Commit graph

8922 commits

Author SHA1 Message Date
Jaeden Amero d247762a8d Merge remote-tracking branch 'origin/pr/2460' into development 2019-02-22 12:52:51 +00:00
Hanno Becker 85fd913950 Fix typo in check_config.h 2019-02-22 12:50:35 +00:00
Jaeden Amero caca307f42 crypto: Update submodule to Mbed Crypto 1.0.0d6 2019-02-22 10:48:48 +00:00
Jaeden Amero 415620c1f2 Merge remote-tracking branch 'origin/pr/2105' into development
Additional work done as part of merge:
    - Run ./tests/scripts/check-generated-files.sh and check in the
      resulting changes to programs/ssl/query_config.c
2019-02-22 10:33:15 +00:00
Jaeden Amero 461bd3dcca Merge remote-tracking branch 'origin/pr/2454' into development 2019-02-22 10:32:44 +00:00
Jaeden Amero 9f47f82218 Merge remote-tracking branch 'origin/pr/2391' into development 2019-02-22 10:32:44 +00:00
Jaeden Amero 8963b0311c Merge remote-tracking branch 'origin/pr/2411' into development 2019-02-22 10:32:44 +00:00
Jaeden Amero d9516b508a Merge remote-tracking branch 'origin/pr/2407' into development 2019-02-22 10:32:44 +00:00
Jaeden Amero 0ae63f7bc3 Merge remote-tracking branch 'origin/pr/2383' into development 2019-02-22 10:32:43 +00:00
Hanno Becker 241b524964 Disable restartable ECC in full config PSA test in all.sh 2019-02-22 10:26:30 +00:00
Hanno Becker 1ce51e4dc3 Forbid setting MBEDTLS_ECP_RESTARTABLE and MBEDTLS_USE_PSA_CRYPTO_C
Restartable ECC isn't supported in PSA yet.
2019-02-22 10:25:47 +00:00
Jaeden Amero 3497323f79 Initialize PSA Crypto operation contexts
It is now required to initialize PSA Crypto operation contexts before
calling psa_*_setup(). Otherwise, one gets a PSA_ERROR_BAD_STATE error.
2019-02-20 10:58:55 +00:00
Gilles Peskine adc558e289 Fix test data missing some fake-random input
The test function pkcs1_rsaes_v15_encrypt gets its fake-random input
for padding from a test parameter. In one test case, the parameter was
too short, causing a fallback to rand(). The reference output depends
on this random input, so the test data was correct only for a platform
with one particular rand() implementation. Supply sufficient
fake-random input so that rand() isn't called.
2019-02-19 18:33:57 +01:00
Gilles Peskine 85a6dd40ba Add tests for RSA encrypt/decrypt with NULL for empty message 2019-02-19 18:33:57 +01:00
k-stachowiak cddbd01e4f Reword changelog entry 2019-02-19 12:40:34 +01:00
k-stachowiak 6409724059 Update change log 2019-02-19 12:25:24 +01:00
Hanno Becker 135baef1bd Define maximum EC public key length depending on enabled curves 2019-02-18 17:04:24 +00:00
Hanno Becker 28f78440d8 Grep for debug output witnessing use of PSA in ECDHE ssl-opt.sh 2019-02-18 16:47:50 +00:00
Hanno Becker 4af484e29a Regenerate VS2010 project file 2019-02-18 16:42:02 +00:00
Hanno Becker 3b7c4a0ff0 Regenerate VisualStudio project file 2019-02-18 16:42:02 +00:00
Hanno Becker 354e248d81 Add ssl-opt.sh tests for PSA-based ECDH with various ECC curves 2019-02-18 16:42:02 +00:00
Hanno Becker 0a94a64bbd Add debugging output to confirm that PSA was used for ECDHE 2019-02-18 16:42:02 +00:00
Hanno Becker c14a3bb5a6 Make variable in ssl_write_client_key_exchange() more descriptive 2019-02-18 16:42:02 +00:00
Hanno Becker 4a63ed421c Implement ClientKeyExchange writing in PSA-based ECDHE suites
- Populate the ECDH private key slot with a fresh private EC key
  designated for the correct algorithm.
- Export the public part of the ECDH private key from PSA and
  reformat it to suite the format of the ClientKeyExchange message.
- Perform the PSA-based ECDH key agreement and store the result
  as the premaster secret for the connection.
2019-02-18 16:42:01 +00:00
Hanno Becker bb89e2727f Implement ServerKeyExchange parsing for PSA-based ECDHE suites
- Reformat the server's ECDH public key to make it suitable
  for the PSA key agreement API. Currently, the key agreement
  API needs a full SubjectPublicKeyInfo structure, while the
  TLS ServerKeyExchange message only contains a ECPoint structure.
2019-02-18 16:42:01 +00:00
Hanno Becker df51dbe17f Add fields for PSA-based ECDHE to handshake structure
This is the first in a series of commits adding client-side
support for PSA-based ECDHE.

