Commit graph

3866 commits

Author SHA1 Message Date
Benjamin Kaduk
7bc2bddb14 Permit the "supported_groups" extension in ServerHellos
Although this is forbidden by all three(!) relevant specifications,
there seem to be multiple server implementations in the wild that
send it.  Since we didn't check for unexpected extensions in any
given message type until TLS 1.3 support was added, our previous
behavior was to silently accept these extensions and pass them over
to the custom extension callback (if any).  In order to avoid
regression of functionality, relax the check for "extension in
unexpected context" for this specific case, but leave the protocol
enforcment mechanism unchanged for other extensions and in other
extension contexts.

Leave a detailed comment to indicate what is going on.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4463)
2018-01-03 09:43:54 -06:00
Matt Caswell
818b625d6c Fix trace of TLSv1.3 Certificate Request message
A TLSv1.3 Certificate Request message was issuing a "Message length parse
error" using the -trace option to s_server/s_client.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/5008)
2018-01-02 15:51:23 +00:00
Daniel Bevenius
436ad81f05 Fix minor 'the the' typos
Similar to commit 17b602802114d53017ff7894319498934a580b17(
"Remove extra `the` in SSL_SESSION_set1_id.pod"), this commit removes
typos where additional 'the' have been added.

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4999)
2018-01-02 15:30:22 +00:00
Matt Caswell
f7414b0827 Disable partial writes for early data
We don't keep track of the number of bytes written between in the
SSL_write_ex() call and the subsequent flush. If the flush needs to be
retried then we will have forgotten how many bytes actually got written.
The simplest solution is to just disable it for this scenario.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4802)
2017-12-28 17:32:41 +00:00
Matt Caswell
2a8db71713 Don't flush the ClientHello if we're going to send early data
We'd like the first bit of early_data and the ClientHello to go in the
same TCP packet if at all possible to enable things like TCP Fast Open.
Also, if you're only going to send one block of early data then you also
don't need to worry about TCP_NODELAY.

Fixes 

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4802)
2017-12-28 17:32:41 +00:00
Bernd Edlinger
8175af50cc Alternate fix for ../test/recipes/80-test_ssl_old.t with no-ec
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4981)
2017-12-27 16:37:22 +01:00
Paul Yang
f69999262a Remove outdated comments
Variables n, d, p are no longer there.

[skip ci]

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4894)
2017-12-26 09:07:10 -05:00
Dr. Matthias St. Pierre
a93ba405b0 Add master DRBG for reseeding
A third shared DRBG is added, the so called master DRBG. Its sole purpose
is to reseed the two other shared DRBGs, the public and the private DRBG.
The randomness for the master DRBG is either pulled from the os entropy
sources, or added by the application using the RAND_add() call.

The master DRBG reseeds itself automatically after a given number of generate
requests, but can also be reseeded using RAND_seed() or RAND_add().
A reseeding of the master DRBG is automatically propagated to the public
and private DRBG. This construction fixes the problem, that up to now
the randomness provided by RAND_add() was added only to the public and
not to the private DRBG.

Signed-off-by: Dr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/4402)
2017-12-17 23:12:10 +01:00
Paul Yang
56d362881e Remove spaces at end of line in ssl/statem
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Matt Caswell <matt@openssl.org>
GH: 
2017-12-17 23:04:41 +01:00
Daniel Bevenius
b4ff66223b Add comments to NULL func ptrs in bio_method_st
This commit adds comments to bio_method_st definitions where the
function pointers are defined as NULL. Most of the structs have comments
but some where missing and not all consitent.

CLA: trivial

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4881)
2017-12-18 07:04:48 +10:00
Bernd Edlinger
fce78bd4ed Fix invalid function type casts.
Rename bio_info_cb to BIO_info_cb.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4493)
2017-12-15 19:33:48 +01:00
Matt Caswell
758e05c52e Make sure we treat records written after HRR as TLSv1.3
This fixes a bug where some CCS records were written with the wrong TLS
record version.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:38 +00:00
Matt Caswell
318d3c0e62 Issue a CCS from the client if we received an HRR
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:38 +00:00
Matt Caswell
75259b4346 Fix server side HRR flushing
Flush following the CCS after an HRR. Only flush the HRR if middlebox
compat is turned off.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:38 +00:00
Matt Caswell
5cc807da25 Delay flush until after CCS with early_data
Normally we flush immediately after writing the ClientHello. However if
we are going to write a CCS immediately because we've got early_data to
come, then we should move the flush until after the CCS.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:38 +00:00
Matt Caswell
7b0a3ce0f9 Ensure CCS sent before early_data has the correct record version
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:37 +00:00
Matt Caswell
426dfc9ff7 Send supported_versions in an HRR
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:37 +00:00
Matt Caswell
e7dd763e51 Make sure supported_versions appears in an HRR too
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:37 +00:00
Matt Caswell
fc7129dc37 Update state machine to send CCS based on whether we did an HRR
The CCS may be sent at different times based on whether or not we
sent an HRR earlier. In order to make that decision this commit
also updates things to make sure we remember whether an HRR was
used or not.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:37 +00:00
Matt Caswell
6f40214f68 Fix an HRR bug
Ensure that after an HRR we can only negotiate TLSv1.3

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:37 +00:00
Matt Caswell
597c51bc98 Merge HRR into ServerHello
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:37 +00:00
Matt Caswell
db37d32cb8 Send a CCS after ServerHello in TLSv1.3 if using middlebox compat mode
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:37 +00:00
Matt Caswell
066904ccee Send a CCS from a client in an early_data handshake
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:37 +00:00
Matt Caswell
4d02f87063 Send a CCS from the client in a non-early_data handshake
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:37 +00:00
Matt Caswell
fa9f9350f3 Remove TLSv1.3 specific write transition for ClientHello
Since we no longer do version negotiation during the processing of an HRR
we do not need the TLSv1.3 specific write transition for ClientHello

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:37 +00:00
Matt Caswell
fdd9236747 Drop CCS messages received in the TLSv1.3 handshake
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:37 +00:00
Matt Caswell
2d729db2f0 Send TLSv1.2 as the record version when using TLSv1.3
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:37 +00:00
Matt Caswell
a5816a5ab9 Implement session id TLSv1.3 middlebox compatibility mode
Clients will send a "fake" session id and servers must echo it back.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:37 +00:00
Matt Caswell
88050dd196 Update ServerHello to new draft-22 format
The new ServerHello format is essentially now the same as the old TLSv1.2
one, but it must additionally include supported_versions. The version
field is fixed at TLSv1.2, and the version negotiation happens solely via
supported_versions.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4701)
2017-12-14 15:06:37 +00:00
Daniel Bevenius
27ab91951c Make BIO_METHOD struct definitions consistent
I noticed that some of the BIO_METHOD structs are placing the name on
the same line as the type and some don't. This commit places the name
on a separate line for consistency (which looks like what the majority
do)

CLA: trivial

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4878)
2017-12-09 21:27:29 +01:00
Daniel Bevenius
6aff543b9b Correct minor typo in ssl_locl.h comment
CLA: trivial

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4880)
2017-12-08 12:02:29 -05:00
Matt Caswell
921d84a0ad Convert the remaining functions in the record layer to use SSLfatal()
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4841)
2017-12-08 16:42:02 +00:00
Matt Caswell
5591a6132e Convert dlts1_write_bytes() to use SSLfatal()
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4841)
2017-12-08 16:42:02 +00:00
Matt Caswell
c285338293 More record layer conversions to use SSLfatal()
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4841)
2017-12-08 16:42:02 +00:00
Matt Caswell
99dd374055 Convert ssl3_read_bytes() to use SSLfatal()
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4841)
2017-12-08 16:42:01 +00:00
Matt Caswell
196f2cbb78 Update ssl3_get_record() to use SSLfatal()
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4841)
2017-12-08 16:42:01 +00:00
Richard Levitte
0a90a6831e In OPENSSL_init_ssl(), run the base ssl init before OPENSSL_init_crypto()
IF OPENSSL_init_ssl() is called with the option flag
OPENSSL_INIT_LOAD_CONFIG, any SSL config will be handled wrongly
(i.e. there will be an attempt to load libssl_conf.so or whatever
corresponds to that on non-Unix platforms).  Therefore, at least
SSL_add_ssl_module() MUST be called before OPENSSL_init_crypto() is
called.  The base ssl init does that, plus adds all kinds of ciphers
and digests, which is harmless.

Fixes 

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4792)
2017-12-08 16:08:39 +01:00
Rich Salz
cbe2964821 Consistent formatting for sizeof(foo)
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4872)
2017-12-07 19:11:49 -05:00
Matt Caswell
e1dd8fa00a Fix the check_fatal macro
The check_fatal macro is supposed to only be called if we are already
expecting to be in the fatal state. The macro asserts that we are and
puts us into the fatal state if not.

This issue combined with the problem fixed in the previous commit meant
that the fuzzer detected a crash at a point in the processing when we
should have already been in the fatal state.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4847)
2017-12-06 14:56:28 +00:00
Matt Caswell
2894e9cb6d Fix bug in TLSv1.3 PSK processing
The recent SSL error overhaul left a case where an error occurs but
SSLfatal() is not called.

Credit to OSSfuzz for finding this issue.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4847)
2017-12-06 14:56:28 +00:00
Matt Caswell
dd5a4279f9 Fix some formatting nits
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4778)
2017-12-04 13:37:01 +00:00
Matt Caswell
e7d961e994 Remove spurious whitespace
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4778)
2017-12-04 13:31:48 +00:00
Matt Caswell
8e7677aeac Update an error reason code to be ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED
The most likely explanation for us ending up at this point in the code
is that we were called by the user application incorrectly - so use an
appropriate error code.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4778)
2017-12-04 13:31:48 +00:00
Matt Caswell
29bfd5b79a Add some more cleanups
Follow up from the conversion to use SSLfatal() in the state machine to
clean things up a bit more.

[extended tests]

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4778)
2017-12-04 13:31:48 +00:00
Matt Caswell
f9f674eb76 Assert that SSLfatal() only gets called once
We shouldn't call SSLfatal() multiple times for the same error condition.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4778)
2017-12-04 13:31:48 +00:00
Matt Caswell
47e2ee0722 Add some sanity checks for the fatal error condition
Sometimes at the top level of the state machine code we know we are
supposed to be in a fatal error condition. This commit adds some sanity
checks to ensure that SSLfatal() has been called.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4778)
2017-12-04 13:31:48 +00:00
Matt Caswell
635c8f7715 Fix up a few places in the state machine that got missed with SSLfatal()
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4778)
2017-12-04 13:31:48 +00:00
Matt Caswell
d273b60b41 Convert more functions in ssl/statem/statem_dtls.c to use SSLfatal()
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4778)
2017-12-04 13:31:48 +00:00
Matt Caswell
d4d2f3a4c1 Convert more functions in ssl/statem/statem.c to use SSLfatal()
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4778)
2017-12-04 13:31:48 +00:00
Matt Caswell
3ec8d113a5 Convert remaining functions in statem_srvr.c to use SSLfatal()
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4778)
2017-12-04 13:31:48 +00:00
Matt Caswell
a2c2e00050 Convert remaining functions in statem_clnt.c to use SSLfatal()
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4778)
2017-12-04 13:31:48 +00:00
Matt Caswell
f63a17d66d Convert the state machine code to use SSLfatal()
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4778)
2017-12-04 13:31:48 +00:00
Matt Caswell
4752c5deb2 Replace some usage of SSLerr with SSLfatal()
This is an initial step towards using SSLfatal() everywhere. Initially in
this commit and in subsequent commits we focus on the state machine code.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4778)
2017-12-04 13:31:48 +00:00
Matt Caswell
1f359471f9 Provide an SSLfatal() macro
Typically if a fatal error occurs three things need to happen:

- Put an error on the error queue
- Send an alert
- Put the state machine into the error state

Although all 3 of these things need to be done every time we hit a fatal
error the responsibilities for doing this are distributed throughout the
code. The place where the error goes on the queue, where the alert gets
sent and where the state machine goes into the error state are almost
invariably different. It has been a common pattern to pass alert codes up
and down the stack to get the alert information from the point in the code
where the error is detected to the point in the code where the alert gets
sent.

