You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

11489 lines
486 KiB

Changes between 1.0.2e and 1.1.0 [xx XXX xxxx]
*) Remove support for all 40 and 56 bit ciphers. This includes all the export
ciphers who are no longer supported and drops support the ephemeral RSA key
exchange. The LOW ciphers currently doesn't have any ciphers in it.
[Kurt Roeckx]
*) Make EVP_MD_CTX, EVP_MD and HMAC_CTX opaque. For HMAC_CTX, the
following constructors and destructors were added:
HMAC_CTX *HMAC_CTX_new(void);
void HMAC_CTX_free(HMAC_CTX *ctx);
For EVP_MD, a complete API to create, fill and destroy such
methods has been added. See EVP_MD_meth_new(3) for
Additional changes:
1) HMAC_CTX_cleanup() and EVP_MD_CTX_cleanup() were removed,
HMAC_CTX_init() and EVP_MD_CTX_init() should be called instead
to reinitialise and already created structure. Also,
HMAC_CTX_init() and EVP_MD_CTX_init() now return 0 for failure
and 1 for success (they previously had the return type void).
2) For consistency with the majority of our object creators and
destructors, EVP_MD_CTX_(create|destroy) were renamed to
EVP_MD_CTX_(new|free). The old names are retained as macros
for deprecated builds.
[Richard Levitte]
*) Added ASYNC support. Libcrypto now includes the async sub-library to enable
cryptographic operations to be performed asynchronously as long as an
asynchronous capable engine is used. See the ASYNC_start_job() man page for
further details. Libssl has also had this capability integrated with the
introduction of the new mode SSL_MODE_ASYNC and associated error
SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
pages. This work was developed in partnership with Intel Corp.
[Matt Caswell]
*) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
always enabled now. If you want to disable the support you should
exclude it using the list of supported ciphers.
[Kurt Roeckx]
*) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
SSL_{CTX_}set1_curves() which can set a list.
[Kurt Roeckx]
*) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
curve you want to support using SSL_{CTX_}set1_curves().
[Kurt Roeckx]
*) State machine rewrite. The state machine code has been significantly
refactored in order to remove much duplication of code and solve issues
with the old code (see ssl/statem/README for further details). This change
does have some associated API changes. Notably the SSL_state() function
has been removed and replaced by SSL_get_state which now returns an
"OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
altogether. The previous handshake states defined in ssl.h and ssl3.h have
also been removed.
[Matt Caswell]
*) All instances of the string "ssleay" in the public API were replaced
with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
Some error codes related to internal RSA_eay API's were renamed.
[Rich Salz]
*) The demo files in crypto/threads were moved to demo/threads.
[Rich Salz]
*) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron and sureware.
[Matt Caswell]
*) New ASN.1 embed macro.
New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
structure is not allocated: it is part of the parent. That is instead of
FOO *x;
it must be:
FOO x;
This reduces memory fragmentation and make it impossible to accidentally
set a mandatory field to NULL.
This currently only works for some fields specifically a SEQUENCE, CHOICE,
or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
[Steve Henson]
*) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
[Emilia Käsper]
*) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
DES and RC4 ciphersuites.
[Matt Caswell]
*) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
This changes the decoding behaviour for some invalid messages,
though the change is mostly in the more lenient direction, and
legacy behaviour is preserved as much as possible.
[Emilia Käsper]
*) Fix no-stdio build.
[ David Woodhouse <> and also
Ivan Nestlerode <> ]
*) New testing framework
The testing framework has been largely rewritten and is now using
perl and the perl modules Test::Harness and an extended variant of
Test::More called OpenSSL::Test to do its work. All test scripts in
test/ have been rewritten into test recipes, and all direct calls to
executables in test/Makefile have become individual recipes using the
simplified testing OpenSSL::Test::Simple.
For documentation on our testing modules, do:
perldoc test/testlib/OpenSSL/Test/
perldoc test/testlib/OpenSSL/
[Richard Levitte]
*) In DSA_generate_parameters_ex, if the provided seed is too short,
return an error
[Rich Salz and Ismo Puustinen <>]
*) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
from RFC4279, RFC4785, RFC5487, RFC5489.
Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
original RSA_PSK patch.
[Steve Henson]
Remove support for SSL3_FLAGS_DELAY_CLIENT_FINISHED This flag was not set anywhere within the codebase (only read). It could only be set by an app reaching directly into s->s3->flags and setting it directly. However that method became impossible when libssl was opaquified. Even in 1.0.2/1.0.1 if an app set the flag directly it is only relevant to ssl3_connect(), which calls SSL_clear() during initialisation that clears any flag settings. Therefore it could take effect if the app set the flag after the handshake has started but before it completed. It seems quite unlikely that any apps really do this (especially as it is completely undocumented). The purpose of the flag is suppress flushing of the write bio on the client side at the end of the handshake after the client has written the Finished message whilst resuming a session. This enables the client to send application data as part of the same flight as the Finished message. This flag also controls the setting of a second flag SSL3_FLAGS_POP_BUFFER. There is an interesting comment in the code about this second flag in the implementation of ssl3_write: /* This is an experimental flag that sends the * last handshake message in the same packet as the first * use data - used to see if it helps the TCP protocol during * session-id reuse */ It seems the experiment did not work because as far as I can tell nothing is using this code. The above comment has been in the code since SSLeay. This commit removes support for SSL3_FLAGS_DELAY_CLIENT_FINISHED, as well as the associated SSL3_FLAGS_POP_BUFFER. Reviewed-by: Rich Salz <>
8 years ago
*) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
era flag was never set throughout the codebase (only read). Also removed
SSL3_FLAGS_POP_BUFFER which was only used if
[Matt Caswell]
*) Changed the default name options in the "ca", "crl", "req" and "x509"
to be "oneline" instead of "compat".
[Richard Levitte]
*) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
not aware of clients that still exhibit this bug, and the workaround
hasn't been working properly for a while.
[Emilia Käsper]
*) The return type of BIO_number_read() and BIO_number_written() as well as
the corresponding num_read and num_write members in the BIO structure has
changed from unsigned long to uint64_t. On platforms where an unsigned
long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
[Matt Caswell]
*) Given the pervasive nature of TLS extensions it is inadvisable to run
OpenSSL without support for them. It also means that maintaining
the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
[Matt Caswell]
*) Removed support for the two export grade static DH ciphersuites
EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
were newly added (along with a number of other static DH ciphersuites) to
1.0.2. However the two export ones have *never* worked since they were
introduced. It seems strange in any case to be adding new export
ciphersuites, and given "logjam" it also does not seem correct to fix them.
[Matt Caswell]
*) Version negotiation has been rewritten. In particular SSLv23_method(),
SSLv23_client_method() and SSLv23_server_method() have been deprecated,
and turned into macros which simply call the new preferred function names
TLS_method(), TLS_client_method() and TLS_server_method(). All new code
should use the new names instead. Also as part of this change the ssl23.h
header file has been removed.
[Matt Caswell]
*) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
code and the associated standard is no longer considered fit-for-purpose.
[Matt Caswell]
*) RT2547 was closed. When generating a private key, try to make the
output file readable only by the owner. This behavior change might
be noticeable when interacting with other software.
*) Documented all exdata functions. Added CRYPTO_free_ex_index.
Added a test.
[Rich Salz]
*) Added HTTP GET support to the ocsp command.
[Rich Salz]
*) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
[Matt Caswell]
Deprecate RAND_pseudo_bytes The justification for RAND_pseudo_bytes is somewhat dubious, and the reality is that it is frequently being misused. RAND_bytes and RAND_pseudo_bytes in the default implementation both end up calling ssleay_rand_bytes. Both may return -1 in an error condition. If there is insufficient entropy then both will return 0, but RAND_bytes will additionally add an error to the error queue. They both return 1 on success. Therefore the fundamental difference between the two is that one will add an error to the error queue with insufficient entory whilst the other will not. Frequently there are constructions of this form: if(RAND_pseudo_bytes(...) <= 1) goto err; In the above form insufficient entropy is treated as an error anyway, so RAND_bytes is probably the better form to use. This form is also seen: if(!RAND_pseudo_bytes(...)) goto err; This is technically not correct at all since a -1 return value is incorrectly handled - but this form will also treat insufficient entropy as an error. Within libssl it is required that you have correctly seeded your entropy pool and so there seems little benefit in using RAND_pseudo_bytes. Similarly in libcrypto many operations also require a correctly seeded entropy pool and so in most interesting cases you would be better off using RAND_bytes anyway. There is a significant risk of RAND_pseudo_bytes being incorrectly used in scenarios where security can be compromised by insufficient entropy. If you are not using the default implementation, then most engines use the same function to implement RAND_bytes and RAND_pseudo_bytes in any case. Given its misuse, limited benefit, and potential to compromise security, RAND_pseudo_bytes has been deprecated. Reviewed-by: Richard Levitte <>
8 years ago
*) Added support for TLS extended master secret from
draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
initial patch which was a great help during development.
