Fix enable-zlib

The enable-zlib option was broken by the recent "const" changes.

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

VMS: Use strict refdef extern model when building library object files

Most of the time, this isn't strictly needed.  However, in the default
extern model (called relaxed refdef), symbols are treated as weak
common objects unless they are initialised.  The librarian doesn't
include weak symbols in the (static) libraries, which renders them
invisible when linking a program with said those libraries, which is a
problem at times.

Using the strict refdef model is much more like standard C on all
other platforms, and thereby avoid the issues that come with the
relaxed refdef model.

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

Add test/bio_enc_test.c.

RT#4628

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

evp/bio_enc.c: refine non-overlapping logic.

RT#4628

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

ecp_nistz256.c: get is_one on 32-bit platforms right.

Thanks to Brian Smith for reporting this.

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

Move BIO index lock creation

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

update ordinals

Reviewed-by: Viktor Dukhovni <viktor@openssl.org>

Add X509_getm_notBefore, X509_getm_notAfter

Add mutable versions of X509_get0_notBefore and X509_get0_notAfter.

Rename X509_SIG_get0_mutable to X509_SIG_getm.

Reviewed-by: Viktor Dukhovni <viktor@openssl.org>

Duplicate includes

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1475)

Fix off by 1 in ASN1_STRING_set()

Reviewed-by: Rich Salz <rsalz@openssl.org>
MR: #3176

Add BIO_get_new_index()

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

fix warning about trailing comma

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

make update

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

rename ordinals

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

Constify certificate and CRL time routines.

Update certificate and CRL time routines to match new standard.

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

Add -dane_ee_no_namechecks s_client(1) option

The DANE API supports a DANE_FLAG_NO_DANE_EE_NAMECHECKS option, but
there was no way to exercise/enable it via s_client.  This commit
addresses that gap.

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

Set certificate times in one function.

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

Avoid duplicated code.

The certificate and CRL time setting functions used similar code,
combine into a single utility function.

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

RT3940: For now, just document the issue.

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

MEMPACKET is typedef'd in ssltestlib.h, don't do so again in ssltestlib.c

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

Allow to run all speed test when async_jobs active

... without any interruption.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1468)

make update

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

Convert PKCS12* functions to use const getters

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

Update function error code

A function error code needed updating due to merge issues.

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

Fix some clang warnings

Clang was complaining about some unused functions. Moving the stack
declaration to the header seems to sort it. Also the certstatus variable
in dtlstest needed to be declared static.

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

Fix DTLS replay protection

The DTLS implementation provides some protection against replay attacks
in accordance with RFC6347 section 4.1.2.6.

A sliding "window" of valid record sequence numbers is maintained with
the "right" hand edge of the window set to the highest sequence number we
have received so far. Records that arrive that are off the "left" hand
edge of the window are rejected. Records within the window are checked
against a list of records received so far. If we already received it then
we also reject the new record.

If we have not already received the record, or the sequence number is off
the right hand edge of the window then we verify the MAC of the record.
If MAC verification fails then we discard the record. Otherwise we mark
the record as received. If the sequence number was off the right hand edge
of the window, then we slide the window along so that the right hand edge
is in line with the newly received sequence number.

Records may arrive for future epochs, i.e. a record from after a CCS being
sent, can arrive before the CCS does if the packets get re-ordered. As we
have not yet received the CCS we are not yet in a position to decrypt or
validate the MAC of those records. OpenSSL places those records on an
unprocessed records queue. It additionally updates the window immediately,
even though we have not yet verified the MAC. This will only occur if
currently in a handshake/renegotiation.

This could be exploited by an attacker by sending a record for the next
epoch (which does not have to decrypt or have a valid MAC), with a very
large sequence number. This means the right hand edge of the window is
moved very far to the right, and all subsequent legitimate packets are
dropped causing a denial of service.

A similar effect can be achieved during the initial handshake. In this
case there is no MAC key negotiated yet. Therefore an attacker can send a
message for the current epoch with a very large sequence number. The code
will process the record as normal. If the hanshake message sequence number
(as opposed to the record sequence number that we have been talking about
so far) is in the future then the injected message is bufferred to be
handled later, but the window is still updated. Therefore all subsequent
legitimate handshake records are dropped. This aspect is not considered a
security issue because there are many ways for an attacker to disrupt the
initial handshake and prevent it from completing successfully (e.g.
injection of a handshake message will cause the Finished MAC to fail and
the handshake to be aborted). This issue comes about as a result of trying
to do replay protection, but having no integrity mechanism in place yet.
Does it even make sense to have replay protection in epoch 0? That
issue isn't addressed here though.