This commit provides an SSLfatal() macro (backed by an ossl_statem_fatal
function) that does all 3 of the above error tasks. This is largely a drop
in replacement for SSLerr, but takes a couple of extra parameters (the SSL
object, and an alert code).

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4778)
2017-12-04 13:31:48 +00:00
Todd Short
e670e90319 Add sk_new_reserve support
This is a specific 1.1.1 change; do not squash if the chacha
prioritization code is to be backported

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4436)
2017-11-30 07:13:09 +10:00
Todd Short
e1c7871de8 Use ChaCha only if prioritized by clnt
IFF the client has ChaCha first, and server cipher priority is used,
and the new SSL_OP_PRIORITIZE_CHACHA_FOR_MOBILE option is used,
then reprioritize ChaCha above everything else. This way, A matching
ChaCha cipher will be selected if there is a match. If no ChaCha ciphers
match, then the other ciphers are used.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4436)
2017-11-30 07:13:08 +10:00
Matt Caswell
881dfedeb5 Fix SSL_state_string() and SSL_state_string_long()
These functions needed updates for the various state machine states that
have been added for TLSv1.3.

Fixes 

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4801)
2017-11-28 10:24:23 +00:00
Matt Caswell
281bf2332c If a server is not acknowledging SNI then don't reject early_data
SNI needs to be consistent before we accept early_data. However a
server may choose to not acknowledge SNI. In that case we have to
expect that a client may send it anyway. We change the consistency
checks so that not acknowledging is treated more a like a "wild card",
accepting any SNI as being consistent.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4738)
2017-11-21 17:46:22 +00:00
Matt Caswell
bfab12bb7d Allow a client to send early_data with SNI if the session has no SNI
We can only send early_data if the SNI is consistent. However it is valid
for the client to set SNI and the server to not use it. This would still be
counted as consistent. OpenSSL client was being overzealous in this check
and disallowing this scenario.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4738)
2017-11-21 17:46:22 +00:00
Andy Polyakov
3a63c0edab Resolve warnings in VC-WIN32 build, which allows to add /WX.
It's argued that /WX allows to keep better focus on new code, which
motivates its comeback...

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4721)
2017-11-13 10:58:57 +01:00
Andy Polyakov
802127e8fc ssl/ssl_asn1.c: resolve warnings in VC-WIN32 build, which allows to add /WX.
It's argued that /WX allows to keep better focus on new code, which
motivates its comeback...

[Keep this commit separate as reminder for time overhaul.]

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4721)
2017-11-13 10:58:21 +01:00
Josh Soref
46f4e1bec5 Many spelling fixes/typo's corrected.
Around 138 distinct errors found and fixed; thanks!

Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3459)
2017-11-11 19:03:10 -05:00
FdaSilvaYY
f479eab227 style : fix some if(...
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4457)
2017-11-07 17:09:24 +01:00
Matt Caswell
1c47d35a03 Mark a zero length record as read
If SSL_read() is called with a zero length buffer, and we read a zero length
record then we should mark that record as read.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4685)
2017-11-07 11:07:17 +00:00
FdaSilvaYY
cf72c75792 Implement Maximum Fragment Length TLS extension.
Based on patch from Tomasz Moń:
https://groups.google.com/forum/#!topic/mailing.openssl.dev/fQxXvCg1uQY

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/1008)
2017-11-05 17:46:48 +01:00
Benjamin Kaduk
3be08e3011 Provide SSL_CTX.stats.sess_accept for switched ctxs
We currently increment the SSL_CTX stats.sess_accept field in
tls_setup_handshake(), which is invoked from the state machine well
before ClientHello processing would have had a chance to switch
the SSL_CTX attached to the SSL object due to a provided SNI value.
However, stats.sess_accept_good is incremented in tls_finish_handshake(),
and uses the s->ctx.stats field (i.e., the new SSL_CTX that was switched
to as a result of SNI processing).  This leads to the confusing
(nonsensical) situation where stats.sess_accept_good is larger than
stats.sess_accept, as the "sess_accept" value was counted on the
s->session_ctx.

In order to provide some more useful numbers, increment
s->ctx.stats.sess_accept after SNI processing if the SNI processing
changed s->ctx to differ from s->session_ctx.  To preserve the
property that any given accept is counted only once, make the
corresponding decrement to s->session_ctx.stats.sess_accept when
doing so.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/4549)
2017-10-30 10:21:10 -05:00
Benjamin Kaduk
0e6161bcae Normalize on session_ctx for stats where possible
For client SSL objects and before any callbacks have had a chance
to be called, we can write the stats accesses using the session_ctx,
which makes sense given that these values are all prefixed with
"sess_".

For servers after a client_hello or servername callback has been
called, retain the existing behavior of modifying the statistics
for the current (non-session) context.  This has some value,
in that it allows the statistics to be viewed on a per-vhost level.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/4549)
2017-10-30 10:21:10 -05:00
Benjamin Kaduk
1fcb4e4d52 Use atomics for SSL_CTX statistics
It is expected that SSL_CTX objects are shared across threads,
and as such we are responsible for ensuring coherent data accesses.
Aligned integer accesses ought to be atomic already on all supported
architectures, but we can be formally correct.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/4549)
2017-10-30 10:18:09 -05:00
Paul Yang
7a908204ed Simplify the stack reservation
Use the newly introduced sk_TYPE_new_reserve API to simplify the
reservation of stack as creating it.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4592)
2017-10-26 14:34:35 -04:00
Bernd Edlinger
e6b10c344e Fix error handling in SSL_new
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4580)
2017-10-25 17:41:46 +02:00
KaoruToda
b2555168ec Various clean-ups
Add a check for NULL return in t1_lib.c.
    Since return type of ssl_cert_lookup_by_idx is pointer and unify coding
    style, I changed from zero to NULL in ssl_cert.c.

Remove unnecessary space for ++.

Fix incorrect condition
    Expression is always false because 'else if' condition matches previous
    condition.  SInce the next line of 'else if' condition has substituted
    TLSEXT_ECPOINTFORMAT_ansiX962_compressed_char2, the 'else if'
    condition should compare with NID_X9_62_characteristic_two_field.

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4562)
2017-10-20 22:58:46 -04:00
KaoruToda
26a7d938c9 Remove parentheses of return.
Since return is inconsistent, I removed unnecessary parentheses and
unified them.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4541)
2017-10-18 16:05:06 +01:00
Benjamin Kaduk
2139145b72 Add missing RAND_DRBG locking
The drbg's lock must be held across calls to RAND_DRBG_generate()
to prevent simultaneous modification of internal state.

This was observed in practice with simultaneous SSL_new() calls attempting
to seed the (separate) per-SSL RAND_DRBG instances from the global
rand_drbg instance; this eventually led to simultaneous calls to
ctr_BCC_update() attempting to increment drbg->bltmp_pos for their
respective partial final block, violating the invariant that bltmp_pos < 16.
The AES operations performed in ctr_BCC_blocks() makes the race window
quite easy to trigger.  A value of bltmp_pos greater than 16 induces
catastrophic failure in ctr_BCC_final(), with subtraction overflowing
and leading to an attempt to memset() to zero a very large range,
which eventually reaches an unmapped page and segfaults.

Provide the needed locking in get_entropy_from_parent(), as well as
fixing a similar issue in RAND_priv_bytes().  There is also an
unlocked call to RAND_DRBG_generate() in ssl_randbytes(), but the
requisite serialization is already guaranteed by the requirements on
the application's usage of SSL objects, and no further locking is
needed for correct behavior.  In that case, leave a comment noting
the apparent discrepancy and the reason for its safety (at present).

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4328)
2017-10-18 08:39:20 -05:00
Dr. Matthias St. Pierre
c16de9d832 Fix reseeding issues of the public RAND_DRBG
Reseeding is handled very differently by the classic RAND_METHOD API
and the new RAND_DRBG api. These differences led to some problems when
the new RAND_DRBG was made the default OpenSSL RNG. In particular,
RAND_add() did not work as expected anymore. These issues are discussed
on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API'
and in Pull Request . This commit fixes the mentioned issues,
introducing the following changes:

- Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which
  facilitates collecting entropy by the get_entropy() callback.
- Don't use RAND_poll()/RAND_add() for collecting entropy from the
  get_entropy() callback anymore. Instead, replace RAND_poll() by
  RAND_POOL_acquire_entropy().
- Add a new function rand_drbg_restart() which tries to get the DRBG
  in an instantiated state by all means, regardless of the current
  state (uninstantiated, error, ...) the DRBG is in. If the caller
  provides entropy or additional input, it will be used for reseeding.
- Restore the original documented behaviour of RAND_add() and RAND_poll()
  (namely to reseed the DRBG immediately) by a new implementation based
  on rand_drbg_restart().
- Add automatic error recovery from temporary failures of the entropy
  source to RAND_DRBG_generate() using the rand_drbg_restart() function.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4328)
2017-10-18 08:39:20 -05:00
Matt Caswell
beb30941d6 Tweak the comment regarding record version check with respect to TLSv1.3
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4527)
2017-10-16 15:52:19 +01:00
Matt Caswell
61278ff3f9 Sanity check the HRR version field
The previous commit removed version negotiation on an HRR. However we should
still sanity check the contents of the version field.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4527)
2017-10-16 15:52:19 +01:00
Matt Caswell
a2b97bdf3d Don't do version neg on an HRR
Previously if a client received an HRR then we would do version negotiation
immediately - because we know we are going to get TLSv1.3. However this
causes a problem when we emit the 2nd ClientHello because we start changing
a whole load of stuff to ommit things that aren't relevant for < TLSv1.3.
The spec requires that the 2nd ClientHello is the same except for changes
required from the HRR. Therefore the simplest thing to do is to defer the
version negotiation until we receive the ServerHello.

Fixes 

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4527)
2017-10-16 15:52:18 +01:00
Matt Caswell
9fb6cb810b Fix bug where early_data does not work if no SNI callback is present
Fixes 

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4519)
2017-10-12 15:19:49 +01:00
Benjamin Kaduk
f9df0a7775 Move supportedgroup ext-block fields out of NO_EC
Now that we are moving to support named FFDH groups, these fields are not
ec-specific, so we need them to always be available.

