Commit graph

19079 commits

Author SHA1 Message Date
Matt Caswell
3982063756 Add documentation for SSL_key_update() and SSL_get_key_update_type()
This also adds documentation for the pre-existing and related
SSL_renegotiate*() functions.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
2017-02-17 10:28:01 +00:00
Matt Caswell
c2fd15f6a7 Fix a shadowed global variable warning
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
2017-02-17 10:28:01 +00:00
Matt Caswell
f14afcaa42 Updates following review feedback
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
2017-02-17 10:28:01 +00:00
Matt Caswell
82f992cbe0 Limit the number of KeyUpdate messages we can process
Too many KeyUpdate message could be inicative of a problem (e.g. an
infinite KeyUpdate loop if the peer always responds to a KeyUpdate message
with an "update_requested" KeyUpdate response), or (conceivably) an attack.
Either way we limit the number of KeyUpdate messages we are prepared to
handle.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
2017-02-17 10:28:00 +00:00
Matt Caswell
57389a3261 Actually update the keys when a KeyUpdate message is sent or received
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
2017-02-17 10:28:00 +00:00
Matt Caswell
5bf4793378 If we receive an "update_requested" KeyUpdate then respond with a KeyUpdate
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
2017-02-17 10:28:00 +00:00
Matt Caswell
8cdc8c5105 Add the ability for a server to receive a KeyUpdate message
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
2017-02-17 10:28:00 +00:00
Matt Caswell
b07b2a1b44 Add the ability for s_client to send a KeyUpdate message
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
2017-02-17 10:28:00 +00:00
Matt Caswell
53d1d07d30 Add a SSL_get_key_update_type() function
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
2017-02-17 10:28:00 +00:00
Matt Caswell
9412b3ad34 Add the ability for a client to send a KeyUpdate message
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
2017-02-17 10:28:00 +00:00
Matt Caswell
e1c3de4450 Add the ability for a client to receive a KeyUpdate message
This just receives the message. It doesn't actually update any keys yet.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
2017-02-17 10:28:00 +00:00
Matt Caswell
7d8c2dfa64 Add SSL_trace() support for KeyUpdate messages
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
2017-02-17 10:28:00 +00:00
Matt Caswell
34df45b531 Add a capability to s_server to send KeyUpdate messages
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
2017-02-17 10:28:00 +00:00
Matt Caswell
44c04a2e06 Provide a function to send a KeyUpdate message
This implements the server side KeyUpdate sending capability as well.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
2017-02-17 10:28:00 +00:00
David Benjamin
25b802bb85 Fix typo in x86_64-mont5.pl CFI directives
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2655)
2017-02-17 10:32:13 +01:00
Richard Levitte
bb0f7eca75 Add a test of the X509_STORE / X509_LOOKUP API
Fortunately, "openssl verify" makes good use of that API

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2652)
2017-02-16 21:09:09 +01:00
Richard Levitte
532e7b36d9 test/README: clarify test number groups
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2652)
2017-02-16 21:09:09 +01:00
Matt Caswell
d605fc3a0c Fix a mem leak in ssl_test_ctx.c
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2650)
2017-02-16 17:14:57 +00:00
Dr. Stephen Henson
86de658a84 Add client auth TLS 1.3 certificate selection tests
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2643)
2017-02-16 16:43:44 +00:00
Dr. Stephen Henson
b4cb7eb7df Add ECDSA client certificates
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2643)
2017-02-16 16:43:44 +00:00
Dr. Stephen Henson
ad4dd362e0 Use tls_choose_sigalg for client auth.
For client auth call tls_choose_sigalg to select the certificate
and signature algorithm. Use the selected algorithm in
tls_construct_cert_verify.

Remove obsolete tls12_get_sigandhash.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2643)
2017-02-16 16:43:44 +00:00
Dr. Stephen Henson
717a265aa5 Add client side support to tls_choose_sigalg.
Select appropriate signature algorithm and certificate for client
authentication using tls_choose_sigalg.