[Steve Henson]
*) All libssl internal structures have been removed from the public header
files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
now redundant). Users should not attempt to access internal structures
directly. Instead they should use the provided API functions.
[Matt Caswell]
10 years ago
*) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
Access to deprecated functions can be re-enabled by running config with
"enable-deprecated". In addition applications wishing to use deprecated
functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
will, by default, disable some transitive includes that previously existed
in the header files (e.g. ec.h will no longer, by default, include bn.h)
[Matt Caswell]
*) Added support for OCB mode. OpenSSL has been granted a patent license
compatible with the OpenSSL license for use of OCB. Details are available
at Support
for OCB can be removed by calling config with no-ocb.
[Matt Caswell]
*) SSLv2 support has been removed. It still supports receiving a SSLv2
compatible client hello.
[Kurt Roeckx]
*) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
done while fixing the error code for the key-too-small case.
[Annie Yousar <>]
*) has been removmed; use instead.
[Rich Salz]
*) Removed old DES API.
[Rich Salz]
*) Remove various unsupported platforms:
Sony NEWS4
Sinix/ReliantUNIX RM400
16-bit platforms such as WIN16
[Rich Salz]
*) Clean up OPENSSL_NO_xxx #define's
Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
Remove MS_STATIC; it's a relic from platforms <32 bits.
[Rich Salz]
*) Cleaned up dead code
Remove all but one '#ifdef undef' which is to be looked at.
[Rich Salz]
*) Clean up calling of xxx_free routines.
Just like free(), fix most of the xxx_free routines to accept
NULL. Remove the non-null checks from callers. Save much code.
[Rich Salz]
*) Add secure heap for storage of private keys (when possible).
Add BIO_s_secmem(), CBIGNUM, etc.
Contributed by Akamai Technologies under our Corporate CLA.
[Rich Salz]
*) Experimental support for a new, fast, unbiased prime candidate generator,
bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
[Felix Laurie von Massenbach <>]
*) New output format NSS in the sess_id command line tool. This allows
exporting the session id and the master key in NSS keylog format.
[Martin Kaiser <>]
*) Harmonize version and its documentation. -f flag is used to display
compilation flags.
[mancha <>]
*) Fix eckey_priv_encode so it immediately returns an error upon a failure
in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
[mancha <>]
*) Fix some double frees. These are not thought to be exploitable.
[mancha <>]
*) A missing bounds check in the handling of the TLS heartbeat extension
can be used to reveal up to 64k of memory to a connected client or
Thanks for Neel Mehta of Google Security for discovering this bug and to
Adam Langley <> and Bodo Moeller <> for
preparing the fix (CVE-2014-0160)
[Adam Langley, Bodo Moeller]
*) Fix for the attack described in the paper "Recovering OpenSSL
ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
by Yuval Yarom and Naomi Benger. Details can be obtained from:
Thanks to Yuval Yarom and Naomi Benger for discovering this
flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
[Yuval Yarom and Naomi Benger]
*) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
this fixes a limitation in previous versions of OpenSSL.
[Steve Henson]
*) Experimental encrypt-then-mac support.
Experimental support for encrypt then mac from
To enable it set the appropriate extension number (0x42 for the test
server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
For non-compliant peers (i.e. just about everything) this should have no
[Steve Henson]
*) Add EVP support for key wrapping algorithms, to avoid problems with
existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
algorithms and include tests cases.
