Commit Graph

3078 Commits

Author SHA1 Message Date
David Benjamin
cec7344bba Add a CBS version of SSL_early_callback_ctx_extension_get.
Save a little bit of typing at the call site.

Change-Id: I818535409b57a694e5e0ea0e9741d89f2be89375
Reviewed-on: https://boringssl-review.googlesource.com/9090
Reviewed-by: Adam Langley <agl@google.com>
2016-08-03 20:47:05 +00:00
Steven Valdez
1e6f11a7ff Adding NewSessionTicket.
We will now send tickets as a server and accept them as a
client. Correctly offering and resuming them in the handshake will be
implemented in a follow-up.

Now that we're actually processing draft 14 tickets, bump the draft
version.

Change-Id: I304320a29c4ffe564fa9c00642a4ace96ff8d871
Reviewed-on: https://boringssl-review.googlesource.com/8982
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-03 20:03:20 +00:00
David Benjamin
e8e84b9008 Reject warning alerts in TLS 1.3.
As of https://github.com/tlswg/tls13-spec/pull/530, they're gone.
They're still allowed just before the ClientHello or ServerHello, which
is kind of odd, but so it goes.

BUG=86

Change-Id: I3d556ab45e42d0755d23566e006c0db9af35b7b6
Reviewed-on: https://boringssl-review.googlesource.com/9114
Reviewed-by: Steven Valdez <svaldez@google.com>
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-03 19:58:01 +00:00
Steven Valdez
7259f2fd08 Prefix ext_key_share methods.
Change-Id: Id6a7443246479c62cbe0024e2131a2013959e21e
Reviewed-on: https://boringssl-review.googlesource.com/9078
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 23:13:29 +00:00
Steven Valdez
7b689f6b9e Using NewSessionCallback for bssl client.
TLS 1.3 requires callers use the callback rather than SSL_get_session.

Change-Id: I2caae70e641b102ce93256c847c178871bf78bac
Reviewed-on: https://boringssl-review.googlesource.com/9076
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 23:04:11 +00:00
David Benjamin
a70de147ff Check for trailing data in key_share extension.
Change-Id: I057e19a9547a14b3950395db4318eaf0da01ec13
Reviewed-on: https://boringssl-review.googlesource.com/9079
Reviewed-by: Steven Valdez <svaldez@google.com>
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 21:37:39 +00:00
David Benjamin
ce079fda12 Add SSL_is_dtls.
OpenSSL 1.1.0 added a function to tell if an SSL* is DTLS or not. This
is probably a good idea, especially since SSL_version returns
non-normalized versions.

BUG=91

Change-Id: I25c6cf08b2ebabf0c610c74691de103399f729bc
Reviewed-on: https://boringssl-review.googlesource.com/9077
Reviewed-by: Steven Valdez <svaldez@google.com>
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 20:43:58 +00:00
Brian Smith
dc7a786d31 Use BN_nnmod instead of BN_mod in BN_mod_exp_mont_consttime.
|BN_mod_exp_mont| uses |BN_nnmod| so it seems like
|BN_mod_exp_mont_consttime| should too. Further, I created
these test vectors by doing the math by hand, and the tests
passed for |BN_mod_exp_mont| but failed for
|BN_mod_exp_mont_consttime| without this change.

Change-Id: I7cffa1375e94dd8eaee87ada78285cd67fff1bac
Reviewed-on: https://boringssl-review.googlesource.com/9032
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 20:24:58 +00:00
David Benjamin
da2630c190 Remove redundant SSL_VERIFY_PEER check.
None of the SSL_VERIFY_FAIL_IF_NO_PEER_CERT codepaths will ever be
reached if SSL_VERIFY_PEER is unset. If we've gotten as far as getting a
Certificate message, consider SSL_VERIFY_FAIL_IF_NO_PEER_CERT alone
significant grounds for rejecting no peer certificate.

Change-Id: I2c6be4269d65b2467b86b1fc7d76ac47ca735553
Reviewed-on: https://boringssl-review.googlesource.com/9070
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 20:09:04 +00:00
Nick Harper
0b3625bcfd Add support for TLS 1.3 PSK resumption in Go.
Change-Id: I998f69269cdf813da19ccccc208b476f3501c8c4
Reviewed-on: https://boringssl-review.googlesource.com/8991
Reviewed-by: Steven Valdez <svaldez@google.com>
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 19:37:07 +00:00
David Benjamin
afc64dec74 Add tests to ensure our ClientHello does not change.
We'll need to update it on occasion, but we should not update our
default ClientHello without noticing.

