oweals/openssl.git
7 years agoUpdates following review feedback
Matt Caswell [Tue, 14 Feb 2017 11:20:44 +0000 (11:20 +0000)]
Updates following review feedback

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)

7 years agoLimit the number of KeyUpdate messages we can process
Matt Caswell [Mon, 13 Feb 2017 11:55:38 +0000 (11:55 +0000)]
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)

7 years agoActually update the keys when a KeyUpdate message is sent or received
Matt Caswell [Fri, 10 Feb 2017 17:43:09 +0000 (17:43 +0000)]
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)

7 years agoIf we receive an "update_requested" KeyUpdate then respond with a KeyUpdate
Matt Caswell [Thu, 9 Feb 2017 16:00:12 +0000 (16:00 +0000)]
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)

7 years agoAdd the ability for a server to receive a KeyUpdate message
Matt Caswell [Thu, 9 Feb 2017 15:29:45 +0000 (15:29 +0000)]
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)

7 years agoAdd the ability for s_client to send a KeyUpdate message
Matt Caswell [Thu, 9 Feb 2017 13:33:09 +0000 (13:33 +0000)]
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)

7 years agoAdd a SSL_get_key_update_type() function
Matt Caswell [Thu, 9 Feb 2017 15:05:46 +0000 (15:05 +0000)]
Add a SSL_get_key_update_type() function

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)

7 years agoAdd the ability for a client to send a KeyUpdate message
Matt Caswell [Thu, 9 Feb 2017 13:12:00 +0000 (13:12 +0000)]
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)

7 years agoAdd the ability for a client to receive a KeyUpdate message
Matt Caswell [Thu, 9 Feb 2017 12:07:31 +0000 (12:07 +0000)]
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)

7 years agoAdd SSL_trace() support for KeyUpdate messages
Matt Caswell [Wed, 8 Feb 2017 17:08:48 +0000 (17:08 +0000)]
Add SSL_trace() support for KeyUpdate messages

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)

7 years agoAdd a capability to s_server to send KeyUpdate messages
Matt Caswell [Wed, 8 Feb 2017 16:52:23 +0000 (16:52 +0000)]
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)

7 years agoProvide a function to send a KeyUpdate message
Matt Caswell [Wed, 8 Feb 2017 09:15:22 +0000 (09:15 +0000)]
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)

7 years agoFix typo in x86_64-mont5.pl CFI directives
David Benjamin [Thu, 16 Feb 2017 22:39:40 +0000 (17:39 -0500)]
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)

7 years agoAdd a test of the X509_STORE / X509_LOOKUP API
Richard Levitte [Thu, 16 Feb 2017 20:07:33 +0000 (21:07 +0100)]
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)

7 years agotest/README: clarify test number groups
Richard Levitte [Thu, 16 Feb 2017 20:06:42 +0000 (21:06 +0100)]
test/README: clarify test number groups

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2652)

7 years agoFix a mem leak in ssl_test_ctx.c
Matt Caswell [Thu, 16 Feb 2017 14:47:26 +0000 (14:47 +0000)]
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)

7 years agoAdd client auth TLS 1.3 certificate selection tests
Dr. Stephen Henson [Wed, 15 Feb 2017 22:16:09 +0000 (22:16 +0000)]
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)

7 years agoAdd ECDSA client certificates
Dr. Stephen Henson [Wed, 15 Feb 2017 22:03:04 +0000 (22:03 +0000)]
Add ECDSA client certificates

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2643)

7 years agoUse tls_choose_sigalg for client auth.
Dr. Stephen Henson [Wed, 15 Feb 2017 16:23:49 +0000 (16:23 +0000)]
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)

7 years agoAdd client side support to tls_choose_sigalg.
Dr. Stephen Henson [Wed, 15 Feb 2017 16:19:43 +0000 (16:19 +0000)]
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)

7 years agoChange req_check_len error message, it also accepts 20 bytes, but states 'less than...
lrns [Thu, 16 Feb 2017 11:27:55 +0000 (12:27 +0100)]
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)