[Steve Henson]
10 years ago
*) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
enveloped data.
[Steve Henson]
*) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
MGF1 digest and OAEP label.
[Steve Henson]
*) Make openssl verify return errors.
[Chris Palmer <> and Ben Laurie]
*) New function ASN1_TIME_diff to calculate the difference between two
ASN1_TIME structures or one structure and the current time.
[Steve Henson]
*) Update fips_test_suite to support multiple command line options. New
test to induce all self test errors in sequence and check expected
[Steve Henson]
*) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
sign or verify all in one operation.
[Steve Henson]
*) Add fips_algvs: a multicall fips utility incorporating all the algorithm
test programs and fips_test_suite. Includes functionality to parse
the minimal script output of directly.
[Steve Henson]
*) Add authorisation parameter to FIPS_module_mode_set().
[Steve Henson]
*) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
[Steve Henson]
*) Use separate DRBG fields for internal and external flags. New function
FIPS_drbg_health_check() to perform on demand health checking. Add
generation tests to fips_test_suite with reduced health check interval to
demonstrate periodic health checking. Add "nodh" option to
fips_test_suite to skip very slow DH test.
[Steve Henson]
*) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
based on NID.
[Steve Henson]
*) More extensive health check for DRBG checking many more failure modes.
New function FIPS_selftest_drbg_all() to handle every possible DRBG
combination: call this in fips_test_suite.
[Steve Henson]
*) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
and POST to handle Dual EC cases.
[Steve Henson]
*) Add support for canonical generation of DSA parameter 'g'. See
FIPS 186-3 A.2.3.
*) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
POST to handle HMAC cases.
[Steve Henson]
*) Add functions FIPS_module_version() and FIPS_module_version_text()
to return numerical and string versions of the FIPS module number.
[Steve Henson]
*) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
outside the validated module in the FIPS capable OpenSSL.
[Steve Henson]
*) Minor change to DRBG entropy callback semantics. In some cases
there is no multiple of the block length between min_len and
max_len. Allow the callback to return more than max_len bytes
of entropy but discard any extra: it is the callback's responsibility
to ensure that the extra data discarded does not impact the
requested amount of entropy.
[Steve Henson]
*) Add PRNG security strength checks to RSA, DSA and ECDSA using
information in FIPS186-3, SP800-57 and SP800-131A.
[Steve Henson]
*) CCM support via EVP. Interface is very similar to GCM case except we
must supply all data in one chunk (i.e. no update, final) and the
message length must be supplied if AAD is used. Add algorithm test
[Steve Henson]
*) Initial version of POST overhaul. Add POST callback to allow the status
of POST to be monitored and/or failures induced. Modify fips_test_suite
to use callback. Always run all selftests even if one fails.
[Steve Henson]
*) XTS support including algorithm test driver in the fips_gcmtest program.
Note: this does increase the maximum key length from 32 to 64 bytes but
there should be no binary compatibility issues as existing applications
will never use XTS mode.
[Steve Henson]
*) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
performs algorithm blocking for unapproved PRNG types. Also do not
set PRNG type in FIPS_mode_set(): leave this to the application.
Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
12 years ago
the standard OpenSSL PRNG: set additional data to a date time vector.
[Steve Henson]
*) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
This shouldn't present any incompatibility problems because applications
shouldn't be using these directly and any that are will need to rethink
anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
[Steve Henson]
*) Extensive self tests and health checking required by SP800-90 DRBG.
Remove strength parameter from FIPS_drbg_instantiate and always
instantiate at maximum supported strength.
[Steve Henson]
*) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
[Steve Henson]
*) New algorithm test program fips_dhvs to handle DH primitives only testing.
[Steve Henson]
*) New function DH_compute_key_padded() to compute a DH key and pad with
leading zeroes if needed: this complies with SP800-56A et al.
[Steve Henson]
*) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
anything, incomplete, subject to change and largely untested at present.
[Steve Henson]
*) Modify fipscanisteronly build option to only build the necessary object
files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
[Steve Henson]
*) Add experimental option FIPSSYMS to give all symbols in
fipscanister.o and FIPS or fips prefix. This will avoid
conflicts with future versions of OpenSSL. Add perl script
util/ to preprocess assembly language source files
and rename any affected symbols.