This addressed an OCAP Audit issue.

CVE-2016-2181

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

Add DTLS replay protection test

Injects a record from epoch 1 during epoch 0 handshake, with a record
sequence number in the future, to test that the record replay protection
feature works as expected. This is described more fully in the next commit.

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

Fix DTLS unprocessed records bug

During a DTLS handshake we may get records destined for the next epoch
arrive before we have processed the CCS. In that case we can't decrypt or
verify the record yet, so we buffer it for later use. When we do receive
the CCS we work through the queue of unprocessed records and process them.

Unfortunately the act of processing wipes out any existing packet data
that we were still working through. This includes any records from the new
epoch that were in the same packet as the CCS. We should only process the
buffered records if we've not got any data left.

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

Add a DTLS unprocesed records test

Add a test to inject a record from the next epoch during the handshake and
make sure it doesn't get processed immediately.

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

Split create_ssl_connection()

Split the create_ssl_connection() helper function into two steps: one to
create the SSL objects, and one to actually create the connection. This
provides the ability to make changes to the SSL object before the
connection is actually made.

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

Add a DTLS packet mem BIO

This adds a BIO similar to a normal mem BIO but with datagram awareness.
It also has the capability to inject additional packets at arbitrary
locations into the BIO, for testing purposes.

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

Add a (D)TLS dumper BIO

Dump out the records passed over the BIO. Only works for DTLS at the
moment but could easily be extended to TLS.

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

Add more details on how to add a new SSL test

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

make update

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

Add X509_get0_serialNumber() and constify OCSP_cert_to_id()

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

constify X509_REQ_get0_signature()

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

constify i2o_ECPublicKey

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

Sort %disabled in Configure

@disablables is sorted, but these were just added at the end of
%disabled in commits c2e27310 and 22e3dcb7.

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

Fix missing dane_tlsa_rrdata option error message

The error message said "dane_tlsa_rrset" instead of "dane_tlsa_rrdata".

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

Constify i2a*

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

Simplify indentation of DECLARE_ and IMPLEMENT_ lines

There's no reason we should enumerate every type of IMPLEMENT_ and
DECLARE_ line (and forget the ones we add a little now and then).
They all start with the same first word, let's just take'm all.

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

Indent ssl/

Run util/openssl-format-source on ssl/

Some comments and hand-formatted tables were fixed up
manually by disabling auto-formatting.

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

Convert X509_REVOKED* functions to use const getters

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

Test that the peers send at most one fatal alert

Duplicate alerts have happened, see
70c22888c1648fe8652e77107f3c74bf2212de36

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

Port multi-buffer tests

Make maximum fragment length configurable and add various fragmentation
tests, in addition to the existing multi-buffer tests.

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

Fix some doc nits.

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

Don't try to init dasync internally

Since dasync isn't installed, and is only ever used as a dynamic
engine, there's no reason to consider it for initialization when
building static engines.

Reviewed-by: Ben Laurie <ben@openssl.org>

make update

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

Constify X509_SIG.

Constify X509_SIG_get0() and order arguments to mactch new standard.

Add X509_SIG_get0_mutable() to support modification or initialisation
of an X509_SIG structure.

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

Small nits and cleanups

using util/openssl-format-source on s_derver, s_client, ca.c, speed.c only...

Fix/merge some #ifndef

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

Constify char* input parameters in apps code

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

Simplify and add help about OPT_PVK* options

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

Improve error message

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

Relocalise some globals variables

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

Constify ssl_cert_type()

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

Constify X509_certificate_type()

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

Constify X509_get0_signature()

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

Convert X509* functions to use const getters

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

Convert X509_CRL* functions to use const getters

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

Make X509_NAME_get0_der() conform to OpenSSL style

Put the main object first in the params list.

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

Corrupt signature in place.

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

Convert OCSP* functions to use const getters

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

Constify private key decode.

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

constify X509_ALGOR_get0()

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

Constify ASN1_item_unpack().

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

Add missing session id and tlsext_status accessors

 * SSL_SESSION_set1_id()
 * SSL_SESSION_get0_id_context()
 * SSL_CTX_get_tlsext_status_cb()
 * SSL_CTX_get_tlsext_status_arg()

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

dasync is an internal testing engine, so don't install it

Unfortunately, it means that the VMS IVP gets a bit crippled.  This
will be fixed later on.