Change-Id: I19ca52fdbe10e3ac14413fecd16be2e58af5a1f6
Reviewed-on: https://boringssl-review.googlesource.com/9075
Reviewed-by: Steven Valdez <svaldez@google.com>
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 19:34:18 +00:00
David Benjamin
3ce4389e96 Move some client/server special-cases out of tls13_process_certificate.
Where we can move uncommon logic to the caller, we probably ought to.

Change-Id: I54a09fffffc20290be05295137ccb605d562cad0
Reviewed-on: https://boringssl-review.googlesource.com/9069
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 19:20:24 +00:00
Brian Smith
78f84f4e03 Document a conservative input range for Montgomery math functions.
The functions appear to try to handle negative inputs, but it isn't
clear how negative inputs are supposed to work and/or if these
functions work the way they are supposed to given negative inputs.
There seems to be no legitimate reason to pass these functions negative
inputs, so just document that negative inputs shouldn't be used. More
specifically, document that the inputs should be in the range [0, n)
where |n| is the Montgomery modulus.

Change-Id: Id8732fb89616f10e673704e6fa09d78926c402d8
Reviewed-on: https://boringssl-review.googlesource.com/9033
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 19:19:35 +00:00
David Benjamin
899b9b19a4 Ensure |BN_div| never gives negative zero in the no_branch code.
Have |bn_correct_top| fix |bn->neg| if the input is zero so that we
don't have negative zeros lying around.

Thanks to Brian Smith for noticing.

Change-Id: I91bcadebc8e353bb29c81c4367e85853886c8e4e
Reviewed-on: https://boringssl-review.googlesource.com/9074
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 18:53:45 +00:00
Eric Roman
875bf04237 Update comments for HMAC to give a more accurate bound than EVP_MD_MAX_SIZE
BUG=59

Change-Id: If3a788ec1328226d69293996845fa1d14690bf40
Reviewed-on: https://boringssl-review.googlesource.com/9068
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 18:20:42 +00:00
David Benjamin
4501bd5118 Align with OpenSSL on SSL_set_bio behavior.
SSL_set_bio is a nightmare.

In f715c42322, we noticed that, among
other problems, SSL_set_bio's actual behavior did not match how
SSL_set_rfd was calling it due to an asymmetry in the rbio/wbio
handling. This resulted in SSL_set_fd/SSL_set_rfd calls to crash.  We
decided that SSL_set_rfd's believed semantics were definitive and
changed SSL_set_bio.

Upstream, in 65e2d672548e7c4bcb28f1c5c835362830b1745b, decided that
SSL_set_bio's behavior, asymmetry and all, was definitive and that the
SSL_set_rfd crash was a bug in SSL_set_rfd. Accordingly, they switched
the fd callers to use the side-specific setters, new in 1.1.0.

Align with upstream's behavior and add tests for all of SSL_set_bio's
insanity. Also export the new side-specific setters in anticipation of
wanting to be mostly compatible with OpenSSL 1.1.0.

Change-Id: Iceac9508711f79750a3cc2ded081b2bb2cbf54d8
Reviewed-on: https://boringssl-review.googlesource.com/9064
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 17:50:39 +00:00
David Benjamin
e76cdde77d Use newest CRL.
If two CRLs are equivalent then use the one with a later lastUpdate field:
this will result in the newest CRL available being used.

(Imported from upstream's 325da8231c8d441e6bb7f15d1a5a23ff63c842e5 and
3dc160e9be6dcaeec9345fbb61b1c427d7026103.)

Change-Id: I8c722663b979dfe08728d091697d8b8204dc265c
Reviewed-on: https://boringssl-review.googlesource.com/8947
Commit-Queue: David Benjamin <davidben@google.com>
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 17:45:15 +00:00
David Benjamin
2b314fa3a9 Tolerate -0 better in BN_bn2{dec,hex}
Negative zeros are nuts, but it will probably be a while before we've
fixed everything that can create them. Fix both to consistently print
'-0' rather than '0' so failures are easier to diagnose (BN_cmp believes
the values are different.)