7 years agoUse _WIN32 over WIN32 for preprocessor conditional
Benjamin Kaduk [Thu, 29 Dec 2016 17:38:24 +0000 (11:38 -0600)]
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)

7 years agoUpdate CHANGES and NEWS for new release
Matt Caswell [Thu, 16 Feb 2017 09:51:56 +0000 (09:51 +0000)]
Update CHANGES and NEWS for new release

Reviewed-by: Richard Levitte <levitte@openssl.org>
7 years agoRemove an OPENSSL_assert() and replace with a soft assert and check
Matt Caswell [Fri, 3 Feb 2017 14:45:49 +0000 (14:45 +0000)]
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>
7 years agoDon't change the state of the ETM flags until CCS processing
Matt Caswell [Fri, 3 Feb 2017 14:06:20 +0000 (14:06 +0000)]
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>
7 years agoProvide a test for the Encrypt-Then-Mac renegotiation crash
Matt Caswell [Fri, 3 Feb 2017 11:21:07 +0000 (11:21 +0000)]
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>
7 years agoFix warning
Dr. Stephen Henson [Thu, 16 Feb 2017 01:29:14 +0000 (01:29 +0000)]
Fix warning

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2644)

7 years agoHandle negative reply for NNTP STARTTLS in s_client
Robert Scheck [Tue, 14 Feb 2017 20:47:25 +0000 (21:47 +0100)]
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)

7 years agoProperly zero cipher_data for ChaCha20-Poly1305 on cleanup
Kazuki Yamaguchi [Thu, 26 Jan 2017 04:01:30 +0000 (13:01 +0900)]
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)

7 years agocrypto/armcap.c: short-circuit processor capability probe in iOS builds.
Andy Polyakov [Wed, 15 Feb 2017 11:01:09 +0000 (12:01 +0100)]
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)

7 years agoARMv4 assembly pack: harmonize Thumb-ification of iOS build.
Andy Polyakov [Mon, 13 Feb 2017 17:16:16 +0000 (18:16 +0100)]
ARMv4 assembly pack: harmonize Thumb-ification of iOS build.

Three modules were left behind in a285992763f3961f69a8d86bf7dfff020a08cef9.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2617)

7 years agoSet current certificate to selected certificate.
Dr. Stephen Henson [Wed, 15 Feb 2017 15:28:56 +0000 (15:28 +0000)]
Set current certificate to selected certificate.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2637)

7 years agosha/asm/*-x86_64.pl: add CFI annotations.
Andy Polyakov [Mon, 13 Feb 2017 21:34:51 +0000 (22:34 +0100)]
sha/asm/*-x86_64.pl: add CFI annotations.

Reviewed-by: Rich Salz <rsalz@openssl.org>
7 years agoRework error handling of custom_ext_meth_add towards strong exception safety.
Bernd Edlinger [Wed, 15 Feb 2017 10:36:17 +0000 (11:36 +0100)]
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)

7 years agoFix merge issue
Matt Caswell [Wed, 15 Feb 2017 09:44:46 +0000 (09:44 +0000)]
Fix merge issue

Causes make update to fail.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2634)

7 years agoSkip curve check if sigalg doesn't specify a curve.
Dr. Stephen Henson [Tue, 14 Feb 2017 14:27:15 +0000 (14:27 +0000)]
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)

7 years agoUse CERT_PKEY pointer instead of index
Dr. Stephen Henson [Tue, 14 Feb 2017 00:35:26 +0000 (00:35 +0000)]
Use CERT_PKEY pointer instead of index

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2623)

7 years agoSimplify tls_construct_server_key_exchange
Dr. Stephen Henson [Mon, 13 Feb 2017 18:07:00 +0000 (18:07 +0000)]
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)

7 years agoUse cert_index and sigalg
Dr. Stephen Henson [Mon, 13 Feb 2017 16:32:06 +0000 (16:32 +0000)]
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)