This fixes the no-ec --strict-warnings build, since gcc
5.4.0-6ubuntu1~16.04.4 appears to always try to compile the static inline
functions from ssl_locl.h, even when they are not used in the current
compilation unit.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4518)
2017-10-11 08:25:40 -05:00
Andy Polyakov
32f3b98d13 crypto/x509v3/v3_utl.c, ssl/ssl_cert.c: fix Coverity problems.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4492)
2017-10-10 20:00:35 +02:00
Tatsuhiro Tsujikawa
751b26b132 Don't change client random in Client Hello in its second flight
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4490)
2017-10-10 09:51:53 +01:00
KaoruToda
208fb891e3 Since return is inconsistent, I removed unnecessary parentheses and
unified them.
- return (0); -> return 0;
- return (1); -> return 1;
- return (-1); -> return -1;

Reviewed-by: Stephen Henson <steve@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4500)
2017-10-09 13:17:09 +01:00
Dr. Stephen Henson
6447e8184c Merge tls1_check_curve into tls1_check_group_id
The function tls_check_curve is only called on clients and contains
almost identical functionaity to tls1_check_group_id when called from
a client. Merge the two.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4475)
2017-10-06 19:09:51 +01:00
Dr. Stephen Henson
f48d826e33 Change curves to groups where relevant
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4475)
2017-10-06 13:23:45 +01:00
Dr. Stephen Henson
ff6d20a67b Use separate functions for supported and peer groups lists
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4475)
2017-10-06 13:23:45 +01:00
Matt Caswell
786b4df402 Remove an incorrect comment
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4456)
2017-10-04 15:56:56 +01:00
Todd Short
a84e5c9aa8 Session resume broken switching contexts
When an SSL's context is swtiched from a ticket-enabled context to
a ticket-disabled context in the servername callback, no session-id
is generated, so the session can't be resumed.

If a servername callback changes the SSL_OP_NO_TICKET option, check
to see if it's changed to disable, and whether a session ticket is
expected (i.e. the client indicated ticket support and the SSL had
tickets enabled at the time), and whether we already have a previous
session (i.e. s->hit is set).

In this case, clear the ticket-expected flag, remove any ticket data
and generate a session-id in the session.

If the SSL hit (resumed) and switched to a ticket-disabled context,
assume that the resumption was via session-id, and don't bother to
update the session.

Before this fix, the updated unit-tests in 06-sni-ticket.conf would
fail test  (server1 = SNI, server2 = no SNI).

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/1529)
2017-10-04 10:21:08 +10:00
Pauli
f32b0abe26 Remove unnecessary #include <openssl/lhash.h> directives.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4431)
2017-09-29 07:38:56 +10:00
Pauli
1f5e0f9280 Use safestack.h exclusively internally.
Remove all stack headers from some includes that don't use them.

Avoid a genearic untyped stack use.

Update stack POD file to include the OPENSSL_sk_ API functions in the notes
section.  They were mentioned in the name section but not defined anywhere.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4430)
2017-09-28 13:14:33 +10:00
Pauli
e431363f8c Add stack space reservations.
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4386)
2017-09-28 06:53:40 +10:00
Dr. Stephen Henson
b50951d36c Add and use function tls1_in_list to avoid code duplication.
[extended tests]

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/=4412)
2017-09-26 13:00:26 +01:00
Dr. Stephen Henson
5ce5f787e3 Use tls1_group_id_lookup in tls1_curve_allowed
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/=4412)
2017-09-26 13:00:26 +01:00
Dr. Stephen Henson
34e5292c57 Rename tls1_get_curvelist.
Rename tls1_get_curvelist to tls1_get_grouplist, change to void as
it can never fail and remove unnecessary return value checks. Clean
up the code.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/=4412)
2017-09-26 13:00:26 +01:00
Dr. Stephen Henson
4a1b428019 Rewrite compression and group checks.
Replace existing compression and groups check with two functions.

tls1_check_pkey_comp() checks a keys compression algorithms is consistent
with extensions.

tls1_check_group_id() checks is a group is consistent with extensions
and preferences.

Rename tls1_ec_nid2curve_id() to tls1_nid2group_id() and make it static.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/=4412)
2017-09-26 13:00:26 +01:00
Dr. Stephen Henson
612f9d2227 New function ssl_generate_param_group
Setup EVP_PKEY structure from a group ID in ssl_generate_param_group,
replace duplicate code with this function.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/=4412)
2017-09-26 13:00:26 +01:00
Dr. Stephen Henson
43b95d7365 Replace tls1_ec_curve_id2nid.
Replace tls1_ec_curve_id2nid() with tls_group_id_lookup() which returns
the TLS_GROUP_INFO for the group.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/=4412)
2017-09-26 13:00:26 +01:00
Dr. Stephen Henson
0e464d9ddd Rename tls_curve_info to TLS_GROUP_INFO, move to ssl_locl.h
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/=4412)
2017-09-26 13:00:26 +01:00
Dr. Stephen Henson
8841154a90 Return group id in tls1_shared_group
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/=4412)
2017-09-26 13:00:26 +01:00
Dr. Stephen Henson
1483b8582c Return correct Suite B curve, fix comment.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/=4412)
2017-09-26 13:00:26 +01:00
Kurt Roeckx
b92d7b62f5 Use size of entries, not size of the pointer.
Reviewed-by: Andy Polyakov <appro@openssl.org>
GH: 
2017-09-23 17:20:05 +02:00
Kurt Roeckx
9b01779cbf Use curve_id not the nid
Found by OSS-Fuzz and the tests

Reviewed-by: Andy Polyakov <appro@openssl.org>
GH: 
2017-09-23 17:20:04 +02:00
Dr. Stephen Henson
9e84a42db4 Store groups as uint16_t
Instead of storing supported groups in on-the-wire format store
them as parsed uint16_t values. This simplifies handling of groups
as the values can be directly used instead of being converted.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4406)
2017-09-22 22:59:42 +01:00
Dr. Matthias St. Pierre
79b4444d81 Cleanup whitespace in ssl_lib.c (tabs to spaces)
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4383)
2017-09-21 23:42:10 +02:00
Patrick Steuer
b9ff0483bd Fix strict-warnings build
Compilation failed due to -Werror=misleading-indentation.

Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4395)
2017-09-21 16:01:22 -05:00
Dr. Stephen Henson
6aaa29fb35 Allow use of RSA-PSS certificates in TLS 1.2
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4368)
2017-09-20 12:50:23 +01:00
Dr. Stephen Henson
b46867d771 Allow RSA certificates to be used for RSA-PSS
Allo RSA certificate to be used for RSA-PSS signatures: this needs
to be explicit because RSA and RSA-PSS certificates are now distinct
types.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4368)
2017-09-20 12:50:23 +01:00
Dr. Stephen Henson
045d078aef Add RSA-PSS key certificate type.
Recognise RSA-PSS certificate algorithm and add a new certificate
type.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4368)
2017-09-20 12:50:23 +01:00
Christian Heimes
3edabd3ccb Provide getters for min/max proto version
OpenSSL 1.1.0 made SSL_CTX and SSL structs opaque and introduced a new
API to set the minimum and maximum protocol version for SSL_CTX with
TLS_method(). Add getters to introspect the configured versions:

  int SSL_CTX_get_min_proto_version(SSL_CTX *ctx);
  int SSL_CTX_get_max_proto_version(SSL_CTX *ctx);
  int SSL_get_min_proto_version(SSL *ssl);
  int SSL_get_max_proto_version(SSL *ssl);

NOTE: The getters do not resolv the version in case when the minimum or
maxium version are configured as '0' (meaning auto-select lowest and
highst version number).

Signed-off-by: Christian Heimes <christian@python.org>

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4364)
2017-09-15 10:04:47 -05:00
Benjamin Kaduk
f5d270cab5 Fix no-ec no-dh build
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4369)
2017-09-14 10:23:20 -05:00
Pauli
4cacc9d510 Revert "GH614: Use memcpy()/strdup() when possible"
This reverts commit a89c9a0d85.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4357)
2017-09-14 10:26:54 +10:00
FdaSilvaYY
06312c4769 SSL Trace improvements
A fix formatting fixes.

SSL Trace: internal constification
-  static trace tables
-  trace methods arguments

SSL Trace: enhance error message when tracing an invalid extension packet
... instead of just "Message length parse error!".

SSL trace: add Maximum-Fragment-Length TLS extension log support

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4353)
2017-09-10 23:33:37 -04:00
David Benjamin
f1b97da1fd Introduce named constants for the ClientHello callback.
It is otherwise unclear what all the magic numbers mean.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4349)
2017-09-08 13:58:59 -05:00
David Benjamin
a9c0d8beea Rename SSL_CTX_set_early_cb to SSL_CTX_set_client_hello_cb.
"Early callback" is a little ambiguous now that early data exists.
Perhaps "ClientHello callback"?

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4349)
2017-09-08 13:58:59 -05:00
Benjamin Kaduk
1c259bb518 Restore historical behavior for absent ServerHello extensions
In OpenSSL 1.1.0, when there were no extensions added to the ServerHello,
we did not write the extension data length bytes to the end of the
ServerHello; this is needed for compatibility with old client implementations
that do not support TLS extensions (such as the default configuration of
OpenSSL 0.9.8).  When ServerHello extension construction was converted
to the new extensions framework in commit
7da160b0f4, this behavior was inadvertently
limited to cases when SSLv3 was negotiated (and similarly for ClientHellos),
presumably since extensions are not defined at all for SSLv3.  However,
extensions for TLS prior to TLS 1.3 have been defined in separate
RFCs (6066, 4366, and 3546) from the TLS protocol specifications, and as such
should be considered an optional protocol feature in those cases.

Accordingly, be conservative in what we send, and skip the extensions block
when there are no extensions to be sent, regardless of the TLS/SSL version.
(TLS 1.3 requires extensions and can safely be treated differently.)

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/4296)
2017-09-07 09:53:21 -05:00
Alfred E. Heggestad
fa4b82cc7c add callback handler for setting DTLS timer interval
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/4011)
2017-09-06 08:30:00 +02:00
Matt Caswell
3d85c7f408 Don't attempt to add a zero length padding extension
The padding extension should always be at least 1 byte long

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/4331)
2017-09-04 15:15:30 +01:00
Andy Polyakov
89bc9cf682 ssl/statem/extensions_clnt.c: fix return code buglet.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4297)
2017-09-01 08:44:41 +02:00
Andy Polyakov
eb5fd03bb2 ssl/statem/*.c: address "enum mixed with another type" warnings.
This is actually not all warnings, only return values.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4297)
2017-09-01 08:43:45 +02:00
Pauli
4cff10dcbf struct timeval include guards
Move struct timeval includes into e_os.h (where the Windows ones were).
Enaure that the include is guarded canonically.

Refer 

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4312)
2017-09-01 09:55:43 +10:00
Rich Salz
ccb76685e0 Avoid possible uninitialized variable.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4307)
2017-08-31 18:27:06 -04:00
Matt Caswell
0ef2802165 Various review fixes for PSK early_data support
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3926)
2017-08-31 15:03:35 +01:00
Matt Caswell
4be3a7c7aa Client side sanity check of ALPN after server has accepted early_data
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3926)
2017-08-31 15:03:35 +01:00
Matt Caswell
fff202e5f7 Add some fixes for Travis failures
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3926)
2017-08-31 15:03:35 +01:00
Matt Caswell
630369d9ce Add server side sanity checks of SNI/ALPN for use with early_data
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3926)
2017-08-31 15:03:35 +01:00
Matt Caswell
ae8d7d994a Make sure we save ALPN data in the session
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3926)
2017-08-31 15:03:35 +01:00
Matt Caswell
ffc5bbaaee Complain if we are writing early data but SNI or ALPN is incorrect
SNI and ALPN must be set to be consistent with the PSK. Otherwise this is
an error.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3926)
2017-08-31 15:03:34 +01:00
Matt Caswell
67738645dc Add functions for getting/setting SNI/ALPN info in SSL_SESSION
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3926)
2017-08-31 15:03:34 +01:00
Matt Caswell
0871754499 Make sure we use the correct cipher when using the early_secret
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3926)
2017-08-31 15:02:58 +01:00
Matt Caswell
98e1d93454 Add SSL_SESSION_set_max_early_data()
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3926)
2017-08-31 15:02:58 +01:00
Matt Caswell
add8d0e9e0 Enable the ability to use an external PSK for sending early_data
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3926)
2017-08-31 15:02:22 +01:00
Pauli
ea78d1ec2c Add ARIA as an alias for all ARIA based modes.
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4300)
2017-08-31 08:40:11 +10:00
Paul Yang
50966bfa11 Introduce SSL_CIPHER_get_protocol_id
The returned ID matches with what IANA specifies (or goes on the
wire anyway, IANA notwithstanding).