Change-Id: Ic38d90601b43f66219d8f44ca085432106cf98e3
Reviewed-on: https://boringssl-review.googlesource.com/9073
Commit-Queue: David Benjamin <davidben@google.com>
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 17:35:16 +00:00
Brian Smith
7fcbfdbdf3 Calculate inverse in |BN_MONT_CTX_set| in constant time w.r.t. modulus.
Simplify the calculation of the Montgomery constants in
|BN_MONT_CTX_set|, making the inversion constant-time. It should also
be faster by avoiding any use of the |BIGNUM| API in favor of using
only 64-bit arithmetic.

Now it's obvious how it works. /s

Change-Id: I59a1e1c3631f426fbeabd0c752e0de44bcb5fd75
Reviewed-on: https://boringssl-review.googlesource.com/9031
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-02 16:26:44 +00:00
David Benjamin
0375127606 Promise more accurate bounds than EVP_MD_MAX_SIZE.
A caller using EVP_Digest* which a priori knows tighter bounds on the
hash function used (perhaps because it is always a particular hash) can
assume the function will not write more bytes than the size of the hash.

The letter of the rules before vaguely[*] allowed for more than
EVP_MD_MAX_SIZE bytes written which made for some unreasonable code in
Chromium. Officially clarify this and add tests which, when paired with
valgrind and ASan prove it.

BUG=59

[*] Not really. I think it already promised the output length will be
both the number of bytes written and the size of the hash and the size
of the hash is given by what the function promises to compute. Meh.

Change-Id: I736d526e81cca30475c90897bca896293ff30278
Reviewed-on: https://boringssl-review.googlesource.com/9066
Reviewed-by: Eric Roman <ericroman@google.com>
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-01 23:24:04 +00:00
Adam Langley
d4aae0f965 Minor typo fixes.
Change-Id: Idf9db184348140972e57b2a8fa30dc9cb8b2e0f2
Reviewed-on: https://boringssl-review.googlesource.com/9065
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-01 19:49:06 +00:00
David Benjamin
4890165509 Empty signature algorithms in TLS 1.3 CertificateRequest is illegal.
In TLS 1.2, this was allowed to be empty for the weird SHA-1 fallback
logic. In TLS 1.3, not only is the fallback logic gone, but omitting
them is a syntactic error.

   struct {
       opaque certificate_request_context<0..2^8-1>;
       SignatureScheme
         supported_signature_algorithms<2..2^16-2>;
       DistinguishedName certificate_authorities<0..2^16-1>;
       CertificateExtension certificate_extensions<0..2^16-1>;
   } CertificateRequest;

Thanks to Eric Rescorla for pointing this out.

Change-Id: I4991e59bc4647bb665aaf920ed4836191cea3a5a
Reviewed-on: https://boringssl-review.googlesource.com/9062
Reviewed-by: Steven Valdez <svaldez@google.com>
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-01 19:47:26 +00:00
David Benjamin
0c40a96455 Send unsupported_extension on unexpected ServerHello extensions.
We were sending decode_error, but the spec explicitly says (RFC 5246):

   unsupported_extension
      sent by clients that receive an extended server hello containing
      an extension that they did not put in the corresponding client
      hello.  This message is always fatal.

Also add a test for this when it's a known but unoffered extension. We
actually end up putting these in different codepaths now due to the
custom extensions stuff.

Thanks to Eric Rescorla for pointing this out.

Change-Id: If6c8033d4cfe69ef8af5678b873b25e0dbadfc4f
Reviewed-on: https://boringssl-review.googlesource.com/9061
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-01 18:56:31 +00:00
David Benjamin
636ff1cb7e Convert rsa_1024_key.pem to a PKCS#8 PEM blob.
I missed one.

Change-Id: I311776efd1b2e5da7dca4c88b59a4a4c3e7df94b
Reviewed-on: https://boringssl-review.googlesource.com/9042
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-01 18:42:17 +00:00
David Benjamin
c5aa8414da Fix up header file handling.
As of a recent change, test_support always included the headers, which
causes Android's new build-system to be unhappy. It doesn't want to
include headers. Split them into test_support_headers and test_support
to match the other keys.

Then fix up references:

- Android's new build system only wants the sources. Fix this.

- Chromium's GN and GYP theoretically want the sources and headers, but
  we've never supplied the headers because this isn't enforced at all.
  Fix this. Headers are selected based on what target the header
  "belongs to".

- Bazel has no change except to sort test_support_sources.

Change-Id: I85809e70a71236b5e91d87f87bb73bc2ea289251
Reviewed-on: https://boringssl-review.googlesource.com/9044
Reviewed-by: Adam Langley <agl@google.com>
2016-08-01 18:38:22 +00:00
Adam Langley
9498e74a92 Don't have the default value of |verify_result| be X509_V_OK.
It seems much safer for the default value of |verify_result| to be an
error value.