A lot of selection logic is very similar except not finding a
certificate is not a fatal error: we just do not present a
certificate.

For TLS 1.2 and earlier we only check the current certificate
is suitable (for compatibility with previous logic) for TLS 1.3
(where there are no compatibility issues) we support multiple
client certificates for different algorithms.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2643)
2017-02-16 16:43:44 +00:00
lrns
0cb8c9d85e Change req_check_len error message, it also accepts 20 bytes, but states 'less than' in the error message
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2648)
2017-02-16 09:18:11 -05:00
Benjamin Kaduk
ac879ed62a Use _WIN32 over WIN32 for preprocessor conditional
The intent seems to be that the WIN32 symbol is for things that are a direct
byproduct of being a windows-variant configuration and should be used for
feature en/disablement on windows systems.  Use of the _WIN32 symbol is more
widespread, being used to implement platform portability of more generic code.

We do define WIN32 in some situations in e_os.h, but that is not included
universally.

Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2642)
2017-02-16 08:59:47 -05:00
Matt Caswell
d4da1bb5ab Update CHANGES and NEWS for new release
Reviewed-by: Richard Levitte <levitte@openssl.org>
2017-02-16 10:10:05 +00:00
Matt Caswell
2c55b28a34 Remove an OPENSSL_assert() and replace with a soft assert and check
Following on from CVE-2017-3733, this removes the OPENSSL_assert() check
that failed and replaces it with a soft assert, and an explicit check of
value with an error return if it fails.

Reviewed-by: Richard Levitte <levitte@openssl.org>
2017-02-16 09:35:56 +00:00
Matt Caswell
28a31a0a10 Don't change the state of the ETM flags until CCS processing
In 1.1.0 changing the ciphersuite during a renegotiation can result in
a crash leading to a DoS attack. In master this does not occur with TLS
(instead you get an internal error, which is still wrong but not a security
issue) - but the problem still exists in the DTLS code.

The problem is caused by changing the flag indicating whether to use ETM
or not immediately on negotiation of ETM, rather than at CCS. Therefore,
during a renegotiation, if the ETM state is changing (usually due to a
change of ciphersuite), then an error/crash will occur.

Due to the fact that there are separate CCS messages for read and write
we actually now need two flags to determine whether to use ETM or not.

CVE-2017-3733

Reviewed-by: Richard Levitte <levitte@openssl.org>
2017-02-16 09:35:56 +00:00
Matt Caswell
cc22cd546b Provide a test for the Encrypt-Then-Mac renegotiation crash
In 1.1.0 changing the ciphersuite during a renegotiation can result in
a crash leading to a DoS attack. In master this does not occur with TLS
(instead you get an internal error, which is still wrong but not a security
issue) - but the problem still exists in the DTLS code.

This commit provides a test for the issue.