Doc is added.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4107)
2017-08-30 14:01:13 -05:00
Jon Spillett
bc32673869 Implement Aria GCM/CCM Modes and TLS cipher suites
AEAD cipher mode implementation is based on that used for AES:

  https://tools.ietf.org/html/rfc5116

TLS GCM cipher suites as specified in:

  https://tools.ietf.org/html/rfc6209

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4287)
2017-08-30 12:33:53 +02:00
Pauli
198c42f5af Move OPENSSL_CONF from e_os.h to cryptlib.h
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4188)
2017-08-30 07:20:44 +10:00
Pauli
07016a8a31 Move e_os.h to be the very first include.
cryptilib.h is the second.

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4188)
2017-08-30 07:20:44 +10:00
Pauli
cd420b0b1f Move the REF_PRINT support from e_os.h to internal/refcount.h.
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4188)
2017-08-30 07:20:44 +10:00
Pauli
677963e5a4 e_os.h removal from other headers and source files.
Removed e_os.h from all bar three headers (apps/apps.h crypto/bio/bio_lcl.h and
ssl/ssl_locl.h).

Added e_os.h into the files that need it now.

Directly reference internal/nelem.h when required.

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4188)
2017-08-30 07:20:43 +10:00
Kurt Roeckx
0b14a5b7cc Don't auto-instantiate a DRBG when trying to use it and it's not
The one creating the DRBG should instantiate it, it's there that we
know which parameters we should use to instantiate it.

This splits the rand init in two parts to avoid a deadlock
because when the global drbg is created it wands to call
rand_add on the global rand method.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
GH: 
2017-08-28 23:15:52 +02:00
Richard Levitte
b379fe6cd0 NO_SYS_TYPES_H isn't defined anywhere, stop using it as a guard
This is a vestige from pre-1.1.0 OpenSSL

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4256)
2017-08-25 15:50:59 +02:00
Kazuki Yamaguchi
0afca8113e Do not lookup zero-length session ID
A condition was removed by commit 1053a6e2281d; presumably it was an
unintended change. Restore the previous behavior so the get_session_cb
won't be called with zero-length session ID.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/4236)
2017-08-25 14:36:07 +02:00
Rich Salz
0e97f1e1a7 (Re)move some things from e_os.h
Remove GETPID_IS_MEANINGLESS and osslargused.

Move socket-related things to new file internal/sockets.h; this is now
only needed by four(!!!) files.  Compiles should be a bit faster.
Remove USE_SOCKETS ifdef's

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4209)
2017-08-22 14:15:40 -04:00
Pauli
00dfbaad88 Fix ctype arguments.
Cast arguments to the various ctype functions to unsigned char to match their
documentation.

Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4203)
2017-08-22 07:35:08 +10:00
Todd Short
10ed1b7239 Reorder extensions to put SigAlgs last
Force non-empty padding extension.
When enabled, force the padding extension to be at least 1 byte long.
WebSphere application server cannot handle having an empty
extension (e.g. EMS/EtM) as the last extension in a client hello.
This moves the SigAlgs extension last for TLSv1.2 to avoid this
issue.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3921)
2017-08-18 09:52:17 -04:00
Benjamin Kaduk
0aed6e449d Add SSL_get_pending_cipher()
The existing function SSL_get_current_cipher() queries the
current session for the ciphersuite in use, but there is no way
for application code to determine what ciphersuite has been
negotiated and will be used in the future, prior to ChangeCipherState
(or the TLS 1.3 equivalent) causing the new cipher to take effect and
become visible in the session information.  Expose this information
to appropriate application callbacks to use during the handshake.

The name SSL_get_pending_cipher() was chosen for compatibility with
BoringSSL's routine of that name.

Improve the note on macro implementations in SSL_get_current_cipher.pod
while here.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4070)
2017-08-15 10:52:21 -05:00
Benjamin Kaduk
5626f634c3 Move ALPN handling from finalizer to delayed call
Commit 02f0274e8c moved ALPN processing
into an extension finalization function, as the only documented ordering
requirement from previous commits was that ALPN processing occur after
SNI processing, and SNI processing is performed before the extension
finalization step.  However, it is useful for applications'
alpn_select callbacks to run after ciphersuite selection as well -- at
least one application protocol specification (HTTP/2) imposes restrictions
on which ciphersuites are usable with that protocol.  Since it is generally
more preferrable to have a successful TLS connection with a default application
protocol than to fail the TLS connection and not be able to have the preferred
application protocol, it is good to give the alpn_select callback information
about the ciphersuite to be used, so that appropriate restrctions can be
enforced in application code.

Accordingly, split the ALPN handling out into a separate tls_handl_alpn()
function akin to tls_handle_status_request(), called from
tls_post_process_client_hello().  This is an alternative to resuscitating
ssl_check_clienthello_tlsext_late(), something of an awkwward name itself.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4070)
2017-08-15 10:52:21 -05:00
FdaSilvaYY
44e6995155 Fix some Typos and indents
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4108)
2017-08-11 10:16:33 -04:00
Benjamin Kaduk
e3743355e8 Don't modify resumed session objects
If s->hit is set, s->session corresponds to a session created on
a previous connection, and is a data structure that is potentially
shared across other SSL objects.  As such, there are thread-safety
issues with modifying the structure without taking its lock (and
of course all corresponding read accesses would also need to take
the lock as well), which have been observed to cause double-frees.

Regardless of thread-safety, the resumed session object is intended
to reflect parameters of the connection that created the session,
and modifying it to reflect the parameters from the current connection
is confusing.  So, modifications to the session object during
ClientHello processing should only be performed on new connections,
i.e., those where s->hit is not set.

The code mostly got this right, providing such checks when processing
SNI and EC point formats, but the supported groups (formerly
supported curves) extension was missing it, which is fixed by this commit.

However, TLS 1.3 makes the suppported_groups extension mandatory
(when using (EC)DHE, which is the normal case), checking for the group
list in the key_share extension processing.  But, TLS 1.3 only [0] supports
session tickets for session resumption, so the session object in question
is the output of d2i_SSL_SESSION(), and will not be shared across SSL
objects.  Thus, it is safe to modify s->session for TLS 1.3 connections.

[0] A psk_find_session callback can also be used, but the restriction that
each callback execution must produce a distinct SSL_SESSION structure
can be documented when the psk_find_session callback documentation is
completed.

Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4123)
2017-08-09 14:54:47 -05:00
Matt Caswell
07927bedf3 Add an SSL_SESSION_dup() function
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/4072)
2017-08-09 13:37:06 +10:00
Rich Salz
5f8dd0f849 Add missing include of cryptlib.h
Also use "" not <> for all include cryptlib

Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4082)
2017-08-06 17:06:19 -04:00
Rich Salz
ae3947de09 Add a DRBG to each SSL object
Give each SSL object it's own DRBG, chained to the parent global
DRBG which is used only as a source of randomness into the per-SSL
DRBG.  This is used for all session, ticket, and pre-master secret keys.
It is NOT used for ECDH key generation which use only the global
DRBG. (Doing that without changing the API is tricky, if not impossible.)

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/4050)
2017-08-03 10:24:03 -04:00
Rich Salz
75e2c87765 Switch from ossl_rand to DRBG rand
If RAND_add wraps around, XOR with existing. Add test to drbgtest that
does the wrap-around.

Re-order seeding and stop after first success.

Add RAND_poll_ex()

Use the DF and therefore lower RANDOMNESS_NEEDED.  Also, for child DRBG's,
mix in the address as the personalization bits.

Centralize the entropy callbacks, from drbg_lib to rand_lib.
(Conceptually, entropy is part of the enclosing application.)
Thanks to Dr. Matthias St Pierre for the suggestion.

Various code cleanups:
    -Make state an enum; inline RANDerr calls.
    -Add RAND_POLL_RETRIES (thanks Pauli for the idea)
    -Remove most RAND_seed calls from rest of library
    -Rename DRBG_CTX to RAND_DRBG, etc.
    -Move some code from drbg_lib to drbg_rand; drbg_lib is now only the
     implementation of NIST DRBG.
    -Remove blocklength

Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/4019)
2017-08-03 09:23:28 -04:00
Matt Caswell
67dc995eaf Move ossl_assert
Move the definition of ossl_assert() out of e_os.h which is intended for OS
specific things. Instead it is moved into internal/cryptlib.h.

This also changes the definition to remove the (int) cast.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4073)
2017-08-03 10:48:00 +01:00
Lingmo Zhu
5bd05e5799 remove horrible pragma macro and remove __owur from SSL_CTX_add_session() declaration
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4014)
2017-08-03 10:14:46 +01:00
Lingmo Zhu
e29bb83479 Remove the obsolete misleading comment and code related to it.
The comment "The following should not return 1, otherwise, things
are very strange" is from the very first commit of OpenSSL. The
really meaning of the comment is if the identical session can be
found from internal cache after calling get_session_cb but not
found before calling get_session_cb, it is just strange.

The value 1 was originated from the old doc of SSLeay, reversed
from the actual return value of SSL_CTX_add_session().

Anyway either return value of SSL_CTX_add_session() should not
interrupt the session resumption process. So the checking of
return value of SSL_CTX_add_session() is not necessary.

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4014)
2017-08-03 10:14:46 +01:00
Matt Caswell
5d61491c88 Fix new_session_cb calls in TLSv1.3
If a new_session_cb is set then it was only ever getting invoked if !s->hit
is true. This is sensible for <=TLSv1.2 but does not work for TLSv1.3.

Fixes 

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/4068)
2017-08-01 13:09:31 +01:00
Todd Short
b93a295a36 Fix SSL_set_tlsext_debug_callback/-tlsextdebug
Some extensions were being displayed twice, before they were parsed, and
again after they were parsed.
The supported_versions extension was not being fully displayed, as it
was processed differently than other extensions.

Move the debug callback to where the extensions are first collected, to
catch all the extensions as they come in, so they are ordered correctly.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3911)
2017-07-31 13:38:05 -04:00
David Benjamin
ee1ed1d380 Fix the names of older ciphers.
The names of these ciphers have an "SSL_" prefix, but the RFC names use
"TLS_":
https://www.iana.org/assignments/tls-parameters/tls-parameters.xhtml#tls-parameters-4

This dates back to these ciphers being originally defined in SSLv3. As
SSLv3 is on its way out anyway and this is a new set of APIs,
consistently use the TLS names.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4007)
2017-07-31 09:11:18 -04:00
Xiaoyin Liu
f978f2b8af Fix errors in SSL_state_string_long
TLS_ST_SR_NEXT_PROTO means "SSLv3/TLS read next proto"
Fix typo in the message for TLS_ST_SW_CERT_STATUS

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4054)
2017-07-31 08:55:37 -04:00
Xiaoyin Liu
3519bae518 Fix typos in files in ssl directory
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4052)
2017-07-30 18:43:19 -04:00
Xiaoyin Liu
f00d0fd783 Update copyright header
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4049)
2017-07-30 17:42:00 -04:00
Xiaoyin Liu
3c051806ce Remove redundant declarations in ssl_locl.h
Remove the function prototypes for ssl_cert_get0_next_certificate, ssl_set_default_md, tls1_shared_list,
dtls1_send_newsession_ticket, tls1_ctrl, and tls1_callback_ctrl, all of which are not defined.