7 years agoAdd sigalg for earlier TLS versions
Dr. Stephen Henson [Mon, 13 Feb 2017 16:04:07 +0000 (16:04 +0000)]
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)

7 years agoChange tls_choose_sigalg so it can set errors and alerts.
Dr. Stephen Henson [Mon, 13 Feb 2017 15:50:43 +0000 (15:50 +0000)]
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)

7 years agoadd ssl_has_cert
Dr. Stephen Henson [Mon, 13 Feb 2017 15:40:21 +0000 (15:40 +0000)]
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)

7 years agoFix a few typos
FdaSilvaYY [Mon, 6 Feb 2017 23:05:06 +0000 (00:05 +0100)]
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)

7 years agoRemove obsolete comment
Guido Vranken [Sat, 11 Feb 2017 21:41:38 +0000 (22:41 +0100)]
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)

7 years agoPrevents that OPENSSL_gmtime incorrectly signals success if gmtime_r fails, and that...
Guido Vranken [Thu, 22 Sep 2016 20:48:44 +0000 (22:48 +0200)]
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)

7 years agoUse TLSEXT_KEYNAME_LENGTH in tls_decrypt_ticket.
Bernd Edlinger [Mon, 13 Feb 2017 17:36:13 +0000 (18:36 +0100)]
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)

7 years agoPrevent allocations of size 0 in sh_init, which are not possible with the default...
Guido Vranken [Mon, 13 Feb 2017 00:36:43 +0000 (01:36 +0100)]
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)

7 years agoAdd Sieve support (RFC 5804) to s_client ("-starttls sieve")
Robert Scheck [Thu, 9 Feb 2017 21:20:59 +0000 (22:20 +0100)]
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)

7 years agoAdd no-ec build
Rich Salz [Tue, 14 Feb 2017 16:51:22 +0000 (11:51 -0500)]
Add no-ec build

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2626)

7 years agoMake -xcert work again.
Dr. Stephen Henson [Tue, 14 Feb 2017 17:18:00 +0000 (17:18 +0000)]
Make -xcert work again.

When a certificate is prepended update the list pointer.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2628)

7 years agoFix no-ec compilation
Matt Caswell [Tue, 14 Feb 2017 15:57:50 +0000 (15:57 +0000)]
Fix no-ec compilation

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2624)

7 years agoRemove a double call to ssl3_send_alert()
Matt Caswell [Wed, 8 Feb 2017 17:27:09 +0000 (17:27 +0000)]
Remove a double call to ssl3_send_alert()

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)

7 years agoFix a bogus uninit variable warning
Matt Caswell [Wed, 8 Feb 2017 17:16:25 +0000 (17:16 +0000)]
Fix a bogus uninit variable warning

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)

7 years agoAdd a bytestogroup macro
Matt Caswell [Mon, 6 Feb 2017 16:52:38 +0000 (16:52 +0000)]
Add a bytestogroup macro

For converting the 2 byte group id into an unsigned int.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)

7 years agoVarious style fixes following review feedback
Matt Caswell [Mon, 6 Feb 2017 16:47:29 +0000 (16:47 +0000)]
Various style fixes following review feedback

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)

7 years agoUpdate the tls13messages test to add some HRR scenarios
Matt Caswell [Thu, 2 Feb 2017 16:28:45 +0000 (16:28 +0000)]
Update the tls13messages test to add some HRR scenarios

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)

7 years agoUpdate the kex modes tests to check various HRR scenarios
Matt Caswell [Thu, 2 Feb 2017 16:06:50 +0000 (16:06 +0000)]
Update the kex modes tests to check various HRR scenarios

Make sure we get an HRR in the right circumstances based on kex mode.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)

7 years agoUpdate TLSProxy to know about HelloRetryRequest messages
Matt Caswell [Thu, 2 Feb 2017 16:06:28 +0000 (16:06 +0000)]
Update TLSProxy to know about HelloRetryRequest messages

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)

7 years agoUpdate test counting in checkhandshake.pm
Matt Caswell [Thu, 2 Feb 2017 14:54:24 +0000 (14:54 +0000)]
Update test counting in checkhandshake.pm