Previously, the state of an ECDHE key agreement was maintained
in the field mbedtls_ssl_handshake_params::ecdh_ctx, of type
::mbedtls_ecdh_context and manipulated through the ECDH API.

The ECDH API will be superseeded by the PSA Crypto API for key
agreement, which needs the following data:
(a) A raw buffer holding the public part of the key agreement
    received from our peer.
(b) A key slot holding the private part of the key agreement.
(c) The algorithm to use.
The commit adds fields to ::mbedtls_ssl_handshake_params
representing these three inputs to PSA-based key agreement.

Specifically, it adds a field for the key slot holding the
ECDH private key, a field for the EC curve identifier, and
a buffer holding the peer's public key.

Note: Storing the peer's public key buffer is slightly
inefficient, as one could perform the ECDH computation
as soon as the peer sends its public key, either working
with in-place or using a stack-buffer to reformat the
public key before passing it to PSA. This optimization
is left for a later commit.
2019-02-18 16:41:55 +00:00
Hanno Becker f75f912c31 Add functions to psa_util module to convert EC public keys 2019-02-18 16:37:12 +00:00
k-stachowiak 17a38d3a74 Reenable GnuTLS next based tests 2019-02-18 15:29:56 +01:00
k-stachowiak 28cb6fbd47 Unbump version to 0.0.0 2019-02-18 12:01:03 +01:00
Manuel Pégourié-Gonnard 9c99dc862c
Merge pull request #2395 from ARMmbed/development-psa-merged-dev-8e76332
Merge updated development-psa into development
2019-02-18 11:55:54 +01:00
Andrzej Kurek 064128c1b0 Update submodule pointer to commit 0574e6a of mbed-crypto/development 2019-02-15 10:20:05 -05:00
Andrzej Kurek 1b20be59e2 Write documentation for TEST_ASSERT 2019-02-15 10:13:35 -05:00
Jaeden Amero 2d7926674d psa_utils: Make fallthrough clear for compilers
Silence a compiler warning about implicit fallthrough by using a comment
format the compiler understand to mean that the fallthrough is
intentional.

  In file included from library/cipher.c:63:0:
  include/mbedtls/psa_util.h: In function ‘mbedtls_psa_translate_cipher_mode’:
  include/mbedtls/psa_util.h:91:15: error: this statement may fall through [-Werror=implicit-fallthrough=]
               if( taglen == 0 )
                 ^
  include/mbedtls/psa_util.h:94:9: note: here
           default:
           ^~~~~~~
  cc1: all warnings being treated as errors

  $ gcc --version
  gcc (Ubuntu 7.3.0-27ubuntu1~18.04) 7.3.0
  Copyright (C) 2017 Free Software Foundation, Inc.
  This is free software; see the source for copying conditions.  There is NO
  warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
2019-02-15 09:50:38 -05:00
Andrzej Kurek eb5087126f all.sh: add a seedfile generation step
When using PSA with MBEDTLS_ENTROPY_NV_SEED, some test suites
require the seed file for PSA initialization, which was normally generated
later, when entropy tests were run. This change creates an initial seedfile
in all.sh.
2019-02-14 07:18:59 -05:00
Andrzej Kurek c058773798 Update submodule pointer to dea02cc8b 2019-02-14 05:28:21 -05:00
Jaeden Amero 3ea2687053 psa: Initialize crypto in tests that need to
Don't unconditionally enable PSA Crypto for all tests. Only enable it in
tests that require it. This allows crypto tests to check that
psa_crypto_init() fails when it is supposed to fail, since we want to
perform some action in a test, and then call psa_crypto_init() and check
the result without it having been called previously.
2019-02-13 07:34:54 -05:00
Ron Eldor 3467dcf452 Use certificates from data_files and refer them
Use the server certificate from `data_files` folder, for formality,
and refer to the source, for easier reproduction.
2019-02-12 15:30:26 +02:00
Andres Amaya Garcia 54efcb7ff0 Update query config with VSNPRINTF macro 2019-02-11 21:55:09 +00:00
Andrzej Kurek 08b807a161 Update crypto submodule version to commit f54597144874
Use mbed-crypto repository in gitmodules file
2019-02-11 08:45:41 -05:00
Andrzej Kurek f389629556 Move a restartable ecp context to a conditional compilation block
This was an unused variable when compiling with parameter validation
but without ecp_restartable
2019-02-11 05:15:54 -05:00
Jaeden Amero 6f7703df3a rsa: Enable use of zero-length null output
Enable handling of zero-length null output in PKCS1 v1.5 decryption.
Prevent undefined behavior by avoiding a memcpy() to zero-length null
output buffers.
2019-02-11 03:39:51 -05:00
Gilles Peskine 004f87b98d RSA encryption: accept input=NULL if ilen=0
In mbedtls_rsa_rsaes_oaep_encrypt and
mbedtls_rsa_rsaes_pkcs1_v15_encrypt, if the input length is 0 (which
is unusual and mostly useless, but permitted) then it is fine for the
input pointer to be NULL. Don't return an error in this case.

