUF_MEM_grow fails

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Kurt Roeckx <kurt@openssl.org>

Fixed memory leak if BUF_MEM_grow fails

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Kurt Roeckx <kurt@openssl.org>

Fixed memory leak in the event of a failure of BUF_MEM_grow

Reviewed-by: Richard Levitte <levitte@openssl.org>
(cherry picked from commit 41bf25013032da0eeb111ce3c8fc0946c0e61c41)

Fix memory leak in SSL_new if errors occur.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(cherry picked from commit 76e6509085ea96df0ca542568ee2596343711307)

Clarify the return values for SSL_get_shared_curve.

Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 376e2ca3e3525290619602dc6013c97c9653c037)

Add extra checks for odd-length EC curve lists.

Odd-length lists should be rejected everywhere upon parsing. Nevertheless,
be extra careful and add guards against off-by-one reads.

Also, drive-by replace inexplicable double-negation with an explicit comparison.

Reviewed-by: Matt Caswell <matt@openssl.org>

Make 'make update' succeed and run it

Reviewed-by: Dr Stephen Henson <steve@openssl.org>

Reject elliptic curve lists of odd lengths.

The Supported Elliptic Curves extension contains a vector of NamedCurves
of 2 bytes each, so the total length must be even. Accepting odd-length
lists was observed to lead to a non-exploitable one-byte out-of-bounds
read in the latest development branches (1.0.2 and master). Released
versions of OpenSSL are not affected.

Thanks to Felix Groebert of the Google Security Team for reporting this issue.

Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 33d5ba862939ff8db70a9e36fc9a326fab3e8d98)

s_client and s_server take -verify_{host,email,ip}, not -check*

RT3596
Reviewed-by: Matt Caswell <matt@openssl.org>

Remove incorrect code inadvertently introduced through commit 59669b6ab.

Reviewed-by: Tim Hudson <tjh@openssl.org>

Remove "#if 0" code

Reviewed-by: Tim Hudson <tjh@openssl.org>
(cherry picked from commit 4bb8eb9ce4f794fecf020a15b54e8505fced0edf)

Only use the fallback mtu after 2 unsuccessful retransmissions if it is less
than the mtu we are already using

Reviewed-by: Tim Hudson <tjh@openssl.org>
(cherry picked from commit 047f21593eebbc617a410a208ded01e65ca11028)

Updates to s_client and s_server to remove the constant 28 (for IPv4 header
and UDP header) when setting an mtu. This constant is not always correct (e.g.
if using IPv6). Use the new DTLS_CTRL functions instead.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(cherry picked from commit 464ce92026bd0c79186cbefa75470f39607110be)

If we really get a situation where the underlying mtu is less than the minimum
we will support then dtls1_do_write can go into an infinite loop. This commit
fixes that.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(cherry picked from commit d3d9eef31661633f5b003a9e115c1822f79d1870)

Fix dtls_query_mtu so that it will always either complete with an mtu that is
at least the minimum or it will fail.
There were some instances in dtls1_query_mtu where the final mtu can end up
being less than the minimum, i.e. where the user has set an mtu manually. This
shouldn't be allowed. Also remove dtls1_guess_mtu that, despite having
logic for guessing an mtu, was actually only ever used to work out the minimum
mtu to use.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(cherry picked from commit 1620a2e49c777f31f2ce57966ae74006b48ad759)

Remove instances in libssl of the constant 28 (for size of IPv4 header + UDP)
and instead use the value provided by the underlying BIO. Also provide some
new DTLS_CTRLs so that the library user can set the mtu without needing to
know this constant. These new DTLS_CTRLs provide the capability to set the
link level mtu to be used (i.e. including this IP/UDP overhead). The previous
DTLS_CTRLs required the library user to subtract this overhead first.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(cherry picked from commit 59669b6abf620d1ed2ef4d1e2df25c998b89b64d)

Conflicts:
ssl/d1_both.c

There are a number of instances throughout the code where the constant 28 is
used with no explanation. Some of this was introduced as part of RT#1929. The
value 28 is the length of the IP header (20 bytes) plus the UDP header (8
bytes). However use of this constant is incorrect because there may be
instances where a different value is needed, e.g. an IPv4 header is 20 bytes
but an IPv6 header is 40. Similarly you may not be using UDP (e.g. SCTP).
This commit introduces a new BIO_CTRL that provides the value to be used for
this mtu "overhead". It will be used by subsequent commits.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(cherry picked from commit 0d3ae34df573f477b6b1aaf614d52dcdfcff5fce)