Change-Id: I372ec19c41d77516ed12d0169969994f7d23ed70
Reviewed-on: https://boringssl-review.googlesource.com/9063
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-01 18:11:39 +00:00
David Benjamin
0d1b0961f9 Fix mixed comment markers.
We managed to mix two comment styles in the Go license headers and
copy-and-paste it throughout the project.

Change-Id: Iec1611002a795368b478e1cae0b53127782210b1
Reviewed-on: https://boringssl-review.googlesource.com/9060
Reviewed-by: Steven Valdez <svaldez@google.com>
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-08-01 14:52:39 +00:00
Steven Valdez
1dc53d2840 Adding handling for KeyUpdate post-handshake message.
BUG=74

Change-Id: I72d52c1fbc3413e940dddbc0b20c7f22459da693
Reviewed-on: https://boringssl-review.googlesource.com/8981
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-29 23:06:09 +00:00
Steven Valdez
8e1c7be1a7 Adding Post-Handshake message handling.
Change-Id: I5cc194fc0a3ba8283049078e5671c924ee23036c
Reviewed-on: https://boringssl-review.googlesource.com/8980
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-29 22:34:57 +00:00
Steven Valdez
87eab4902d Splitting SSL session state.
To prevent configuration/established session confusion, the handshake
session state is separated into the configured session (ssl->session)
and the newly created session (ssl->s3->new_session). Upon conclusion of
the handshake, the finalized session is stored
in (ssl->s3->established_session). During the handshake, any requests
for the session (SSL_get_session) return a non-resumable session, to
prevent resumption of a partially filled session. Sessions should only
be cached upon the completion of the full handshake, using the resulting
established_session. The semantics of accessors on the session are
maintained mid-renego.

Change-Id: I4358aecb71fce4fe14a6746c5af1416a69935078
Reviewed-on: https://boringssl-review.googlesource.com/8612
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-29 21:22:46 +00:00
David Benjamin
163f29af07 Move post-handshake message handling out of read_app_data.
This finishes getting rid of ssl_read_bytes! Now we have separate
entry-points for the various cases. For now, I've kept TLS handshake
consuming records partially. When we do the BIO-less API, I expect that
will need to change, since we won't have the record buffer available.

(Instead, the ssl3_read_handshake_bytes and extend_handshake_buffer pair
will look more like the DTLS side or Go and pull the entire record into
init_buf.)

This change opts to make read_app_data drive the message to completion
in anticipation of DTLS 1.3. That hasn't been specified, but
NewSessionTicket certainly will exist. Knowing that DTLS necessarily has
interleave seems something better suited for the SSL_PROTOCOL_METHOD
internals to drive.

It needs refining, but SSL_PROTOCOL_METHOD is now actually a half-decent
abstraction boundary between the higher-level protocol logic and
DTLS/TLS-specific record-layer and message dispatchy bits.

BUG=83

Change-Id: I9b4626bb8a29d9cb30174d9e6912bb420ed45aff
Reviewed-on: https://boringssl-review.googlesource.com/9001
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-29 21:05:49 +00:00
David Benjamin
e97fb48fbe Test that V2ClientHello must be the first record.
Regression tests for upstream's
https://github.com/openssl/openssl/issues/1298.

Also, given that we're now on our third generation of V2ClientHello
handling, I'm sure we'll have a fourth and fifth and one of these days
I'm going to mess this one up. :-)

Change-Id: I6fd8f311ed0939fbbfd370448b637ccc06145021
Reviewed-on: https://boringssl-review.googlesource.com/9040
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-29 19:39:31 +00:00
Brian Smith
ec3cb3adbc Add |BN_mod_inverse_blinded| and use it in RSA blinding.
Yo dawg I herd you like blinding so I put inversion blinding in your
RSA blinding so you can randomly mask your random mask.

This improves upon the current situation where we pretend that
|BN_mod_inverse_no_branch| is constant-time, and it avoids the need to
exert a lot of effort to make a actually-constant-time modular
inversion function just for RSA blinding.

Note that if the random number generator weren't working correctly then
the blinding of the inversion wouldn't be very effective, but in that
case the RSA blinding itself would probably be completely busted, so
we're not really losing anything by relying on blinding to blind the
blinding.