It also changed the signature of the function pqueue_next to `pitem *pqueue_next(piterator *item)` in
pqueue.c, making it match the prototype in ssl_locl.h. (`piterator *` is equivalent to `pitem **`.)

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4049)
2017-07-30 17:42:00 -04:00
Xiaoyin Liu
00f3a013c3 Remove redundant declarations in record_locl.h
This patch removes the prototype of function RECORD_LAYER_set_write_sequence from record_locl.h, since this function is not defined.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4051)
2017-07-30 17:40:56 -04:00
Paul Yang
190b9a03b7 Make SSL_set_tlsext_host_name no effect from server side
Documentation and test cases are also updated

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3783)
2017-07-28 11:51:59 -04:00
Todd Short
a58eb06d52 Add support to free/allocate SSL buffers
OpenSSL already has the feature of SSL_MODE_RELEASE_BUFFERS that can
be set to release the read or write buffers when data has finished
reading or writing. OpenSSL will automatically re-allocate the buffers
as needed. This can be quite aggressive in terms of memory allocation.

This provides a manual mechanism. SSL_free_buffers() will free
the data buffers if there's no pending data. SSL_alloc_buffers()
will realloc them; but this function is not strictly necessary, as it's
still done automatically in the state machine.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2240)
2017-07-26 11:42:17 -04:00
Paul Yang
bbb4ceb86e Support converting cipher name to RFC name and vice versa
Fixes: issue 

make SSL_CIPHER_standard_name globally available and introduce a new
function OPENSSL_cipher_name.

A new option '-convert' is also added to 'openssl ciphers' app.

Documentation and test cases are added.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/3859)
2017-07-21 07:20:14 +10:00
Matt Caswell
242525372c Remove session checks from SSL_clear()
We now allow a different protocol version when reusing a session so we can
unconditionally reset the SSL_METHOD if it has changed.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/3954)
2017-07-18 17:35:52 +01:00
Matt Caswell
59ff3f07dc Fix SSL_clear() in TLSv1.3
SSL_clear() does not reset the SSL_METHOD if a session already exists in
the SSL object. However, TLSv1.3 does not have an externally visible
version fixed method (only an internal one). The state machine assumes
that we are always starting from a version flexible method for TLSv1.3.
The simplest solution is to just fix SSL_clear() to always reset the method
if it is using the internal TLSv1.3 version fixed method.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/3954)
2017-07-18 17:35:34 +01:00
Matt Caswell
00848ea842 Tolerate a zero length ticket nonce
TLSv1.3 draft-21 requires the ticket nonce to be at least 1 byte in length.
However NSS sends a zero length nonce. This is actually ok because the next
draft will allow zero length nonces anyway, so we should tolerate this.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3957)
2017-07-18 16:59:37 +01:00
Matt Caswell
d4504fe579 Fix early_data with an HRR
early_data is not allowed after an HRR. We failed to handle that
correctly.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3933)
2017-07-18 16:51:58 +01:00
Tatsuhiro Tsujikawa
c8b93876f1 Fix TLSv1.3 exporter secret
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3941)
2017-07-17 13:42:52 +01:00
Dr. Stephen Henson
cd933ebd57 Move certificate table to header file so it can be tested.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3858)
2017-07-13 12:38:42 +01:00
Dr. Stephen Henson
13cc257423 Use cert tables in ssl_set_sig_mask
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3858)
2017-07-13 12:38:42 +01:00
Dr. Stephen Henson
ebefced904 make errors
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3858)
2017-07-13 12:38:42 +01:00
Dr. Stephen Henson
dd24857b78 Use cert tables instead of X509_certificate_type
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3858)
2017-07-13 12:38:42 +01:00
Dr. Stephen Henson
52fd27f978 Use certificate tables instead of ssl_cert_type
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3858)
2017-07-13 12:38:42 +01:00
Dr. Stephen Henson
7f6b466b2c Use certificate tables instead of ssl_cipher_get_cert_index.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3858)
2017-07-13 12:38:42 +01:00
Dr. Stephen Henson
b8858aec13 Replace tls12_get_pkey_idx
The functiontls12_get_pkey_idx is only used to see if a certificate index is
enabled: call ssl_cert_is_disabled instead.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3858)
2017-07-13 12:38:42 +01:00
Dr. Stephen Henson
e4fb8b471c Add SSL_aCERT: this is used for any ciphersuite with a certificate.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3858)
2017-07-13 12:38:42 +01:00
Dr. Stephen Henson
c04cd72827 Add certificate properties table.
Add certificate table giving properties of each certificate index:
specifically the NID associated with the index and the the auth mask
value for any cipher the certificate can be used with.

This will be used to generalise certificate handling instead of hard coding
algorithm specific cases.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3858)
2017-07-13 12:38:42 +01:00
Richard Levitte
a7ff57965b Fix cipher_compare
Unsigned overflow.  Found by Brian Carpenter

Fixes 

Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3890)
2017-07-08 22:13:24 +02:00
Matt Caswell
e3c0d76bc7 Do not allow non-dhe kex_modes by default
Allow that mode to be configured if desired.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3833)
2017-07-07 16:08:05 +01:00
Matt Caswell
5159821540 Updates following review feedback of TLSv1.3 draft-21 code
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3852)
2017-07-07 15:02:09 +01:00
Matt Caswell
8f81476145 Update SSL_trace() to know about ticket_nonce
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3852)
2017-07-07 15:02:09 +01:00
Matt Caswell
b81bd33680 Update the early_secret generation to use the new ticket_nonce field
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3852)
2017-07-07 15:02:09 +01:00
Matt Caswell
a19ae67d8d Update tls13_hkdf_expand() to take the length of the data
In most scenarios the length of the input data is the hashsize, or 0 if
the data is NULL. However with the new ticket_nonce changes the length can
be different.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3852)
2017-07-07 15:02:09 +01:00
Matt Caswell
9b6a825461 Send and receive the ticket_nonce field in a NewSessionTicket
This just adds the processing for sending and receiving the newly added
ticket_nonce field. It doesn't actually use it yet.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3852)
2017-07-07 15:02:09 +01:00
Benjamin Kaduk
c8feba723a Error out when forcing an unsupported TLS version
If the result of a SSL_{CTX_,}set_{min,max}_proto_version() call
leaves the min and max version identical, and support for that version
is compiled out of the library, return an error.  Such an object has
no hope of successfully completing a handshake, and this error may
be easier to decipher than the resulting handshake failure.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3422)
2017-07-05 12:58:34 -05:00
Bernd Edlinger
b43c376586 Fix potential crash in tls_construct_finished.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3667)
2017-07-03 16:34:33 +02:00
Bernd Edlinger
c31ad0bbf8 Fix a crash in tls_construct_client_certificate.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3667)
2017-07-03 16:34:33 +02:00
Matt Caswell
4d89bf4484 Fix TLSv1.3 exporter
We need to use the hashsize in generating the exportsecret not 0! Otherwise
we end up with random garbage for the secret.

It was pure chance that this passed the tests previously. It so happens
that, because we call SSL_export_keying_material() repeatedly for different
scenarios in the test, we end up in the tls13_export_keying_material() at
exactly the same position in the stack each time and therefore end up using
the same random garbage secret each time!

Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3810)
2017-07-03 10:48:34 +01:00
Matt Caswell
d8c66f5e09 Drop support for OPENSSL_NO_TLS1_3_METHOD
There are no public TLSv1_3_*method() functions so
OPENSSL_NO_TLS1_3_METHOD doesn't make any sense and should be removed.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3800)
2017-06-30 09:41:46 +01:00
Matt Caswell
a599574be1 Updates following review of SSL_export_key_material() changes
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3782)
2017-06-29 10:15:50 +01:00
Matt Caswell
0ca8d1ecf2 Update SSL_export_keying_material() for TLSv1.3
Fixes 

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3782)
2017-06-29 10:15:49 +01:00
Todd Short
eed3ec9047 ssl_session_dup() missing ext.alpn_session
Properly copy ext.alpn_session in ssl_session_dup()
Use OPENSSL_strndup() as that's used in ssl_asn1.c

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3770)
2017-06-28 14:53:23 -05:00
Benjamin Kaduk
6ffeb269a3 Disallow DSA/SHA1/etc. for pure TLS 1.3 ClientHellos
In draft-ietf-tls-tls13-20 Appendix B we find that:

   This section describes protocol types and constants.  Values listed
   as _RESERVED were used in previous versions of TLS and are listed
   here for completeness.  TLS 1.3 implementations MUST NOT send them
   but might receive them from older TLS implementations.

Similarly, in section 4.2.3 we see:

   Legacy algorithms  Indicates algorithms which are being deprecated
      because they use algorithms with known weaknesses, specifically
      SHA-1 which is used in this context with either with RSA using
      RSASSA-PKCS1-v1_5 or ECDSA.  These values refer solely to
      signatures which appear in certificates (see Section 4.4.2.2) and
      are not defined for use in signed TLS handshake messages.
      Endpoints SHOULD NOT negotiate these algorithms but are permitted
      to do so solely for backward compatibility.  Clients offering
      these values MUST list them as the lowest priority (listed after
      all other algorithms in SignatureSchemeList).  TLS 1.3 servers
      MUST NOT offer a SHA-1 signed certificate unless no valid
      certificate chain can be produced without it (see
      Section 4.4.2.2).

However, we are currently sending the SHA2-based DSA signature schemes
and many SHA1-based schemes, which is in contradiction with the specification.

Because TLS 1.3 support will appear in OpenSSL 1.1, we are bound by
stability requirements to continue to offer the DSA signature schemes
and the deprecated hash algorithms.  at least until OpenSSL 1.2.
However, for pure TLS 1.3 clients that do not offer lower TLS versions,
we can be compliant.  Do so, and leave a note to revisit the issue when
we are permitted to break with sacred historical tradition.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3326)
2017-06-24 19:25:43 -05:00
Matt Caswell
25ffeb11ea Fix another EVP_DigestVerify() instance
Following on from the previous commit this fixes another instance where
we need to treat a -ve return from EVP_DigestVerify() as a bad signature.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3756)
2017-06-23 17:23:52 +01:00
Matt Caswell
cfba06758e Treat all failures from EVP_DigestVerify() as a bad signature
Prior to 72ceb6a we treated all failures from the call to
EVP_DigestVerifyFinal() as if it were a bad signature, and failures in
EVP_DigestUpdate() as an internal error. After that commit we replaced
this with the one-shot function EVP_DigestVerify() and treated a 0 return
as a bad signature and a negative return as an internal error. However,
some signature errors can be negative (e.g. according to the docs if the
form of the signature is wrong). Therefore we should treat all <=0
returns as a bad signature.