The first call to query the mtu in dtls1_do_write correctly checks that the
mtu that we have received is not less than the minimum. If its less it uses the
minimum instead. The second call to query the mtu does not do that, but
instead uses whatever comes back. We have seen an instance in RT#3592 where we
have got an unreasonably small mtu come back. This commit makes both query
checks consistent.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(cherry picked from commit 6abb0d1f8e702a0daa9c32b8021d01eda0483018)

The SSL_OP_NO_QUERY_MTU option is supposed to stop the mtu from being
automatically updated, and we should use the one provided instead.
Unfortunately there are a couple of locations where this is not respected.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(cherry picked from commit 001235778a6e9c645dc0507cad6092d99c9af8f5)

Verify that we have a sensible message len and fail if not
RT#3592 provides an instance where the OPENSSL_assert that this commit
replaces can be hit. I was able to recreate this issue by forcing the
underlying BIO to misbehave and come back with very small mtu values. This
happens the second time around the while loop after we have detected that the
MTU has been exceeded following the call to dtls1_write_bytes.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(cherry picked from commit cf75017bfd60333ff65edf9840001cd2c49870a3)

Use the SSLv23 method by default

If SSLv2 and SSLv3 are both disabled we still support SSL/TLS.

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

Check for FindNextFile when defining it rather than FindFirstFile

Reviewed-by: Matt Caswell <matt@openssl.org>

[PR3597] Advance to the next state variant when reusing messages.

Previously, state variant was not advanced, which resulted in state
being stuck in the st1 variant (usually "_A").

This broke certificate callback retry logic when accepting connections
that were using SSLv2 ClientHello (hence reusing the message), because
their state never advanced to SSL3_ST_SR_CLNT_HELLO_C variant required
for the retry code path.

Reported by Yichun Zhang (agentzh).

Signed-off-by: Piotr Sikora <piotr@cloudflare.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>

Correct some layout issues, convert all remaining tabs to appropriate amounts of spaces.

Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 8123d158ab6f0a6a468748c133e33c2063ff36b5)

Improves the proxy certificates howto doc.

The current documentation contains a bunch of spelling and grammar mistakes. I also
found it hard to understand some paragraphs, so here is my attempt to improve its
readability.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 03b637a730e4a298c360cc143de7564060c06324)

Fix warning in ssl2_enc

Reviewed-by: Richard Levitte <levitte@openssl.org>
(cherry picked from commit f9ea4deba006ee45a61e5f406f12c936b9bc162e)

Remove more references to dtls1_enc

Reviewed-by: Richard Levitte <levitte@openssl.org>
(cherry picked from commit eceef8fb865eb5de329b27ea472d4fdea4c290fe)

Check EVP_Cipher return values for SSL2

Reviewed-by: Richard Levitte <levitte@openssl.org>
(cherry picked from commit 81ec01b21767da5e5f0c03ad79b4d3dc9b632393)

Delete unused file

Reviewed-by: Richard Levitte <levitte@openssl.org>
(cherry picked from commit 4b87706d20f0a2fdf2e8f1b90256e141c487ef47)

Conflicts:
ssl/d1_enc.c

Add checks to the return value of EVP_Cipher to prevent silent encryption failure.

PR#1767

Reviewed-by: Richard Levitte <levitte@openssl.org>
(cherry picked from commit fe78f08d1541211566a5656395186bfbdc61b6f8)

Remove redundant checks in ssl_cert_dup. This was causing spurious error messages when using GOST

PR#3613

Reviewed-by: Richard Levitte <levitte@openssl.org>
(cherry picked from commit fc3968a25ce0c16cab8730ec0d68a59856158029)

Remove duplicated code

Reviewed-by: Dr. Stephen Henson <steve@openssl.org>

Tidy up ocsp help output

Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
(cherry picked from commit 5e31a40f47c6bfd09c718d2af42ba8d8fe6bb932)

Conflicts:
apps/ocsp.c

Add documentation on -timeout option in the ocsp utility

PR#3612

Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
(cherry picked from commit de87dd46c1283f899a9ecf4ccc72db74f36afbf2)

NetWare compilation fix.

Workaround for NetWare CodeWarrior compiler which doesn't properly lookup
includes when in same directory as the C file which includes it.