Previously counting the number of tests in checkhandshake.pm took an
initial guess and then modified it based on various known special
cases. That is becoming increasingly untenable, so this changes it to
properly calculate the number of tests we expect to run.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)

7 years agoUpdate the key_share tests for HelloRetryRequest
Matt Caswell [Thu, 2 Feb 2017 13:12:08 +0000 (13:12 +0000)]
Update the key_share tests for HelloRetryRequest

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)

7 years agoAdd trace support for HelloRetryRequest
Matt Caswell [Thu, 2 Feb 2017 11:16:25 +0000 (11:16 +0000)]
Add trace support for HelloRetryRequest

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)

7 years agoImplement support for resumption with a HelloRetryRequest
Matt Caswell [Wed, 1 Feb 2017 17:10:45 +0000 (17:10 +0000)]
Implement support for resumption with a HelloRetryRequest

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)

7 years agoAdd client side support for parsing Hello Retry Request
Matt Caswell [Wed, 1 Feb 2017 13:31:27 +0000 (13:31 +0000)]
Add client side support for parsing Hello Retry Request

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)

7 years agoAdd server side support for creating the Hello Retry Request message
Matt Caswell [Mon, 30 Jan 2017 16:16:28 +0000 (16:16 +0000)]
Add server side support for creating the Hello Retry Request message

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)

7 years agoMake the context available to the extensions parse and construction funcs
Matt Caswell [Tue, 31 Jan 2017 17:00:12 +0000 (17:00 +0000)]
Make the context available to the extensions parse and construction funcs

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)

7 years agomem leak on error path and error propagation fix
Yuchi [Mon, 6 Feb 2017 00:33:47 +0000 (19:33 -0500)]
mem leak on error path and error propagation fix

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2559)

7 years agoaes/asm/*-x86_64.pl: add CFI annotations.
Andy Polyakov [Fri, 10 Feb 2017 10:02:09 +0000 (11:02 +0100)]
aes/asm/*-x86_64.pl: add CFI annotations.

Reviewed-by: Rich Salz <rsalz@openssl.org>
7 years agoperlasm/x86_64-xlate.pl: recognize even offset(%reg) in cfa_expression.
Andy Polyakov [Fri, 10 Feb 2017 09:59:45 +0000 (10:59 +0100)]
perlasm/x86_64-xlate.pl: recognize even offset(%reg) in cfa_expression.

This is handy when "offset(%reg)" is a perl variable.

Reviewed-by: Rich Salz <rsalz@openssl.org>
7 years agoec/asm/ecp_nistz256-x86_64.pl: add CFI directives.
Andy Polyakov [Fri, 10 Feb 2017 11:20:18 +0000 (12:20 +0100)]
ec/asm/ecp_nistz256-x86_64.pl: add CFI directives.

Reviewed-by: Rich Salz <rsalz@openssl.org>
7 years agoec/asm/ecp_nistz256-x86_64.pl: fix typo-bug in Win64 SE handler.
Andy Polyakov [Fri, 10 Feb 2017 10:43:42 +0000 (11:43 +0100)]
ec/asm/ecp_nistz256-x86_64.pl: fix typo-bug in Win64 SE handler.

Thanks to Jun Sun for spotting this.

Reviewed-by: Rich Salz <rsalz@openssl.org>
7 years agoFurther improvements to ASYNC_WAIT_CTX_clear_fd
Andrea Grandi [Fri, 10 Feb 2017 10:23:21 +0000 (10:23 +0000)]
Further improvements to ASYNC_WAIT_CTX_clear_fd

Remove call to cleanup function
Use only one loop to find previous element

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2581)

7 years agoRemove fd from the list when the engine clears the wait context before pause
Andrea Grandi [Fri, 3 Feb 2017 05:46:17 +0000 (05:46 +0000)]
Remove fd from the list when the engine clears the wait context before pause

This fixes the num of fds added/removed returned by ASYNC_WAIT_CTX_get_changed_fds