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

VMS: no ENDIF on one line IF statements, in config.com

Correct small error from last config.com change

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

Convert SSL_SESSION* functions to use const getters

Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
Reviewed-by: Stephen Henson <steve@openssl.org>

Convert PKCS8* functions to use const getters

Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
Reviewed-by: Stephen Henson <steve@openssl.org>

Convert TS_STATUS_INFO* functions to use const getters

Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
Reviewed-by: Stephen Henson <steve@openssl.org>

two typo fixes

Reviewed-by: Emilia Käsper <emilia@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1461)

Fix compilation when using MASM on x86

The generated asm code from x86cpuid.pl contains CMOVE instructions
which are only available on i686 and later CPUs.

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1459)

Provide compat macros for SSL_CTX_set_ecdh_auto() and SSL_set_ecdh_auto()

These functions are no longer relevant to 1.1.0 (we always have auto ecdh
on) - but no reason to break old code that tries to call it. The macros will
only return a dummy "success" result if the app was trying to enable ecdh.
Disabling can't be done in quite this way any more.

Fixes Github Issue #1437

Reviewed-by: Viktor Dukhovni <viktor@openssl.org>

Ensure we unpad in constant time for read pipelining

The read pipelining code broke constant time unpadding. See GitHub
issue #1438

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

Corrupt signature earlier.

If -badsig is selected corrupt the signature before printing out
any details so the output reflects the modified signature.

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

make update

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

Add ASN1_STRING_get0_data(), deprecate ASN1_STRING_data().

Deprecate the function ASN1_STRING_data() and replace with a new function
ASN1_STRING_get0_data() which returns a constant pointer. Update library
to use new function.

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

Remove duplicate ordinals

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

ARMv8 assembly pack: add Samsung Mongoose results.

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

Configure: recognize -static as link option and disable incompatible options.

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

test/ssl_test.tmpl: make it work with elderly perl.

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

Fix satsub64be() to unconditionally use 64-bit integers

Now we support (u)int64_t this can be very much simpler.

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

SSL tests: send some application data

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

Add a "config" for verbosity and use it with Travis

Modify VMS config.com to match

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

Make "make" less verbose in Travis, except for the build only case

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

Limit reads in do_b2i_bio()

Apply a limit to the maximum blob length which can be read in do_d2i_bio()
to avoid excessive allocation.

Thanks to Shi Lei for reporting this.

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

Check for errors in a2d_ASN1_OBJECT()

Check for error return in BN_div_word().

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

Check for errors in BN_bn2dec()

If an oversize BIGNUM is presented to BN_bn2dec() it can cause
BN_div_word() to fail and not reduce the value of 't' resulting
in OOB writes to the bn_data buffer and eventually crashing.

Fix by checking return value of BN_div_word() and checking writes
don't overflow buffer.

Thanks to Shi Lei for reporting this bug.

CVE-2016-2182

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

Avoid truncating the pointer on x32 platform.

The 64 bit pointer must not be cast to 32bit unsigned long on
x32 platform.

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

Add a comment for the added cast with explanation.

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

Fix af_alg engine failure on 32 bit architectures.

Add extra cast to unsigned long to avoid sign extension when
converting pointer to 64 bit data.

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

Remove a stray unneeded line in 70-test_sslrecords.t

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

Address feedback on SSLv2 ClientHello processing

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

Add some SSLv2 ClientHello tests

Test that we handle a TLS ClientHello in an SSLv2 record correctly.

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

Send an alert if we get a non-initial record with the wrong version

If we receive a non-initial record but the version number isn't right then
we should send an alert.

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

Address feedback on SSLv2 ClientHello processing

Feedback on the previous SSLv2 ClientHello processing fix was that it
breaks layering by reading init_num in the record layer. It also does not
detect if there was a previous non-fatal warning.

This is an alternative approach that directly tracks in the record layer
whether this is the first record.

GitHub Issue #1298

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

Replaces CT_POLICY_EVAL_CTX_set0 entries with new setters in libcrypto.num

Reviewed-by: Emilia Käsper <emilia@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1408)

Make CT_POLICY_EVAL_CTX_set1_{cert,issuer} into boolean functions

They may fail if they cannot increment the reference count of the
certificate they are storing a pointer for. They should return 0 if this
occurs.

Reviewed-by: Emilia Käsper <emilia@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1408)

Improves CTLOG_STORE setters

Changes them to have clearer ownership semantics, as suggested in
https://github.com/openssl/openssl/pull/1372#discussion_r73232196.

Reviewed-by: Emilia Käsper <emilia@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1408)