[Steve Henson]
*) Add selftest checks and algorithm block of non-fips algorithms in
FIPS mode. Remove DES2 from selftests.
[Steve Henson]
*) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
return internal method without any ENGINE dependencies. Add new
tiny fips sign and verify functions.
[Steve Henson]
*) New build option no-ec2m to disable characteristic 2 code.
[Steve Henson]
*) New build option "fipscanisteronly". This only builds fipscanister.o
and (currently) associated fips utilities. Uses the file Makefile.fips
instead of as the prototype.
[Steve Henson]
*) Add some FIPS mode restrictions to GCM. Add internal IV generator.
Update fips_gcmtest to use IV generator.
[Steve Henson]
*) Initial, experimental EVP support for AES-GCM. AAD can be input by
setting output buffer to NULL. The *Final function must be
called although it will not retrieve any additional data. The tag
can be set or retrieved with a ctrl. The IV length is by default 12
bytes (96 bits) but can be set to an alternative value. If the IV
length exceeds the maximum IV length (currently 16 bytes) it cannot be
set before the key.
[Steve Henson]
*) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
underlying do_cipher function handles all cipher semantics itself
including padding and finalisation. This is useful if (for example)
an ENGINE cipher handles block padding itself. The behaviour of
do_cipher is subtly changed if this flag is set: the return value
is the number of characters written to the output buffer (zero is
no longer an error code) or a negative error code. Also if the
input buffer is NULL and length 0 finalisation should be performed.
[Steve Henson]
*) If a candidate issuer certificate is already part of the constructed
path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
[Steve Henson]
*) Improve forward-security support: add functions
void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
for use by SSL/TLS servers; the callback function will be called whenever a
new session is created, and gets to decide whether the session may be
cached to make it resumable (return 0) or not (return 1). (As by the
SSL/TLS protocol specifications, the session_id sent by the server will be
empty to indicate that the session is not resumable; also, the server will
not generate RFC 4507 (RFC 5077) session tickets.)
A simple reasonable callback implementation is to return is_forward_secure.
This parameter will be set to 1 or 0 depending on the ciphersuite selected
by the SSL/TLS server library, indicating whether it can provide forward
[Emilia Käsper <> (Google)]
*) New -verify_name option in command line utilities to set verification
parameters by name.
[Steve Henson]
Add CMAC pkey methods.
[Steve Henson]
*) Experimental renegotiation in s_server -www mode. If the client
browses /reneg connection is renegotiated. If /renegcert it is
renegotiated requesting a certificate.
[Steve Henson]
*) Add an "external" session cache for debugging purposes to s_server. This
should help trace issues which normally are only apparent in deployed
multi-process servers.
[Steve Henson]
*) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
return value is ignored. NB. The functions RAND_add(), RAND_seed(),
BIO_set_cipher() and some obscure PEM functions were changed so they
can now return an error. The RAND changes required a change to the
RAND_METHOD structure.
[Steve Henson]
*) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
a gcc attribute to warn if the result of a function is ignored. This
is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
whose return value is often ignored.
[Steve Henson]
Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
*) BN_mod_exp may produce incorrect results on x86_64
There is a carry propagating bug in the x86_64 Montgomery squaring
procedure. No EC algorithms are affected. Analysis suggests that attacks
against RSA and DSA as a result of this defect would be very difficult to
perform and are not believed likely. Attacks against DH are considered just
feasible (although very difficult) because most of the work necessary to
deduce information about a private key may be performed offline. The amount
of resources required for such an attack would be very significant and
likely only accessible to a limited number of attackers. An attacker would
additionally need online access to an unpatched system using the target
private key in a scenario with persistent DH parameters and a private
key that is shared between multiple clients. For example this can occur by
default in OpenSSL DHE based SSL/TLS ciphersuites.
This issue was reported to OpenSSL by Hanno Böck.