Previously, the numbers were not consistent with the fds actually written in
the buffers since the fds that have been both added and removed are explicitly
ignored in the loop.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2581)

7 years agoAdd test to show wrong behavior of ASYNC_WAIT_CTX
Andrea Grandi [Thu, 26 Jan 2017 03:17:54 +0000 (03:17 +0000)]
Add test to show wrong behavior of ASYNC_WAIT_CTX

This happens when a fd is added and then immediately removed from the
ASYNC_WAIT_CTX before pausing the job.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2581)

7 years ago{md5,rc4}/asm/*-x86_64.pl: add CFI annotations.
Andy Polyakov [Sat, 11 Feb 2017 21:31:47 +0000 (22:31 +0100)]
{md5,rc4}/asm/*-x86_64.pl: add CFI annotations.

Reviewed-by: Rich Salz <rsalz@openssl.org>
7 years agomodes/asm/*-x86_64.pl: add CFI annotations.
Andy Polyakov [Sat, 11 Feb 2017 21:08:12 +0000 (22:08 +0100)]
modes/asm/*-x86_64.pl: add CFI annotations.

Reviewed-by: Rich Salz <rsalz@openssl.org>
7 years agoDES keys are not 7 days long.
Darren Tucker [Sun, 12 Feb 2017 23:36:29 +0000 (10:36 +1100)]
DES keys are not 7 days long.

CLA: trivial
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2604)

7 years agotest_rehash does nothing, have it do something
Richard Levitte [Fri, 10 Feb 2017 21:50:24 +0000 (22:50 +0100)]
test_rehash does nothing, have it do something

test/recipes/40-test_rehash.t uses test files from certs/demo, which
doesn't exist any longer.  Have it use PEM files from test/ instead.

Because rehash wants only one certificate or CRL per file, we must
also filter those PEM files to produce test files with a single object
each.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2594)

7 years agosha/asm/sha1-x86_64.pl: add CFI annotations.
Adam Langley [Fri, 10 Feb 2017 18:19:06 +0000 (10:19 -0800)]
sha/asm/sha1-x86_64.pl: add CFI annotations.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2590)

7 years agoReplace SSL_PKEY_RSA_ENC, SSL_PKEY_RSA_SIGN
Dr. Stephen Henson [Fri, 10 Feb 2017 04:23:53 +0000 (04:23 +0000)]
Replace SSL_PKEY_RSA_ENC, SSL_PKEY_RSA_SIGN

The original intent of SSL_PKEY_RSA_SIGN and SSL_PKEY_RSA_ENC was to
support two different keys for RSA signing and decrypt. However this
was never implemented and we only ever set one key and the other was
always NULL. Replace with single SSL_PKEY_RSA type.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2587)

7 years agoCheck index >= 0 as 0 is a valid index.
Dr. Stephen Henson [Fri, 10 Feb 2017 04:22:18 +0000 (04:22 +0000)]
Check index >= 0 as 0 is a valid index.

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2587)

7 years agoperlasm/x86_64-xlate.pl: fix pair of typo-bugs in the new cfi_directive.
Andy Polyakov [Fri, 10 Feb 2017 10:27:54 +0000 (11:27 +0100)]
perlasm/x86_64-xlate.pl: fix pair of typo-bugs in the new cfi_directive.

.cfi_{start|end}proc and .cfi_def_cfa were not tracked.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2585)

7 years agoperlasm/x86_64-xlate.pl: typo fix in comment.
Adam Langley [Thu, 9 Feb 2017 22:37:01 +0000 (14:37 -0800)]
perlasm/x86_64-xlate.pl: typo fix in comment.

CLA: trivial

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2584)

7 years agoFix copy-pasteism in CFI directives.
Adam Langley [Thu, 9 Feb 2017 20:02:08 +0000 (12:02 -0800)]
Fix copy-pasteism in CFI directives.

I don't think this actually affects anything since the cfi_restore
directives aren't strictly needed anyway. (The old values are still in
memory so either will do.)

