Commit graph

621 commits

Author SHA1 Message Date
Andres Amaya Garcia 01692531c6 Document code silently discarding invalid records 2017-09-14 20:20:31 +01:00
Andres Amaya Garcia 2fad94b193 Dont send alert on invalid DTLS record type
Do not send fatal alerts when receiving a record with an invalid header
while running DTLS as this is not compliant behaviour.
2017-09-14 20:18:37 +01:00
Hanno Becker bdf3905fff Ensure application data records are not kept when fully processed
This commit fixes the following case: If a client is both expecting a
SERVER_HELLO and has an application data record that's partially
processed in flight (that's the situation the client gets into after
receiving a ServerHelloRequest followed by ApplicationData), a
subsequent call to mbedtls_ssl_read will set keep_current_message = 1
when seeing the unexpected application data, but not reset it to 0
after the application data has been processed. This commit fixes this.

It also documents and suggests how the problem might be solved in a
more structural way on the long run.
2017-06-09 10:42:03 +01:00
Hanno Becker bb9dd0c044 Add hard assertion to mbedtls_ssl_read_record_layer
This commit adds a hard assertion to mbedtls_ssl_read_record_layer
triggering if both ssl->in_hslen and ssl->in_offt are not 0. This
should never happen, and if it does, there's no sensible way of
telling whether the previous message was a handshake or an application
data message.
2017-06-08 11:55:34 +01:00
Hanno Becker 4a810fba69 Fix mbedtls_ssl_read
Don't fetch a new record in mbedtls_ssl_read_record_layer as long as an application data record is being processed.
2017-06-08 10:12:16 +01:00
Hanno Becker af0665d8b0 Simplify retaining of messages for future processing
There are situations in which it is not clear what message to expect
next. For example, the message following the ServerHello might be
either a Certificate, a ServerKeyExchange or a CertificateRequest. We
deal with this situation in the following way: Initially, the message
processing function for one of the allowed message types is called,
which fetches and decodes a new message. If that message is not the
expected one, the function returns successfully (instead of throwing
an error as usual for unexpected messages), and the handshake
continues to the processing function for the next possible message. To
not have this function fetch a new message, a flag in the SSL context
structure is used to indicate that the last message was retained for
further processing, and if that's set, the following processing
function will not fetch a new record.

This commit simplifies the usage of this message-retaining parameter
by doing the check within the record-fetching routine instead of the
specific message-processing routines. The code gets cleaner this way
and allows retaining messages to be used in other situations as well
without much effort. This will be used in the next commits.
2017-06-08 10:12:16 +01:00
Hanno Becker e6706e62d8 Add tests for missing CA chains and bad curves.
This commit adds four tests to tests/ssl-opt.sh:
(1) & (2): Check behaviour of optional/required verification when the
trusted CA chain is empty.
(3) & (4): Check behaviour of optional/required verification when the
client receives a server certificate with an unsupported curve.
2017-06-07 11:26:59 +01:00
Hanno Becker 39ae8cd207 Fix implementation of VERIFY_OPTIONAL verification mode
This commit changes the behaviour of mbedtls_ssl_parse_certificate
to make the two authentication modes MBEDTLS_SSL_VERIFY_REQUIRED and
MBEDTLS_SSL_VERIFY_OPTIONAL be in the following relationship:

    Mode == MBEDTLS_SSL_VERIFY_REQUIRED
<=> Mode == MBEDTLS_SSL_VERIFY_OPTIONAL + check verify result

Also, it changes the behaviour to perform the certificate chain
verification even if the trusted CA chain is empty. Previously, the
function failed in this case, even when using optional verification,
which was brought up in #864.
2017-06-07 11:13:19 +01:00
Manuel Pégourié-Gonnard 383a118338 Merge remote-tracking branch 'gilles/IOTSSL-1330/development' into development
* gilles/IOTSSL-1330/development:
  Changelog entry for the bug fixes
  SSLv3: when refusing renegotiation, stop processing
  Ignore failures when sending fatal alerts
  Cleaned up double variable declaration
  Code portability fix
  Added changelog entry
  Send TLS alerts in many more cases
  Skip all non-executables in run-test-suites.pl
  SSL tests: server requires auth, client has no certificate
  Balanced braces across preprocessor conditionals
  Support setting the ports on the command line
2017-06-06 19:22:41 +02:00
Gilles Peskine 5d2511c4d4 SHA-1 deprecation: allow it in key exchange
By default, keep allowing SHA-1 in key exchange signatures. Disabling
it causes compatibility issues, especially with clients that use
TLS1.2 but don't send the signature_algorithms extension.