Change-Id: I771100f0ad8ed3c24e80dd859ec22463ef2a194f
Reviewed-on: https://boringssl-review.googlesource.com/8923
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-29 18:30:34 +00:00
EKR
173bf93827 Accept the special token 'UNTRANSLATED_ERROR' instead of the expected error code when -loose-errors argument is used. Usable for non-bssl shims
Change-Id: I7e85a2677fe28a22103a975d517bbee900c44ac3
Reviewed-on: https://boringssl-review.googlesource.com/9050
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-29 17:00:24 +00:00
David Benjamin
ccd511e499 Add a test for BN_cmp_word.
This also adds a missing OPENSSL_EXPORT.

Change-Id: I6c2400246280f68f51157e959438644976b1171b
Reviewed-on: https://boringssl-review.googlesource.com/9041
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-29 16:39:04 +00:00
Brian Smith
4edca0b308 Add BN_rand_range_ex and use internally.
There are many cases where we need |BN_rand_range| but with a minimum
value other than 0. |BN_rand_range_ex| provides that.

Change-Id: I564326c9206bf4e20a37414bdbce16a951c148ce
Reviewed-on: https://boringssl-review.googlesource.com/8921
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-29 16:09:26 +00:00
David Benjamin
4792110b2b Forbid interleaving app data in a HelloRequest.
We already forbid renego/app-data interleave. Forbid it within a
HelloRequest too because that's nonsense. No one would ever send:

   [hs:HelloReq-] [app:Hello world] [hs:-uest]

Add tests for this case.

This is in preparation for our more complex TLS 1.3 post-handshake logic
which is going to go through the usual handshake reassembly logic and,
for sanity, will want to enforce this anyway.

BUG=83

Change-Id: I80eb9f3333da3d751f98f25d9469860d1993a97a
Reviewed-on: https://boringssl-review.googlesource.com/9000
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-29 15:44:42 +00:00
David Benjamin
17e1292fe4 Make runner's -test parameter take glob patterns.
Per request from EKR. Also we have a lot of long test names, so this
seems generally a good idea.

Change-Id: Ie463f5367ec7d33005137534836005b571c8f424
Reviewed-on: https://boringssl-review.googlesource.com/9021
Reviewed-by: Adam Langley <agl@google.com>
2016-07-29 00:08:20 +00:00
David Benjamin
4497e58961 Switch finish_handshake to release_current_message.
With the previous DTLS change, the dispatch layer only cares about the
end of the handshake to know when to drop the current message. TLS 1.3
post-handshake messages will need a similar hook, so convert it to this
lower-level one.

BUG=83

Change-Id: I4c8c3ba55ba793afa065bf261a7bccac8816c348
Reviewed-on: https://boringssl-review.googlesource.com/8989
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-28 22:59:18 +00:00
David Benjamin
02edcd0098 Reject stray post-Finished messages in DTLS.
This is in preparation for switching finish_handshake to a
release_current_message hook. finish_handshake in DTLS is also
responsible for releasing any memory associated with extra messages in
the handshake.

Except that's not right and we need to make it an error anyway. Given
that the rest of the DTLS dispatch layer already strongly assumes there
is only one message in epoch one, putting the check in the fragment
processing works fine enough. Add tests for this.

This will certainly need revising when DTLS 1.3 happens (perhaps just a
version check, perhaps bringing finish_handshake back as a function that
can fail... which means we need a state just before SSL_ST_OK), but DTLS
1.3 post-handshake messages haven't really been written down, so let's
do the easy thing for now and add a test for when it gets more
interesting.

This removes the sequence number reset in the DTLS code. That reset
never did anything becase we don't and never will renego. We should make
sure DTLS 1.3 does not bring the reset back for post-handshake stuff.
(It was wrong in 1.2 too. Penultimate-flight retransmits and renego
requests are ambiguous in DTLS.)

BUG=83

Change-Id: I33d645a8550f73e74606030b9815fdac0c9fb682
Reviewed-on: https://boringssl-review.googlesource.com/8988
Reviewed-by: Adam Langley <agl@google.com>
2016-07-28 22:53:04 +00:00
David Benjamin
9fd9580137 Remove ssl->s3->message_complete in favor of ssl->init_msg.
This was only used so we knew when we had a current message to discard
and when we didn't. With init_msg being tracked better, we can use that
instead.

As part of this, switch the V2ClientHello hack to not using
reuse_message. Otherwise we have to fill in init_msg and friends in two
places.