CLA: trivial

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2582)

7 years agoRestore EVP_CIPH_FLAG_LENGTH_BITS working properly
Lukasz Pawelczyk [Thu, 17 Nov 2016 09:31:39 +0000 (10:31 +0100)]
Restore EVP_CIPH_FLAG_LENGTH_BITS working properly

EVP_CIPH_FLAG_LENGTH_BITS flag for CFB1 has been broken with the
introduction of the is_partially_overlapping() check that did not take
it into the account (treating number of bits passed as bytes). This
remedies that and allows this flag to work as intended.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1942)

7 years agospeed.c: simplify aggregation of ecdh --multi results
Cristian Stoica [Tue, 2 Aug 2016 08:41:30 +0000 (11:41 +0300)]
speed.c: simplify aggregation of ecdh --multi results

CLA: trivial

Signed-off-by: Cristian Stoica <cristian.stoica@nxp.com>
Reviewed-by: Geoff Thorpe <geoff@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1377)

7 years agospeed.c: simplify aggregation of ecdsa --multi results
Cristian Stoica [Tue, 2 Aug 2016 08:38:45 +0000 (11:38 +0300)]
speed.c: simplify aggregation of ecdsa --multi results

CLA: trivial

Signed-off-by: Cristian Stoica <cristian.stoica@nxp.com>
Reviewed-by: Geoff Thorpe <geoff@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1377)

7 years agospeed.c: simplify aggregation of dsa --multi results
Cristian Stoica [Tue, 2 Aug 2016 08:22:27 +0000 (11:22 +0300)]
speed.c: simplify aggregation of dsa --multi results

CLA: trivial

Signed-off-by: Cristian Stoica <cristian.stoica@nxp.com>
Reviewed-by: Geoff Thorpe <geoff@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1377)

7 years agospeed.c: simplify aggregation of rsa --multi results
Cristian Stoica [Tue, 2 Aug 2016 08:13:00 +0000 (11:13 +0300)]
speed.c: simplify aggregation of rsa --multi results

CLA: trivial

Signed-off-by: Cristian Stoica <cristian.stoica@nxp.com>
Reviewed-by: Geoff Thorpe <geoff@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1377)

7 years agoDon't read uninitialised data for short session IDs.
David Benjamin [Thu, 9 Feb 2017 20:13:13 +0000 (15:13 -0500)]
Don't read uninitialised data for short session IDs.

While it's always safe to read |SSL_MAX_SSL_SESSION_ID_LENGTH| bytes
from an |SSL_SESSION|'s |session_id| array, the hash function would do
so with without considering if all those bytes had been written to.

This change checks |session_id_length| before possibly reading
uninitialised memory. Since the result of the hash function was already
attacker controlled, and since a lookup of a short session ID will
always fail, it doesn't appear that this is anything more than a clean
up.

In particular, |ssl_get_prev_session| uses a stack-allocated placeholder
|SSL_SESSION| as a lookup key, so the |session_id| array may be
uninitialised.

This was originally found with libFuzzer and MSan in
https://boringssl.googlesource.com/boringssl/+/e976e4349d693b4bbb97e1694f45be5a1b22c8c7,
then by Robert Swiecki with honggfuzz and MSan here. Thanks to both.

Reviewed-by: Geoff Thorpe <geoff@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2583)

7 years agobn/asm/x86_64*: add DWARF CFI directives.
Andy Polyakov [Wed, 8 Feb 2017 09:12:28 +0000 (10:12 +0100)]
bn/asm/x86_64*: add DWARF CFI directives.

Reviewed-by: Rich Salz <rsalz@openssl.org>
7 years agoperlasm/x86_64-xlate.pl: recognize DWARF CFI directives.
Andy Polyakov [Mon, 6 Feb 2017 07:58:34 +0000 (08:58 +0100)]
perlasm/x86_64-xlate.pl: recognize DWARF CFI directives.