SHA-1 is forbidden in certificates by default, since it's vulnerable
to offline collision-based attacks.
2017-06-06 18:44:14 +02:00
Gilles Peskine 5e79cb3662 Remove SHA-1 in TLS by default
Default to forbidding the use of SHA-1 in TLS where it is unsafe: for
certificate signing, and as the signature hash algorithm for the TLS
1.2 handshake signature. SHA-1 remains allowed in HMAC-SHA-1 in the
XXX_SHA ciphersuites and in the PRF for TLS <= 1.1.

For easy backward compatibility for use in controlled environments,
turn on the MBEDTLS_TLS_DEFAULT_ALLOW_SHA1 compiled-time option.
2017-06-06 18:44:13 +02:00
Manuel Pégourié-Gonnard 23b33f8663 Merge remote-tracking branch 'hanno/sig_hash_compatibility' into development
* hanno/sig_hash_compatibility:
  Improve documentation
  Split long lines
  Remember suitable hash function for any signature algorithm.
  Introduce macros and functions to characterize certain ciphersuites.
2017-06-06 18:14:57 +02:00
Janos Follath 088ce43ffe Implement optional CA list suppression in Certificate Request
According to RFC5246 the server can indicate the known Certificate
Authorities or can constrain the aurhorisation space by sending a
certificate list. This part of the message is optional and if omitted,
the client may send any certificate in the response.

The previous behaviour of mbed TLS was to always send the name of all the
CAs that are configured as root CAs. In certain cases this might cause
usability and privacy issues for example:
- If the list of the CA names is longer than the peers input buffer then
  the handshake will fail
- If the configured CAs belong to third parties, this message gives away
  information on the relations to these third parties

Therefore we introduce an option to suppress the CA list in the
Certificate Request message.

Providing this feature as a runtime option comes with a little cost in
code size and advantages in maintenance and flexibility.
2017-05-16 10:22:37 +01:00
Hanno Becker 7e5437a972 Remember suitable hash function for any signature algorithm.
This commit changes `ssl_parse_signature_algorithms_ext` to remember
one suitable ( := supported by client and by our config ) hash
algorithm per signature algorithm.

It also modifies the ciphersuite checking function
`ssl_ciphersuite_match` to refuse a suite if there
is no suitable hash algorithm.

Finally, it adds the corresponding entry to the ChangeLog.
2017-05-15 11:50:11 +01:00
Gilles Peskine 92e4426169 SSLv3: when refusing renegotiation, stop processing
Before the code was sending a fatal alert but then processing any
further data from the peer.

Internal reference: IOTSSL-1384
2017-05-10 17:31:13 +02:00
Gilles Peskine c94f7352fa Ignore failures when sending fatal alerts
In many places in TLS handling, some code detects a fatal error, sends
a fatal alert message, and returns to the caller. If sending the alert
fails, then return the error that triggered the alert, rather than
overriding the return status. This effectively causes alert sending
failures to be ignored. Formerly the code was inconsistently sometimes
doing one, sometimes the other.

In general ignoring the alert is the right thing: what matters to the
caller is the original error. A typical alert failure is that the
connection is already closed.

One case which remains not handled correctly is if the alert remains
in the output buffer (WANT_WRITE). Then it won't be sent, or will be
truncated. We'd need to either delay the application error or record
the write buffering notice; to be done later.
2017-05-10 17:31:02 +02:00
Gilles Peskine 8498cb3687 Cleaned up double variable declaration 2017-05-10 15:39:40 +02:00
Gilles Peskine 064a85ca48 Code portability fix 2017-05-10 10:46:40 +02:00
Gilles Peskine 1cc8e3472a Send TLS alerts in many more cases
The TLS client and server code was usually closing the connection in
case of a fatal error without sending an alert. This commit adds
alerts in many cases.