This fixes a boringssl test failure.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3756)
2017-06-23 17:23:25 +01:00
Alexey Komnin
f6373823c0 Fix tls1_generate_master_secret
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3751)
2017-06-22 11:54:19 -04:00
David Benjamin
32bbf777d0 Fix the names of ChaCha20-Poly1305 cipher suites in t1_trce.c.
Per RFC 7905, the cipher suite names end in "_SHA256". The original
implementation targeted the -03 draft, but there was a -04 draft right
before the RFC was published to make the names consistent.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3748)
2017-06-22 16:43:03 +01:00
Paul Yang
15b1688ac9 Avoid compiler complaining
initialize some local variables

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3741)
2017-06-21 14:36:02 -04:00
Matt Caswell
23cec1f4b4 Add documentation for the SSL_export_keying_material() function
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3735)
2017-06-21 16:18:36 +01:00
Matt Caswell
9924087573 Fix DTLS failure when used in a build which has SCTP enabled
The value of BIO_CTRL_DGRAM_SET_PEEK_MODE was clashing with the value for
BIO_CTRL_DGRAM_SCTP_SET_IN_HANDSHAKE. In an SCTP enabled build
BIO_CTRL_DGRAM_SCTP_SET_IN_HANDSHAKE was used unconditionally with
the reasoning that it would be ignored if SCTP wasn't in use. Unfortunately
due to this clash, this wasn't the case. The BIO ended up going into peek
mode and was continually reading the same data over and over - throwing it
away as a replay.

Fixes 

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3724)
2017-06-21 14:56:24 +01:00
Matt Caswell
72257204bd PSK related tweaks based on review feedback
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3670)
2017-06-21 14:45:36 +01:00
Matt Caswell
011d768aba Fix some bugs in the TLSv1.3 PSK code
Found while developing the PSK tests

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3670)
2017-06-21 14:45:36 +01:00
Matt Caswell
725b0f1e13 Make the input parameters for SSL_SESSION_set1_master_key const
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3670)
2017-06-21 14:45:36 +01:00
Matt Caswell
5a43d5119e Add SSL_SESSION_set_protocol_version()
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3670)
2017-06-21 14:45:35 +01:00
Matt Caswell
1a993d1d56 Add SSL_SESSION_set_cipher()
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3670)
2017-06-21 14:45:35 +01:00
Matt Caswell
911d63f261 Add SSL_SESSION_set1_master_key()
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3670)
2017-06-21 14:45:35 +01:00
Matt Caswell
ba4df68286 Add a function to get the handshake digest for an SSL_CIPHER
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3670)
2017-06-21 14:45:35 +01:00
Matt Caswell
f46184bd90 Add public API functions for setting TLSv1.3 PSK callbacks
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3670)
2017-06-21 14:45:35 +01:00
Matt Caswell
9368f86526 Add TLSv1.3 client side external PSK support
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3670)
2017-06-21 14:45:35 +01:00
Matt Caswell
3a7c56b2c0 Add TLSv1.3 server side external PSK support
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3670)
2017-06-21 14:45:35 +01:00
Dr. Stephen Henson
d2916a5b29 Use EVP_PKEY_X25519, EVP_PKEY_ED25519 instead of NIDs where appropriate.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3585)
2017-06-21 14:11:01 +01:00
Dr. Stephen Henson
b2021556e4 Allow Ed25519 in TLS 1.2
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3585)
2017-06-21 14:11:01 +01:00
Dr. Stephen Henson
72ceb6a692 Convert key exchange to one shot call
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3585)
2017-06-21 14:11:01 +01:00
Dr. Stephen Henson
881d2c5ed2 Add Ed25519 to trace output
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3585)
2017-06-21 14:11:01 +01:00
Dr. Stephen Henson
168067b631 Handle signature algorithms with no associated digest
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3585)
2017-06-21 14:11:01 +01:00
Dr. Stephen Henson
3d234c9e5c Add Ed25519 signature algorithm
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3585)
2017-06-21 14:11:01 +01:00
Dr. Stephen Henson
b04d4e38b7 Add Ed25519 to signature algorithm table
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3585)
2017-06-21 14:11:01 +01:00
Dr. Stephen Henson
07afa3d880 Add index for ED25519
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3585)
2017-06-21 14:11:01 +01:00
Dr. Stephen Henson
65e89736b3 Use X509_get_signature_info to get signature strength.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3585)
2017-06-21 14:11:01 +01:00
Rich Salz
c80149d9f0 Merge Nokia copyright notice into standard
This is done with the kind permission of Nokia.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3722)
2017-06-21 08:59:18 -04:00
Rich Salz
aa8f3d76fc Modify Sun copyright to follow OpenSSL style
Approved by Oracle.

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/3585)
2017-06-20 11:13:45 -04:00
Matt Caswell
abeb2a639b Tweak the check that a ciphersuite has not changed since the HRR
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3623)
2017-06-16 10:57:59 +01:00
Matt Caswell
8acc27998b Fix an uninitialised variable warning
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3623)
2017-06-16 10:57:59 +01:00
Matt Caswell
0de6d66d36 Move ciphersuite selection before session resumption in TLSv1.3
This does things as per the recommendation in the TLSv1.3 spec. It also
means that the server will always choose its preferred ciphersuite.
Previously the server would only select ciphersuites compatible with the
session.

Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/3623)
2017-06-16 10:57:59 +01:00
Matt Caswell
a055a88155 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)
2017-06-16 10:57:59 +01:00
Benjamin Kaduk
1316c9ffab 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)
2017-06-15 17:02:45 -04:00
Bernd Edlinger
71d8c138cf 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)
2017-06-12 15:15:04 -04:00
Rich Salz
5816586ba4 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)
2017-06-12 15:03:40 -04:00
Benjamin Kaduk
193b5d769c 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)
2017-06-12 09:31:47 +01:00
Benjamin Kaduk
cf34d54d82 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)
2017-06-12 09:19:05 +01:00
Rich Salz
fbaf2857cc 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)
2017-06-11 13:44:56 -04:00
Pichulin Dmitrii
f464f9c04b 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)
2017-06-09 13:05:22 -04:00
Paul Yang
bd91e3c870 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)
2017-06-09 12:04:10 -04:00
Tomas Mraz
1c7aa0dbf1 Ignore -named_curve auto value to improve backwards compatibility
Fixes 

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3518)
2017-06-08 21:28:36 +01:00
Diego Santa Cruz
135976b3dd 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)
2017-06-08 20:59:24 +01:00
Diego Santa Cruz
e655f54941 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)
2017-06-08 20:59:00 +01:00
Matt Caswell
d8311fc93a 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 

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3639)
2017-06-08 17:37:02 +01:00
Rich Salz
52df25cf2e 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)
2017-06-07 15:12:03 -04:00
Todd Short
db0f35dda1 Fix 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)
2017-06-06 22:39:41 +01:00
Todd Short
270d65fa34 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)
2017-06-06 22:39:41 +01:00
Todd Short
1ee2125922 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)
2017-06-02 12:11:38 -04:00
Todd Short
dffdcc773a Fix inconsistent check of UNSAFE_LEGACY_RENEGOTIATION
The check for SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION is
inconsistent. Most places check SSL->options, one place is checking
SSL_CTX->options; fix that.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
GH: 
2017-05-26 11:29:18 +02:00
Dr. Stephen Henson
caf2b6b54f Don't use one shot API for SSLv3.
SSLv3 (specifically with client auth) cannot use one shot APIs: the digested
data and the master secret are handled in separate update operations. So
in the special case of SSLv3 use the streaming API.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3527)
2017-05-23 15:51:24 +01:00
Matt Caswell
42bd7a16d0 Add an error to the stack on failure in dtls1_write_bytes()
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3496)
2017-05-22 14:00:52 +01:00
Matt Caswell
b77f3ed171 Convert existing usage of assert() to ossl_assert() in libssl
Provides consistent output and approach.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3496)
2017-05-22 14:00:43 +01:00
Matt Caswell
a89325e41f Fix some style issues in returns
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3496)
2017-05-22 14:00:34 +01:00
Matt Caswell
380a522f68 Replace instances of OPENSSL_assert() with soft asserts in libssl
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3496)
2017-05-22 14:00:19 +01:00
Matt Caswell
aabe3a3591 Fix return code in tls1_mac
The return code from tls1_mac is supposed to be a boolean 0 for fail, 1 for
success. In one place we returned -1 on error. This would cause code calling
the mac function to erroneously see this as a success (because a non-zero
value is being treated as success in all call sites).

Fortunately, AFAICT, the place that returns -1 can only happen on an
internal error so is not under attacker control. Additionally this code only
appears in master. In 1.1.0 the return codes are treated differently.
Therefore there are no security implications.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3495)
2017-05-19 17:02:11 +01:00
Matt Caswell
6944311688 Make SSL_is_server() accept a const SSL
Fixes 

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3360)
2017-05-19 08:54:39 +01:00
Matt Caswell
fb34a0f4e0 Try to be more consistent about the alerts we send
We are quite inconsistent about which alerts get sent. Specifically, these
alerts should be used (normally) in the following circumstances:

SSL_AD_DECODE_ERROR = The peer sent a syntactically incorrect message
SSL_AD_ILLEGAL_PARAMETER = The peer sent a message which was syntactically
correct, but a parameter given is invalid for the context
SSL_AD_HANDSHAKE_FAILURE = The peer's messages were syntactically and
semantically correct, but the parameters provided were unacceptable to us
(e.g. because we do not support the requested parameters)
SSL_AD_INTERNAL_ERROR = We messed up (e.g. malloc failure)

The standards themselves aren't always consistent but I think the above
represents the best interpretation.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3480)
2017-05-19 08:47:08 +01:00
Matt Caswell
d8028b202b Fix EXT_RETURN usage for add_key_share()
add_key_share() is a helper function used during key_share extension
construction. It is expected to be a simple boolean success/fail return.
It shouldn't be using the new EXT_RETURN type but it was partially converted
anyway. This changes it back.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3488)
2017-05-18 10:05:02 +01:00
Matt Caswell
108d45df73 Allow a missing sig algs extension if resuming
The current TLSv1.3 spec says:

'If a server is authenticating via a certificate and the client has not
sent a "signature_algorithms" extension, then the server MUST abort the
handshake with a "missing_extension" alert (see Section 8.2).'

If we are resuming then we are not "authenticating via a certificate" but
we were still aborting with the missing_extension alert if sig algs was
missing.

This commit ensures that we only send the alert if we are not resuming.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3478)
2017-05-17 15:33:42 +01:00
Matt Caswell
355a0d10a9 Add a new unsolicited extension error code and add enum tag
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3418)
2017-05-17 15:32:25 +01:00
Matt Caswell
b186a59283 Fail if we receive a response to an extension that we didn't request
We already did this on an ad-hoc per extension basis (for some extensions).
This centralises it and makes sure we do it for all extensions.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3418)
2017-05-17 15:23:49 +01:00
Matt Caswell
bd990e2535 Don't allow fragmented alerts
An alert message is 2 bytes long. In theory it is permissible in SSLv3 -
TLSv1.2 to fragment such alerts across multiple records (some of which
could be empty). In practice it make no sense to send an empty alert
record, or to fragment one. TLSv1.3 prohibts this altogether and other
libraries (BoringSSL, NSS) do not support this at all. Supporting it adds
significant complexity to the record layer, and its removal is unlikely
to cause inter-operability issues.