PR#3569
Reviewed-by: Stephen Henson <steve@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(cherry picked from commit 333fad9f2de1dea99552fcb424b312ca1a390f85)

Updates to X509_NAME_get_index_by_NID.pod submitted by user Bernardh via the wiki
Minor changes made by Matt Caswell

Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
(cherry picked from commit 648495864513da788246f9b75dbbbce0614ed5e8)

Updates to X509_NAME_add_entry_by_txt.pod submitted by user Bernardh via the wiki
Minor changes made by Matt Caswell.

Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
(cherry picked from commit f281b8df704ce0123fa2193f2890a25da1756528)

Updates to EVP_PKEY_encrypt.pod submitted by user Bernardh via the wiki
Minor changes made by Matt Caswell.

Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
(cherry picked from commit 34890ac18eb5ee7bffe9d460480164e1546b491e)

Add include of ssl.h which is required by srtp.h

Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
(cherry picked from commit f67203836cd4a35c1774b6147e0f2d33eb7b1b6f)

Fixed memory leak due to incorrect freeing of DTLS reassembly bit mask

PR#3608

Reviewed-by: Tim Hudson <tjh@openssl.org>
(cherry picked from commit 8a35dbb6d89a16d792b79b157b3e89443639ec94)

Corrected comments in ssl.h about SSLv23_method and friends

PR#3574

Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
(cherry picked from commit 3a0765882c4b3b67960b7efb203570764dd4ed29)

Print out Suite B status.

When using the -xcert option to test certificate validity print out
if we pass Suite B compliance. We print out "not tested" if we aren't
in Suite B mode.
Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 7d4cdededc371854eb36e773ed18204593e994e1)

Fix SuiteB chain checking logic.

Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 7255ca99df1f2d83d99d113dd5ca54b88d50e72b)

Do not resume a session if the negotiated protocol version does not match
the session's version (server).

See also BoringSSL's commit bdf5e72f50e25f0e45e825c156168766d8442dde.

Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
(cherry picked from commit 9e189b9dc10786c755919e6792e923c584c918a1)

When using EVP_PKEY_derive with a KDF set, a negative error from
ECDH_compute_key is silently ignored and the KDF is run on duff data

Thanks to github user tomykaira for the suggested fix.

Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
(cherry picked from commit 8d02bebddf4b69f7f260adfed4be4f498dcbd16c)

Clean up CHANGES

Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 31832e8ff1a3e731ea9fab41aef071a12709cf33)

Conflicts:
CHANGES

Ensure SSL3_FLAGS_CCS_OK (or d1->change_cipher_spec_ok for DTLS) is reset
once the ChangeCipherSpec message is received. Previously, the server would
set the flag once at SSL3_ST_SR_CERT_VRFY and again at SSL3_ST_SR_FINISHED.
This would allow a second CCS to arrive and would corrupt the server state.

(Because the first CCS would latch the correct keys and subsequent CCS
messages would have to be encrypted, a MitM attacker cannot exploit this,
though.)

Thanks to Joeri de Ruiter for reporting this issue.

Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit e94a6c0ede623960728415b68650a595e48f5a43)

Always require an advertised NewSessionTicket message.

The server must send a NewSessionTicket message if it advertised one
in the ServerHello, so make a missing ticket message an alert
in the client.

An equivalent change was independently made in BoringSSL, see commit
6444287806d801b9a45baf1f6f02a0e3a16e144c.

Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit de2c7504ebd4ec15334ae151a31917753468f86f)

Remove ssl3_check_finished.

The client sends a session ID with the session ticket, and uses
the returned ID to detect resumption, so we do not need to peek
at handshake messages: s->hit tells us explicitly if we're resuming.

An equivalent change was independently made in BoringSSL, see commit
407886f589cf2dbaed82db0a44173036c3bc3317.

Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 980bc1ec6114f5511b20c2e6ca741e61a39b99d6)

Conflicts:
ssl/s3_clnt.c

Set s->hit when resuming from external pre-shared secret.

The same change was independently made in BoringSSL, see commit
9eaeef81fa2d4fd6246dc02b6203fa936a5eaf67

Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 7b3ba508af5c86afe43e28174aa3c53a0a24f4d9)

Reset s->tlsext_ticket_expected in ssl_scan_serverhello_tlsext.
This ensures that it's zeroed even if the SSL object is reused
(as in ssltest.c). It also ensures that it applies to DTLS, too.

Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit a06cd5d056c6a5b1d161786873e21a5e53d554d8)