Added one test case to detect that we send the alert, where a server
complains that the client's certificate is from an unknown CA (case
tracked internally as IOTSSL-1330).
2017-05-03 16:28:34 +02:00
Gilles Peskine f982852bf0 Balanced braces across preprocessor conditionals
This is a cosmetic improvement to ease source code navigation only.
2017-05-03 12:28:43 +02:00
Janos Follath 182013faf4 Prevent SLOTH attacks 2017-02-15 23:29:32 +02:00
Andres AG 2196c7f81c Fix renegotiation at incorrect times in DTLS
Fix an incorrect condition in ssl_check_ctr_renegotiate() that compared
64 bits of record counter instead of 48 bits as described in RFC 6347
Section 4.3.1. This would cause the function's return value to be
occasionally incorrect and the renegotiation routines to be triggered
at unexpected times.
2017-02-15 23:26:35 +02:00
Brian J Murray 1903fb312f Clarify Comments and Fix Typos (#651)
Fixes many typos, and errors in comments.

* Clarifies many comments
* Grammar correction in config.pl help text
* Removed comment about MBEDTLS_X509_EXT_NS_CERT_TYPE.
* Comment typo fix (Dont => Don't)
* Comment typo fix (assure => ensure)
* Comment typo fix (byes => bytes)
* Added citation for quoted standard
* Comment typo fix (one complement => 1's complement)

The is some debate about whether to prefer "one's complement",  "ones'
complement", or "1's complement".  The more recent RFCs related to TLS
(RFC 6347,  RFC 4347, etc) use " 1's complement", so I followed that
convention.

* Added missing ")" in comment
* Comment alignment
* Incorrect comment after #endif
2017-02-15 09:08:26 +00:00
Simon Butcher 8e00410402 Merge fix for AEAD Random IVs 2016-10-14 00:48:33 +01:00
Simon Butcher 99000142cb Merge fix for IE Certificate Compatibility 2016-10-13 17:21:01 +01:00
Janos Follath 23bdca0d63 Fix an x509 compatibility issue
Certificates with unsupported algorithms in the certificate chain
prevented verification even if a certificate before the unsupported
ones was already trusted.

We change the behaviour to ignoring every certificate with unknown
(unsupported) signature algorithm oid when parsing the certificate
chain received from the peer.
2016-10-07 16:12:50 +01:00
Simon Butcher b5b6af2663 Puts platform time abstraction into its own header
Separates platform time abstraction into it's own header from the
general platform abstraction as both depend on different build options.
(MBEDTLS_PLATFORM_C vs MBEDTLS_HAVE_TIME)
2016-07-13 14:46:18 +01:00
Janos Follath 8eb64132da Address issues find by manual coverity scan. 2016-06-09 13:41:28 +01:00
Nicholas Wilson f0021645b0 Refactor slightly to silence a clang-analyze warning
Since the buffer is used in a few places, it seems Clang isn't clever
enough to realise that the first byte is never touched.  So, even though
the function has a correct null check for ssl->handshake, Clang
complains.  Pulling the handshake type out into its own variable is
enough for Clang's analysis to kick in though.
2016-05-23 14:29:28 +01:00
SimonB d5800b7761 Abstracts away time()/stdlib.h into platform
Substitutes time() into a configurable platform interface to allow it to be
easily substituted.
2016-04-26 14:49:59 +01:00
Simon Butcher e846b5128f Use the SSL IO and time callback typedefs consistently
The callback typedefs defined for mbedtls_ssl_set_bio() and
mbedtls_ssl_set_timer_cb() were not used consistently where the callbacks were
referenced in structures or in code.
2016-03-09 19:32:09 +00:00
Manuel Pégourié-Gonnard 214a84889c Update note about hardcoded verify_data_length 2016-03-09 19:32:08 +00:00
Manuel Pégourié-Gonnard 967994a05e Remove unused code.
After the record contents are decompressed, in_len is no longer
accessed directly, only in_msglen is accessed. in_len is only read by
ssl_parse_record_header() which happens before ssl_prepare_record_contents().