CFI directives annotate instructions that are significant for stack
unwinding procedure. In addition to directives recognized by GNU
assembler this module implements three synthetic ones:

- .cfi_push annotates push instructions in prologue and translates to
  .cfi_adjust_cfa_offset (if needed) and .cfi_offset;
- .cfi_pop annotates pop instructions in epilogue and translates to
  .cfi_adjust_cfs_offset (if needed) and .cfi_restore;
- .cfi_cfa_expression encodes DW_CFA_def_cfa_expression and passes it
  to .cfi_escape as byte vector;

CFA expression syntax is made up mix of DWARF operator suffixes [subset
of] and references to registers with optional bias. Following example
describes offloaded original stack pointer at specific offset from
current stack pointer:

.cfi_cfa_expression %rsp+40,deref,+8

Final +8 has everything to do with the fact that CFA, Canonical Frame
Address, is reference to top of caller's stack, and on x86_64 call to
subroutine pushes 8-byte return address.

Triggered by request from Adam Langley.

Reviewed-by: Rich Salz <rsalz@openssl.org>
7 years agoperlasm/x86_64-xlate.pl: remove obsolete .picmeup synthetic directive.
Andy Polyakov [Sat, 4 Feb 2017 12:26:15 +0000 (13:26 +0100)]
perlasm/x86_64-xlate.pl: remove obsolete .picmeup synthetic directive.

Reviewed-by: Rich Salz <rsalz@openssl.org>
7 years agoperlasm/x86_64-xlate.pl: minor readability updates.
Andy Polyakov [Sat, 4 Feb 2017 12:24:42 +0000 (13:24 +0100)]
perlasm/x86_64-xlate.pl: minor readability updates.

Reviewed-by: Rich Salz <rsalz@openssl.org>
7 years agobn/asm/rsaz-avx2.pl: refine Win64 SE handler.
Andy Polyakov [Wed, 8 Feb 2017 09:09:21 +0000 (10:09 +0100)]
bn/asm/rsaz-avx2.pl: refine Win64 SE handler.

Reviewed-by: Rich Salz <rsalz@openssl.org>
7 years agoFix a typo in the X509_get0_subject_key_id() documentation
Matt Caswell [Tue, 7 Feb 2017 14:17:57 +0000 (14:17 +0000)]
Fix a typo in the X509_get0_subject_key_id() documentation

Fixes a copy&paste error

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2568)

7 years agoAdd needed module in 25-test_sid.t
Richard Levitte [Thu, 9 Feb 2017 09:30:44 +0000 (10:30 +0100)]
Add needed module in 25-test_sid.t

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2579)

7 years agoSSL_get_shared_sigalgs: handle negative idx parameter
Peter Wu [Thu, 2 Feb 2017 11:11:10 +0000 (12:11 +0100)]
SSL_get_shared_sigalgs: handle negative idx parameter

When idx is negative (as is the case with do_print_sigalgs in
apps/s_cb.c), AddressSanitizer complains about a buffer overflow (read).
Even if the pointer is not dereferenced, this is undefined behavior.

Change the user not to use "-1" as index since the function is
documented to return 0 on out-of-range values.

Tested with `openssl s_server` and `curl -k https://localhost:4433`.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2349)

7 years agoBecause our test sid file contains EC, don't try it when configured no-ec
Richard Levitte [Mon, 6 Feb 2017 16:49:41 +0000 (17:49 +0100)]
Because our test sid file contains EC, don't try it when configured no-ec

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2564)

7 years agoFix crash in tls13_enc
Matt Caswell [Wed, 8 Feb 2017 09:33:44 +0000 (09:33 +0000)]
Fix crash in tls13_enc

If s->s3->tmp.new_cipher is NULL then a crash can occur. This can happen
if an alert gets sent after version negotiation (i.e. we have selected
TLSv1.3 and ended up in tls13_enc), but before a ciphersuite has been
selected.

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2575)

7 years agoUpdate documentation
Dr. Stephen Henson [Sat, 4 Feb 2017 13:12:49 +0000 (13:12 +0000)]
Update documentation

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2550)