The DTLS code for this never worked anyway and it is not supported at a
protocol level for DTLS. Similarly fragmented DTLS handshake records only
work at a protocol level where at least the handshake message header
exists within the record. DTLS code existed for trying to handle fragmented
handshake records smaller than this size. This code didn't work either so
has also been removed.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3476)
2017-05-17 10:40:04 +01:00
Bernd Edlinger
018fcbec38 Fix gcc-7 warnings.
- Mostly missing fall thru comments
- And uninitialized value used in sslapitest.c

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3440)
2017-05-11 19:39:38 +02:00
Matt Caswell
9010b7bc6e Add some extra comments following alert changes
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3436)
2017-05-11 13:13:04 +01:00
Matt Caswell
26b9172a50 Add some checks for trailing data after extension blocks
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3436)
2017-05-11 13:13:04 +01:00
Matt Caswell
b6fdc12d94 Send a missing_extension alert if key_share/supported groups not present
Only applies if we're not doing psk.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3436)
2017-05-11 13:13:04 +01:00
Matt Caswell
fc4c15faa4 TLSv1.3 alert and handshake messages can never be 0 length
We abort if we read a message like this.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3436)
2017-05-11 13:13:04 +01:00
Matt Caswell
0b367d7955 TLSv1.3 alerts cannot be fragmented and only one per record
We should be validating that.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3436)
2017-05-11 13:13:04 +01:00
Matt Caswell
3c544acc38 Check that a TLSv1.3 encrypted message has an app data content type
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3436)
2017-05-11 13:13:04 +01:00
Matt Caswell
2d871227fa Send an illegal parameter alert if the update type in a KeyUpdate is wrong
Previously we sent a decode_error alert.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3436)
2017-05-11 13:13:04 +01:00
Matt Caswell
1a281aab73 Ensure we fail with a decode error alert if the server sends and empty Cert
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3436)
2017-05-11 13:13:04 +01:00
Matt Caswell
f69fe73a2a Fix more alert codes
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3436)
2017-05-11 13:13:04 +01:00
Matt Caswell
ef57a475a0 Verify that there is no trailing data after the extensions block
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3436)
2017-05-11 13:13:04 +01:00
Matt Caswell
721586eadf Fix some alert codes
Make sure we are using the correct alert codes as per the spec.

Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3436)
2017-05-11 13:13:04 +01:00
Matt Caswell
f66f8a4491 Reject unknown warning alerts in TLSv1.3
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3436)
2017-05-11 13:13:04 +01:00
Dr. Stephen Henson
7539418981 Add EVP_DigestSign and EVP_DigesVerify
Add "single part" digest sign and verify functions. These sign and verify
a message in one function. This simplifies some operations and it will later
be used as the API for algorithms which do not support the update/final
mechanism (e.g. PureEdDSA).

Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3409)
2017-05-11 12:59:25 +01:00
Matt Caswell
21181889d7 Copy custom extension flags in a call to SSL_set_SSL_CTX()
The function SSL_set_SSL_CTX() can be used to swap the SSL_CTX used for
a connection as part of an SNI callback. One result of this is that the
s->cert structure is replaced. However this structure contains information
about any custom extensions that have been loaded. In particular flags are
set indicating whether a particular extension has been received in the
ClientHello. By replacing the s->cert structure we lose the custom
extension flag values, and it appears as if a client has not sent those
extensions.

SSL_set_SSL_CTX() should copy any flags for custom extensions that appear
in both the old and the new cert structure.

Fixes 

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3425)
2017-05-10 16:49:00 +01:00
Richard Levitte
cf53cbea5b Fix 'no-ec'
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3430)
2017-05-10 17:09:35 +02:00
Matt Caswell
66d4bf6b20 Verify that if we have an HRR then something will change
It is invalid if we receive an HRR but no change will result in
ClientHello2.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3414)
2017-05-09 17:23:58 +01:00
Matt Caswell
7b1ec1cfb7 Fix HRR bug
If an HRR gets sent without a key_share (e.g. cookie only) then the code
fails when it should not.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3414)
2017-05-09 17:23:58 +01:00
Matt Caswell
07d447a6fc Don't do the final key_share checks if we are in an HRR
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3414)
2017-05-09 17:23:58 +01:00
Matt Caswell
1fe3549428 Verify that only NULL compression is sent in TLSv1.3 ClientHello
It is illegal in a TLSv1.3 ClientHello to send anything other than the
NULL compression method. We should send an alert if we find anything else
there. Previously we were ignoring this error.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3410)
2017-05-09 17:02:48 +01:00
Matt Caswell
ad5100bc41 Add support to SSL_trace() for inner content types
When using the -trace option with TLSv1.3 all records appear as "application
data". This adds the ability to see the inner content type too.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3408)
2017-05-08 11:42:37 +01:00
Matt Caswell
eee2750bd3 Remove support for OPENSSL_SSL_TRACE_CRYPTO
This trace option does not appear in Configure as a separate option and is
undocumented. It can be switched on using "-DOPENSSL_SSL_TRACE_CRYPTO",
however this does not compile in master or in any 1.1.0 released version.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3408)
2017-05-08 11:42:37 +01:00
Matt Caswell
12635aa09d Updates to supported_groups following review feedback
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3395)
2017-05-08 11:09:02 +01:00
Matt Caswell
6af8754637 Send the supported_groups extension in EE where applicable
The TLSv1.3 spec says that a server SHOULD send supported_groups in the
EE message if there is a group that it prefers to the one used in the
key_share. Clients MAY act on that. At the moment we don't do anything
with it on the client side, but that may change in the future.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3395)
2017-05-08 11:09:02 +01:00
Matt Caswell
7a4e6a1e50 Updates to serverinfo fix based on review feedback
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3382)
2017-05-04 16:02:08 +01:00
Matt Caswell
bade29da33 Fix SSL_CTX_use_serverinfo_ex() et al to properly handle V1 data
SSL_CTX_use_serverinfo_ex() et al were always processing data as if it was
V2 format, even if it was V1. This bug was masked because, although we had
a test which loaded V1 serverinfo data from a file, the function
SSL_CTX_use_serverinfo_file() transparently converts V1 data to V2 before
calling SSL_CTX_use_serverinfo_ex().

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3382)
2017-05-04 16:02:08 +01:00
Matt Caswell
bb78552ee1 Revert "Fix clang compile time error"
This reverts commit 1608d658af.

This is the wrong fix for this issue. The next commit provides a better
fix.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3382)
2017-05-04 16:02:08 +01:00
Todd Short
1608d658af Fix clang compile time error
|version| "could" be used uninitialized here, not really, but the
compiler doesn't understand the flow

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3373)
2017-05-04 05:37:30 +02:00
Matt Caswell
17aa119e7b Update the HKDF labels for draft-20
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3371)
2017-05-03 17:23:02 +01:00
Todd Short
96c9aee2a8 Limit padded record to max plaintext
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3374)
2017-05-03 17:12:00 +01:00
Matt Caswell
f233a9d181 Clarify serverinfo usage with Certificate messages
Ensure that serverinfo only gets added for the first Certificate in a list.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3298)
2017-05-03 14:37:42 +01:00
Matt Caswell
2698bbfec6 Add an SSL_ prefix to SERVERINFOV2 and SERVERINFOV1
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3298)
2017-05-03 14:37:42 +01:00
Matt Caswell
7f533d6fa1 Only send custom extensions where we have received one in the ClientHello
We already did this for ServerHello and EncryptedExtensions. We should be
doing it for Certificate and HelloRetryRequest as well.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3298)
2017-05-03 14:37:42 +01:00
Matt Caswell
84c34ba876 Extend the SERVERINFO file format to include an extensions context
This enables us to know what messages the extensions are relevant for in
TLSv1.3. The new file format is not compatible with the previous one so
we call it SERVERINFOV2.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3298)
2017-05-03 14:37:42 +01:00
Matt Caswell
33564cb749 Fix comment around safari fingerprint check
Fixes 

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3367)
2017-05-02 17:08:41 +01:00
Todd Short
c649d10d3f TLS1.3 Padding
Add padding callback for application control
Standard block_size callback
Documentation and tests included
Configuration file/s_client/s_srver option

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3130)
2017-05-02 09:44:43 +01:00
Graham Edgecombe
e596c68c31 Return success in custom_ext_parse_old_cb_wrap if parse_cb is NULL
This fixes a segfault if a NULL parse_cb is passed to
SSL_CTX_add_{client,server}_custom_ext, which was supported in the
pre-1.1.1 implementation.

This behaviour is consistent with the other custom_ext_*_old_cb_wrap
functions, and with the new SSL_CTX_add_custom_ext function.

CLA: trivial

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3310)
2017-04-26 17:37:30 +01:00
Tatsuhiro Tsujikawa
8e1634ec36 Don't treat PACKET_remaining() as boolean
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3244)
2017-04-26 16:56:37 +01:00
Tatsuhiro Tsujikawa
bf5c84f5d1 Break before && operator
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3244)
2017-04-26 16:56:36 +01:00
Tatsuhiro Tsujikawa
735d5b59df Call init and finalization functions per extension message
Previously, init and finalization function for extensions are called
per extension block, rather than per message.  This commit changes
that behaviour, and now they are called per message.  The parse
function is still called per extension block.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3244)
2017-04-26 16:56:35 +01:00
Matt Caswell
150840b944 Always duplicate the session on NewSessionTicket in TLSv1.3
Because NST messages arrive post-handshake, the session may have already
gone into the cache. Once in the cache a session must be immutable -
otherwise you could get multi-thread issues.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3008)
2017-04-26 16:42:29 +01:00
Matt Caswell
e586eac885 Add support for SSL_SESSION_is_resumable()
Provide a way to test whether the SSL_SESSION object can be used to resume a
sesion or not.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3008)
2017-04-26 16:42:29 +01:00
Matt Caswell
3348fc7e89 Remove TLS1.3 TODO around testing for session id length
TLSv1.3 will do the same thing as TLSv1.2 with tickets with regards to session
ids, i.e. it will create a synthetic session id when the session is established,
so it is reasonable to check the session id length, even in TLSv1.3.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3008)
2017-04-26 16:42:29 +01:00
Matt Caswell
38a7315060 Add a ciphersuite config sanity check for servers
Ensure that there are ciphersuites enabled for the maximum supported
version we will accept in a ClientHello.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3316)
2017-04-26 14:31:00 +01:00
Matt Caswell
aafec89c63 Add a ciphersuite config sanity check for clients
Ensure that there are ciphersuites enabled for the maximum supported
version we are claiming in the ClientHello.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3316)
2017-04-26 14:31:00 +01:00
Matt Caswell
bf846a6d47 Don't overwrite the alert value if there is no alert to send
The function tls_early_post_process_client_hello() was overwriting the
passed "al" parameter even if it was successful. The caller of that
function, tls_post_process_client_hello(), sets "al" to a sensible default
(HANDSHAKE_FAILURE), but this was being overwritten to be INTERNAL_ERROR.
The result is a "no shared cipher" error (and probably other similar errors)
were being reported back to the client with an incorrect INTERNAL_ERROR
alert.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3314)
2017-04-26 14:13:26 +01:00
Matt Caswell
aefb925647 Don't attempt to send fragments > max_send_fragment in DTLS
We were allocating the write buffer based on the size of max_send_fragment,
but ignoring it when writing data. We should fragment handshake messages
if they exceed max_send_fragment and reject application data writes that
are too large.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)
2017-04-25 11:13:39 +01:00
Matt Caswell
bd79bcb42b Remove special case code for SCTP reneg handling
There was code existing which attempted to handle the case where application
data is received after a reneg handshake has started in SCTP. In normal DTLS
we just fail the connection if this occurs, so there doesn't seem any reason
to try and work around it for SCTP. In practice it didn't work properly
anyway and is probably a bad idea to start with.