[Andy Polyakov]
*) Certificate verify crash with missing PSS parameter
The signature verification routines will crash with a NULL pointer
dereference if presented with an ASN.1 signature using the RSA PSS
algorithm and absent mask generation function parameter. Since these
routines are used to verify certificate signature algorithms this can be
used to crash any certificate verification operation and exploited in a
DoS attack. Any application which performs certificate verification is
vulnerable including OpenSSL clients and servers which enable client
This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
[Stephen Henson]
*) X509_ATTRIBUTE memory leak
When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
memory. This structure is used by the PKCS#7 and CMS routines so any
application which reads PKCS#7 or CMS data from untrusted sources is
affected. SSL/TLS is not affected.
This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
[Stephen Henson]
*) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
This changes the decoding behaviour for some invalid messages,
though the change is mostly in the more lenient direction, and
legacy behaviour is preserved as much as possible.
[Emilia Käsper]
*) In DSA_generate_parameters_ex, if the provided seed is too short,
return an error
[Rich Salz and Ismo Puustinen <>]
Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
*) Alternate chains certificate forgery
During certificate verfification, OpenSSL will attempt to find an
alternative certificate chain if the first attempt to build such a chain
fails. An error in the implementation of this logic can mean that an
attacker could cause certain checks on untrusted certificates to be
bypassed, such as the CA flag, enabling them to use a valid leaf
certificate to act as a CA and "issue" an invalid certificate.
This issue was reported to OpenSSL by Adam Langley/David Benjamin
[Matt Caswell]
Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
*) Fix HMAC ABI incompatibility. The previous version introduced an ABI
incompatibility in the handling of HMAC. The previous ABI has now been
[Matt Caswell]
Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
*) Malformed ECParameters causes infinite loop
When processing an ECParameters structure OpenSSL enters an infinite loop
if the curve specified is over a specially malformed binary polynomial
This can be used to perform denial of service against any
system which processes public keys, certificate requests or
certificates. This includes TLS clients and TLS servers with
client authentication enabled.
This issue was reported to OpenSSL by Joseph Barr-Pixton.
[Andy Polyakov]
*) Exploitable out-of-bounds read in X509_cmp_time
X509_cmp_time does not properly check the length of the ASN1_TIME
string and can read a few bytes out of bounds. In addition,
X509_cmp_time accepts an arbitrary number of fractional seconds in the
time string.
An attacker can use this to craft malformed certificates and CRLs of
various sizes and potentially cause a segmentation fault, resulting in
a DoS on applications that verify certificates or CRLs. TLS clients
that verify CRLs are affected. TLS clients and servers with client
authentication enabled may be affected if they use custom verification
This issue was reported to OpenSSL by Robert Swiecki (Google), and
independently by Hanno Böck.
[Emilia Käsper]
*) PKCS7 crash with missing EnvelopedContent
The PKCS#7 parsing code does not handle missing inner EncryptedContent
correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
with missing content and trigger a NULL pointer dereference on parsing.
Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
structures from untrusted sources are affected. OpenSSL clients and
servers are not affected.
This issue was reported to OpenSSL by Michal Zalewski (Google).
[Emilia Käsper]
*) CMS verify infinite loop with unknown hash function
When verifying a signedData message the CMS code can enter an infinite loop
if presented with an unknown hash function OID. This can be used to perform
denial of service against any system which verifies signedData messages using
the CMS code.
This issue was reported to OpenSSL by Johannes Bauer.
[Stephen Henson]
*) Race condition handling NewSessionTicket
If a NewSessionTicket is received by a multi-threaded client when attempting to
reuse a previous ticket then a race condition can occur potentially leading to
a double free of the ticket data.
[Matt Caswell]
*) Only support 256-bit or stronger elliptic curves with the
'ecdh_auto' setting (server) or by default (client). Of supported
curves, prefer P-256 (both).
[Emilia Kasper]
Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
*) ClientHello sigalgs DoS fix
If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
invalid signature algorithms extension a NULL pointer dereference will
occur. This can be exploited in a DoS attack against the server.
This issue was was reported to OpenSSL by David Ramos of Stanford
[Stephen Henson and Matt Caswell]
*) Multiblock corrupted pointer fix
OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
feature only applies on 64 bit x86 architecture platforms that support AES
NI instructions. A defect in the implementation of "multiblock" can cause
OpenSSL's internal write buffer to become incorrectly set to NULL when
using non-blocking IO. Typically, when the user application is using a
socket BIO for writing, this will only result in a failed connection.