When `input` is NULL, `memcpy( p, input, ilen )` has undefined behavior
even if `ilen` is zero. So skip the `memcpy` call in this case.
Likewise, in `mbedtls_rsa_rsaes_oaep_decrypt`, skip the `memcpy` call if
`*olen` is zero.
2019-02-11 03:39:21 -05:00
Peter Kolbus 5da93f8903 Update ChangeLog 2019-02-09 10:21:57 -06:00
Hanno Becker 46f34d0ac0 Fix style issue and wording 2019-02-08 14:26:41 +00:00
Hanno Becker c1e18bdf06 Fix memory leak 2019-02-08 14:26:41 +00:00
Hanno Becker e2734e2be4 Improve formatting of ssl_parse_certificate_chain() 2019-02-08 14:26:41 +00:00
Hanno Becker 84879e32ef Add compile-time guards around helper routine 2019-02-08 14:26:41 +00:00
Hanno Becker bd9d51d969 Adapt ChangeLog 2019-02-08 14:26:41 +00:00
Hanno Becker def9bdc152 Don't store the peer CRT chain twice during renegotiation
Context: During a handshake, the SSL/TLS handshake logic constructs
an instance of ::mbedtls_ssl_session representing the SSL session
being established. This structure contains information such as the
session's master secret, the peer certificate, or the session ticket
issues by the server (if applicable).
During a renegotiation, the new session is constructed aside the existing
one and destroys and replaces the latter only when the renegotiation is
complete. While conceptually clear, this means that during the renegotiation,
large pieces of information such as the peer's CRT or the session ticket
exist twice in memory, even though the original versions are removed
eventually.

This commit removes the simultaneous presence of two peer CRT chains
in memory during renegotiation, in the following way:
- Unlike in the case of SessionTickets handled in the previous commit,
  we cannot simply free the peer's CRT chain from the previous handshake
  before parsing the new one, as we need to verify that the peer's end-CRT
  hasn't changed to mitigate the 'Triple Handshake Attack'.
- Instead, we perform a binary comparison of the original peer end-CRT
  with the one presented during renegotiation, and if it succeeds, we
  avoid re-parsing CRT by moving the corresponding CRT pointer from the
  old to the new session structure.
- The remaining CRTs in the peer's chain are not affected by the triple
  handshake attack protection, and for them we may employ the canonical
  approach of freeing them before parsing the remainder of the new chain.

Note that this commit intends to not change any observable behavior
of the stack. In particular:
- The peer's CRT chain is still verified during renegotiation.
- The tail of the peer's CRT chain may change during renegotiation.
2019-02-08 14:26:41 +00:00
Hanno Becker b2964cbe14 SSL/TLS client: Remove old session ticket on renegotiation
Context: During a handshake, the SSL/TLS handshake logic constructs
an instance of ::mbedtls_ssl_session representing the SSL session
being established. This structure contains information such as the
session's master secret, the peer certificate, or the session ticket
issues by the server (if applicable).

During a renegotiation, the new session is constructed aside the existing
one and destroys and replaces the latter only when the renegotiation is
complete. While conceptually clear, this means that during the renegotiation,
large pieces of information such as the peer's CRT or the session ticket
exist twice in memory, even though the original versions are removed
eventually.

This commit starts removing this memory inefficiency by freeing the old
session's SessionTicket before the one for the new session is allocated.
2019-02-08 14:26:41 +00:00