CVE-2017-3733

Reviewed-by: Richard Levitte <levitte@openssl.org>
2017-02-16 09:35:56 +00:00
Dr. Stephen Henson
7b3a4d6107 Fix warning
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2644)
2017-02-16 01:44:28 +00:00
Robert Scheck
af7e05c7c6 Handle negative reply for NNTP STARTTLS in s_client
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2629)
2017-02-15 20:28:44 -05:00
Kazuki Yamaguchi
a8f9576866 Properly zero cipher_data for ChaCha20-Poly1305 on cleanup
Fix a typo. Probably this has not been found because EVP_CIPHER_CTX is
smaller than EVP_CHACHA_AEAD_CTX and heap overflow does not occur.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2294)
2017-02-15 20:00:34 -05:00
Andy Polyakov
8653e78f43 crypto/armcap.c: short-circuit processor capability probe in iOS builds.
Capability probing by catching SIGILL appears to be problematic
on iOS. But since Apple universe is "monocultural", it's actually
possible to simply set pre-defined processor capability mask.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2617)
2017-02-15 23:16:23 +01:00
Andy Polyakov
c93f06c12f ARMv4 assembly pack: harmonize Thumb-ification of iOS build.
Three modules were left behind in a285992763.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2617)
2017-02-15 23:16:01 +01:00
Dr. Stephen Henson
59088e43b1 Set current certificate to selected certificate.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2637)
2017-02-15 15:33:15 +00:00
Andy Polyakov
399976c7ba sha/asm/*-x86_64.pl: add CFI annotations.
Reviewed-by: Rich Salz <rsalz@openssl.org>
2017-02-15 15:43:05 +01:00
Bernd Edlinger
ed874fac63 Rework error handling of custom_ext_meth_add towards strong exception safety.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2636)
2017-02-15 08:37:52 -05:00
Matt Caswell
bb90d02a71 Fix merge issue
Causes make update to fail.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2634)
2017-02-15 10:50:39 +00:00
Dr. Stephen Henson
a34a9df071 Skip curve check if sigalg doesn't specify a curve.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2623)
2017-02-15 02:23:55 +00:00
Dr. Stephen Henson
a497cf2516 Use CERT_PKEY pointer instead of index
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2623)
2017-02-15 02:23:54 +00:00
Dr. Stephen Henson
f695571e10 Simplify tls_construct_server_key_exchange
Use negotiated signature algorithm and certificate index in
tls_construct_key_exchange instead of recalculating it.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2623)
2017-02-15 02:23:54 +00:00
Dr. Stephen Henson
f365a3e2e5 Use cert_index and sigalg
Now the certificate and signature algorithm is set in one place we
can use it directly insetad of recalculating it. The old functions
ssl_get_server_send_pkey() and ssl_get_server_cert_index() are no
longer required.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2623)
2017-02-15 02:23:54 +00:00
Dr. Stephen Henson
0972bc5ced Add sigalg for earlier TLS versions
Update tls_choose_sigalg to choose a signature algorithm for all
versions of TLS not just 1.3.

For TLS 1.2 we choose the highest preference signature algorithm
for the chosen ciphersuite.

For TLS 1.1 and earlier the signature algorithm is determined by
the ciphersuite alone. For RSA we use a special MD5+SHA1 signature
algorithm.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2623)
2017-02-15 02:23:54 +00:00
Dr. Stephen Henson
4a419f6018 Change tls_choose_sigalg so it can set errors and alerts.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2623)
2017-02-15 02:23:54 +00:00
Dr. Stephen Henson
4020c0b33b add ssl_has_cert
Add inline function ssl_has_cert which checks to see if a certificate and
private key for a given index are not NULL.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2623)
2017-02-15 02:23:54 +00:00
FdaSilvaYY
7e12cdb52e Fix a few typos
[skip ci]

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2571)
2017-02-14 15:48:51 -05:00
Guido Vranken
7c120357e5 Remove obsolete comment
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1613)
2017-02-14 14:52:24 -05:00
Guido Vranken
873019f2c3 Prevents that OPENSSL_gmtime incorrectly signals success if gmtime_r fails, and that struct* tm result's possibly uninitialized content is used
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1613)
2017-02-14 14:52:24 -05:00
Bernd Edlinger
57b0d651f0 Use TLSEXT_KEYNAME_LENGTH in tls_decrypt_ticket.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2618)
2017-02-14 14:37:59 -05:00
Guido Vranken
7f07149d25 Prevent allocations of size 0 in sh_init, which are not possible with the default OPENSSL_zalloc, but are possible if the user has installed their own allocator using CRYPTO_set_mem_functions. If the 0-allocations succeeds, the secure heap code will later access (at least) the first byte of that space, which is technically an OOB access. This could lead to problems with some custom allocators that only return a valid pointer for subsequent free()-ing, and do not expect that the pointer is actually dereferenced.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2605)
2017-02-14 14:28:34 -05:00
Robert Scheck
20967afb7f Add Sieve support (RFC 5804) to s_client ("-starttls sieve")
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2300)
2017-02-14 14:15:28 -05:00