Fixes 

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3286)
2017-04-25 11:13:39 +01:00
Matt Caswell
8af91fd9d0 Don't fail the connection in SSLv3 if server selects ECDHE
ECDHE is not properly defined for SSLv3. Commit fe55c4a2 prevented ECDHE
from being selected in that protocol. However, historically, servers do
still select ECDHE anyway so that commit causes interoperability problems.
Clients that previously worked when talking to an SSLv3 server could now
fail.

This commit introduces an exception which enables a client to continue in
SSLv3 if the server selected ECDHE.

[extended tests]

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3204)
2017-04-24 16:15:40 +01:00
Rich Salz
26fb4b033c if log fails, avoid pms getting double free
doing the pms assignment after log is successful

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3278)
2017-04-24 09:41:51 -04:00
Rich Salz
c0452248ea Ignore dups in X509_STORE_add_*
X509_STORE_add_cert and X509_STORE_add_crl are changed to return
success if the object to be added was already found in the store, rather
than returning an error.

Raise errors if empty or malformed files are read when loading certificates
and CRLs.

Remove NULL checks and allow a segv to occur.
Add error handing for all calls to X509_STORE_add_c{ert|tl}

Refactor these two routines into one.

Bring the unit test for duplicate certificates up to date using the test
framework.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2830)
2017-04-20 15:33:42 -04:00
Todd Short
d1186c30a2 Fix minor compiler issues.
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3220)
2017-04-19 12:51:08 -04:00
Richard Levitte
9612e15760 ASN.1: adapt our use of INTxx et al by making them explicitely embedded
Fixes 

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3199)
2017-04-13 10:23:31 +02:00
Rob Percival
6a71e06d7a CT_POLICY_EVAL_CTX_set_time expects milliseconds, but given seconds
This resulted in the SCT timestamp check always failing, because the
timestamp appeared to be in the future.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3138)
2017-04-12 19:08:57 +02:00
Todd Short
fe55c4a20f Remove ECDH(E) ciphers from SSLv3
SSLv3 does not support TLS extensions, and thus, cannot provide any
curves for ECDH(E). With the removal of the default (all) list of curves
being used for connections that didn't provide any curves, ECDHE is no
longer possible.

Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3181)
2017-04-11 13:25:19 -04:00
Rich Salz
cbbe9186f3 Additional check to handle BAD SSL_write retry
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3124)
2017-04-11 11:59:59 -04:00
Benjamin Kaduk
8313a787d7 Allow an ALPN callback to pretend to not exist
RFC 7301 mandates that the server SHALL respond with a fatal
"no_application_protocol" alert when there is no overlap between
the client's supplied list and the server's list of supported protocols.
In commit 062178678f we changed from
ignoring non-success returns from the supplied alpn_select_cb() to
treating such non-success returns as indicative of non-overlap and
sending the fatal alert.

In effect, this is using the presence of an alpn_select_cb() as a proxy
to attempt to determine whether the application has configured a list
of supported protocols.  However, there may be cases in which an
application's architecture leads it to supply an alpn_select_cb() but
have that callback be configured to take no action on connections that
do not have ALPN configured; returning SSL_TLSEXT_ERR_NOACK from
the callback would be the natural way to do so.  Unfortunately, the
aforementioned behavior change also treated SSL_TLSEXT_ERR_NOACK as
indicative of no overlap and terminated the connection; this change
supplies special handling for SSL_TLSEXT_ERR_NOACK returns from the
callback.  In effect, it provides a way for a callback to obtain the
behavior that would have occurred if no callback was registered at
all, which was not possible prior to this change.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2570)
2017-04-10 11:57:37 -04:00
Richard Levitte
64f11ee888 Publish our INT32, UINT32, INT64, UINT64 ASN.1 types and Z variants
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3126)
2017-04-10 12:10:59 +02:00
Matt Caswell
0f5af6b1b8 Remove an out of date TODO
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3139)
2017-04-07 13:41:04 +01:00
Matt Caswell
787d9ec74c Create an ENDPOINT enum type for use internally
We need it for the custom extensions API

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3139)
2017-04-07 13:41:04 +01:00
Matt Caswell
cd17bb190c Prefix custom extension API callback types with SSL_
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3139)
2017-04-07 13:41:04 +01:00
Matt Caswell
64350ab587 Various style tweaks based on feedback
Style updates for the new custom extensions API

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3139)
2017-04-07 13:41:04 +01:00
Matt Caswell
43ae5eed6f Implement a new custom extensions API
The old custom extensions API was not TLSv1.3 aware. Extensions are used
extensively in TLSv1.3 and they can appear in many different types of
messages. Therefore we need a new API to be able to cope with that.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3139)
2017-04-07 13:41:04 +01:00
Matt Caswell
fe874d27d3 Move the extensions context codes into the public API
This move prepares for the later addition of the new custom extensions
API. The context codes have an additional "SSL_" added to their name to
ensure we don't have name clashes with other applications.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3139)
2017-04-07 13:41:04 +01:00
Matt Caswell
b443c84594 Move ssl/t1_ext.c to ssl/statem/extensions_cust.c
Brings all the extensions code together.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3139)
2017-04-07 13:41:04 +01:00
Richard Levitte
66ecfb5437 Convert SSL_SESSION_ASN1 to use size specific integers
This increases portability of SSL_SESSION files between architectures
where the size of |long| may vary.  Before this, SSL_SESSION files
produced on a 64-bit long architecture may break on a 32-bit long
architecture.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3088)
2017-04-04 11:29:23 +02:00
Dr. Stephen Henson
86135bedd5 Constify SSL_dup_CA_list()
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3015)
2017-04-03 23:47:22 +01:00
Dr. Stephen Henson
9784ec0474 Don't use client specific functions to retrieve CA list
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3015)
2017-04-03 23:47:21 +01:00
Dr. Stephen Henson
be885d5075 SSL_CONF support for certificate_authorities
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3015)
2017-04-03 23:47:21 +01:00
Dr. Stephen Henson
fa7c263747 New certificate_authorities functions
Add functions to add/retrieve the certificate_authorities. The older
client_CA functions mainly just call the new versions now.

Rename fields sice new extension can be generated by client and server.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3015)
2017-04-03 23:47:21 +01:00
Matt Caswell
090c8118e8 Fix calls to SSL_get_server_tmp_key() in TLSv1.3
The macro SSL_get_server_tmp_key() returns information about the temp key
used by the server during a handshake. This was returning NULL for TLSv1.3
and causing s_client to omit this information in its connection summary.

Fixes 

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3114)
2017-04-03 19:18:47 +01:00
Matt Caswell
a0cb628b17 Tweak a style issue
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3091)
2017-04-03 19:07:29 +01:00
Tatsuhiro Tsujikawa
bbea9f2c5f Restore s->early_data_state with the original value
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3091)
2017-04-03 19:07:29 +01:00
Matt Caswell
59cebcf9f6 Don't handle handshake messages when writing early data on server
If we have received the EoED message but not yet had the CF then we are
"in init". Despite that we still want to write application data, so suppress
the "in init" check in ssl3_write_bytes() in that scenario.

Fixes 

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3091)
2017-04-03 19:07:29 +01:00
Kazuki Yamaguchi
a8e75d5680 Fix a typo in the SSL_get_max_early_data() declarations
SSL_get_max_early_data() recently added by 3fc8d85610 ("Construct the
ticket_early_data_info extension", 2017-02-17) is supposed to take an
SSL, but it doesn't.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3113)
2017-04-03 13:45:39 +01:00
Matt Caswell
36abb6a270 Update SSL_extension_supported()
Numerous changes have been made to the supported built-in extensions and
SSL_extension_supported() has not kept up.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3097)
2017-03-31 11:14:13 +01:00
Matt Caswell
116d0da5e8 Fix early data bug with pause between EoED and CF
If the server received EoED then SSL_read_early_data() will return
SSL_READ_EARLY_DATA_FINISH. However if the CF has not yet been processed
then SSL_is_init_finished() will still return 0. Therefore we should still
be able to write early data.

Fixes 

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3089)
2017-03-30 15:45:45 +01:00
Matt Caswell
39ef78210c Fix bug with SSL_read_early_data()
If read_ahead is set, or SSL_MODE_AUTO_RETRY is used then if
SSL_read_early_data() hits an EndOfEarlyData message then it will
immediately retry automatically, but this time read normal data instead
of early data!

Fixes 

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3077)
2017-03-30 09:09:21 +01:00
Qin Long
dc8da7b1b8 Fix the build warning under OPENSSL_NO_GOST
Variable 'pktype' was set but not used under OPENSSL_NO_GOST. This change
will fix the build warning under [-Werror=unused-but-set-variable].

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2961)
2017-03-29 07:37:02 +02:00
Benjamin Kaduk
0acee5045a Further de-obfuscation
A similar change that probably should have been wrapped into
commit e0926ef49d.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3010)
2017-03-29 07:33:39 +02:00
FdaSilvaYY
69687aa829 More typo fixes
Fix some comments too
[skip ci]

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3069)
2017-03-29 07:14:29 +02:00
FdaSilvaYY
1ee4b98e69 Fix a few more typos
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3066)
2017-03-28 16:16:49 -04:00
Matt Caswell
643a358042 Move the downgrade sentinel declarations to a header file
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3022)
2017-03-24 14:07:11 +00:00
Matt Caswell
c3043dcd55 Add client side support for TLSv1.3 downgrade mechanism
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3022)
2017-03-24 14:07:11 +00:00
Matt Caswell
f7f2a01d63 Add server side support for TLSv1.3 downgrade mechanism
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3022)
2017-03-24 14:07:11 +00:00
Matt Caswell
77815a026c Fix resumption after HRR
Commit 6b1bb98fa moved the processing of ClientHello extensions into the
state machine post-processing stage. After processing s->init_num is reset
to 0, so by post-processing we cannot rely on its value. Unfortunately we
were using it to handle the PSK extension. This causes the handshake to
fail.

We were using init_num to figure out the length of ClientHello2 so we can
remove it from the handshake_buffer. The handshake_buffer holds the
transcript of all the messages sent so far. For PSK processing though we
only want to add in a partial ClientHello2. This commit changes things so
we just work out where ClientHello2 starts, working forward from the
beginning of handshake_buffer.

Fixes 

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2996)
2017-03-21 10:00:44 +00:00
Joseph Birr-Pixton
b0c9fc9b48 TLS1.3: Correct intermediate secret derivation
This label for this derivation was incorrectly "derived" or "der" depending
on the pointer size of the build(!). The correct string is "derived secret".

(cherry picked from commit 936dcf272033c1bf59a5e859ec63e2557194f191)

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2989)
2017-03-19 20:59:59 +00:00
Dr. Stephen Henson
45615c5fac Implement certificate_authorities extension
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2918)
2017-03-17 18:41:56 +00:00
Dr. Stephen Henson
32f661079d Support draft-19 TLS certificate request format
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2918)
2017-03-17 18:41:56 +00:00
Dr. Stephen Henson
51c7d3e824 Allow signature algorithms in TLS 1.3 certificate request extensions.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2918)
2017-03-17 18:41:56 +00:00
Dr. Stephen Henson
5d6cca05b0 Move parsing and construction of CA names to separate functions
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2918)
2017-03-17 18:41:56 +00:00
Dr. Stephen Henson
f8f16d8ea4 Remove obsolete version test when returning CA names.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2969)
2017-03-16 18:07:19 +00:00
Todd Short
07518cfb38 Add TLSv1.3 draft-19 messages to trace
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2970)
2017-03-16 13:16:51 -04:00