The next change will require that we have a better handle on the "is
there a current message" boolean.

BUG=83

Change-Id: I917efacbad10806d492bbe51eda74c0779084d60
Reviewed-on: https://boringssl-review.googlesource.com/8987
Reviewed-by: Adam Langley <agl@google.com>
2016-07-28 22:52:47 +00:00
David Benjamin
a950948962 Use SSL3_HM_HEADER_LENGTH a bit more.
Somewhat clearer what it's for than just 4.

Change-Id: Ie7bb89ccdce188d61741da203acd624b49b69058
Reviewed-on: https://boringssl-review.googlesource.com/8986
Reviewed-by: Adam Langley <agl@google.com>
2016-07-28 22:49:57 +00:00
David Benjamin
481b9d2047 Remove begin_handshake and allocate init_buf lazily.
For TLS 1.3, we will need to process more complex post-handshake
messages. It is simplest if we use the same mechanism. In preparation,
allow ssl3_get_message to be called at any point.

Note that this stops reserving SSL3_RT_MAX_PLAIN_LENGTH in init_buf
right off the bat. Instead it will grow as-needed to accomodate the
handshake. SSL3_RT_MAX_PLAIN_LENGTH is rather larger than we probably
need to receive, particularly as a server, so this seems a good plan.

BUG=83

Change-Id: Id7f4024afc4c8a713b46b0d1625432315594350e
Reviewed-on: https://boringssl-review.googlesource.com/8985
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-28 22:07:28 +00:00
David Benjamin
7baf681a8b Convert all of our test private keys to PKCS#8 PEM blobs.
Right now they're RSA PRIVATE KEY or EC PRIVATE KEY which requires a bit
more effort to parse. It means the PEM header is necessary to parse
these. OpenSSL and Go automagically convert the format, but other shims
(namely NSS) may not.

Change-Id: I9fa2767dcf1fe6ceeea546390759e1c364a8f16f
Reviewed-on: https://boringssl-review.googlesource.com/9020
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-28 21:54:02 +00:00
Adam Langley
5a8d48ee8c Fix the comments for |SHA[256|384|512]_Transform|.
Change-Id: I6d552d26b3d72f6fffdc4d4d9fc3b5d82fb4e8bb
Reviewed-on: https://boringssl-review.googlesource.com/9010
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-28 21:49:48 +00:00
David Benjamin
49054544d3 Clear init_msg/init_num whenever we clear the backing store.
This API needs to be improved but, for the time being, keep the
invariant reasonable.

Change-Id: If94d41e7e7936e44de5ecb36da45f89f80df7784
Reviewed-on: https://boringssl-review.googlesource.com/8984
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-28 21:44:39 +00:00
David Benjamin
bd4679d133 Tidy up ssl3_get_message slightly.
Change-Id: Iccd86440bf8721098050fac220dc9bb80bbfc670
Reviewed-on: https://boringssl-review.googlesource.com/8983
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-28 21:36:55 +00:00
David Benjamin
21c0028d40 Implement KeyUpdate in Go.
Implemented in preparation for testing the C implementation. Tested
against itself.

BUG=74

Change-Id: Iec1b9ad22e09711fa4e67c97cc3eb257585c3ae5
Reviewed-on: https://boringssl-review.googlesource.com/8873
Reviewed-by: Nick Harper <nharper@chromium.org>
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-28 18:43:52 +00:00
Brian Smith
92d60c2059 Use Fermat's Little Theorem when converting points to affine.
Fermat's Little Theorem is already used for the custom curve implementations.
Use it, for the same reasons, for the ec_montgomery-based implementations.

I tested the performance (only) on x86-64 Windows.

Change-Id: Ibf770fd3f2d3e2cfe69f06bc12c81171624ff557
Reviewed-on: https://boringssl-review.googlesource.com/8924
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-28 18:29:32 +00:00
Brian Smith
286fbf2ce0 Add tests for |BN_mod_inverse| with modulus 1.
Zero is only a valid input to or output of |BN_mod_inverse| when the
modulus is one. |BN_MONT_CTX_set| actually depends on this, so test
that this works.

Change-Id: Ic18f1fe786f668394951d4309020c6ead95e5e28
Reviewed-on: https://boringssl-review.googlesource.com/8922
Reviewed-by: Adam Langley <agl@google.com>
Commit-Queue: Adam Langley <agl@google.com>
CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2016-07-28 18:07:03 +00:00