This is also made clear by the fact that in_len is not touched after
decrypting anyway, so if it was accessed after that it would be wrong unless
decryption is used - as this is not the case, it show in_len is not accessed.
2016-03-09 19:32:08 +00:00
Janos Follath 4ae5c294a4 Add Changelog entry and improve coding style 2016-02-10 11:27:43 +00:00
Janos Follath ca214b9aaf Updated relevant #ifdef 2016-02-09 16:53:08 +00:00
Simon Butcher c97b697939 Fix for unused variable warning 2015-12-27 23:48:17 +00:00
Simon Butcher 4c2bfdbff6 Merge 'iotssl-558-md5-tls-sigs-restricted' 2015-12-23 18:33:54 +00:00
Simon Butcher 207990dcf5 Added description of change to the Changelog
Also clarified some comments following review.
2015-12-16 01:51:30 +00:00
Manuel Pégourié-Gonnard 47229c7cbb Disable MD5 in handshake signatures by default 2015-12-04 15:02:56 +01:00
Manuel Pégourié-Gonnard e2e25e7427 DTLS: avoid dropping too many records
When the peer retransmits a flight with many record in the same datagram, and
we already saw one of the records in that datagram, we used to drop the whole
datagram, resulting in interoperability failure (spurious handshake timeouts,
due to ignoring record retransmitted by the peer) with some implementations
(issues with Chrome were reported).

So in those cases, we want to only drop the current record, and look at the
following records (if any) in the same datagram. OTOH, this is not something
we always want to do, as sometime the header of the current record is not
reliable enough.

This commit introduces a new return code for ssl_parse_header() that allows to
distinguish if we should drop only the current record or the whole datagram,
and uses it in mbedtls_ssl_read_record()

fixes #345
2015-12-03 16:13:17 +01:00
Manuel Pégourié-Gonnard bd3639852c Merge branch 'iotssl-519-asn1write-overflows-restricted' into development-restricted
* iotssl-519-asn1write-overflows-restricted:
  Fix other int casts in bounds checking
  Fix other occurrences of same bounds check issue
  Fix potential buffer overflow in asn1write
2015-11-02 11:07:30 +09:00
Manuel Pégourié-Gonnard f8b2442e2f Merge branch 'iotssl-517-double-free-restricted' into development-restricted
* iotssl-517-double-free-restricted:
  Fix potential double-free in ssl_conf_psk()
2015-11-02 11:03:32 +09:00
Simon Butcher 204606238c Merge branch 'development' into misc 2015-10-27 16:57:34 +00:00
Simon Butcher 459a950f31 Fixed typo in comment 2015-10-27 16:09:03 +00:00
Simon Butcher 62aab15085 Merge branch 'development' into iotssl-513-alerts 2015-10-27 16:05:34 +00:00
Simon Butcher e357a64355 Merge pull request #328 from ARMmbed/iotssl-461-ecjpake-finalization
Iotssl 461 ecjpake finalization
2015-10-27 00:08:31 +00:00
Manuel Pégourié-Gonnard fbdf06c1a4 Fix handling of non-fatal alerts
fixes #308
2015-10-23 13:11:31 +02:00
Manuel Pégourié-Gonnard e5f3072aed Fix #ifdef inconsistency
fixes #310

Actually all key exchanges that use a certificate use signatures too, and
there is no key exchange that uses signatures but no cert, so merge those two
flags.
2015-10-23 08:40:23 +02:00
Manuel Pégourié-Gonnard bc5e508855 Fix other int casts in bounds checking
Not a security issue as here we know the buffer is large enough (unless
something else if badly wrong in the code), and the value cast to int is less
than 2^16 (again, unless issues elsewhere).

Still changing to a more correct check as a matter of principle
2015-10-21 12:51:16 +02:00
Manuel Pégourié-Gonnard cdea97c1c3 Remove useless code
closes #321
2015-10-20 20:06:36 +02:00