However if some other BIO is used then it is likely that a segmentation
fault will be triggered, thus enabling a potential DoS attack.
This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
[Matt Caswell]
*) Segmentation fault in DTLSv1_listen fix
The DTLSv1_listen function is intended to be stateless and processes the
initial ClientHello from many peers. It is common for user code to loop
over the call to DTLSv1_listen until a valid ClientHello is received with
an associated cookie. A defect in the implementation of DTLSv1_listen means
that state is preserved in the SSL object from one invocation to the next
that can lead to a segmentation fault. Errors processing the initial
ClientHello can trigger this scenario. An example of such an error could be
that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
This issue was reported to OpenSSL by Per Allansson.
[Matt Caswell]
*) Segmentation fault in ASN1_TYPE_cmp fix
The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
certificate signature algorithm consistency this can be used to crash any
certificate verification operation and exploited in a DoS attack. Any
application which performs certificate verification is vulnerable including
OpenSSL clients and servers which enable client authentication.
[Stephen Henson]
*) Segmentation fault for invalid PSS parameters fix
The signature verification routines will crash with a NULL pointer
dereference if presented with an ASN.1 signature using the RSA PSS
algorithm and invalid parameters. Since these routines are used to verify
certificate signature algorithms this can be used to crash any
certificate verification operation and exploited in a DoS attack. Any
application which performs certificate verification is vulnerable including
OpenSSL clients and servers which enable client authentication.
This issue was was reported to OpenSSL by Brian Carpenter.
[Stephen Henson]
*) ASN.1 structure reuse memory corruption fix
Reusing a structure in ASN.1 parsing may allow an attacker to cause
memory corruption via an invalid write. Such reuse is and has been
strongly discouraged and is believed to be rare.
Applications that parse structures containing CHOICE or ANY DEFINED BY
components may be affected. Certificate parsing (d2i_X509 and related
functions) are however not affected. OpenSSL clients and servers are
not affected.
[Stephen Henson]
*) PKCS7 NULL pointer dereferences fix
The PKCS#7 parsing code does not handle missing outer ContentInfo
correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
missing content and trigger a NULL pointer dereference on parsing.
Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
otherwise parse PKCS#7 structures from untrusted sources are
affected. OpenSSL clients and servers are not affected.
This issue was reported to OpenSSL by Michal Zalewski (Google).
[Emilia Käsper]
*) DoS via reachable assert in SSLv2 servers fix
A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
servers that both support SSLv2 and enable export cipher suites by sending
a specially crafted SSLv2 CLIENT-MASTER-KEY message.
This issue was discovered by Sean Burford (Google) and Emilia Käsper
(OpenSSL development team).
[Emilia Käsper]
*) Empty CKE with client auth and DHE fix
If client auth is used then a server can seg fault in the event of a DHE
ciphersuite being selected and a zero length ClientKeyExchange message
being sent by the client. This could be exploited in a DoS attack.
[Matt Caswell]
*) Handshake with unseeded PRNG fix
Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
with an unseeded PRNG. The conditions are:
- The client is on a platform where the PRNG has not been seeded
automatically, and the user has not seeded manually
- A protocol specific client method version has been used (i.e. not
- A ciphersuite is used that does not require additional random data from
the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
If the handshake succeeds then the client random that has been used will
have been generated from a PRNG with insufficient entropy and therefore the
output may be predictable.
For example using the following command with an unseeded openssl will
succeed on an unpatched platform:
openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
[Matt Caswell]
*) Use After Free following d2i_ECPrivatekey error fix
A malformed EC private key file consumed via the d2i_ECPrivateKey function
could cause a use after free condition. This, in turn, could cause a double
free in several private key parsing functions (such as d2i_PrivateKey
or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
for applications that receive EC private keys from untrusted
sources. This scenario is considered rare.
This issue was discovered by the BoringSSL project and fixed in their
commit 517073cd4b.
[Matt Caswell]
*) X509_to_X509_REQ NULL pointer deref fix
The function X509_to_X509_REQ will crash with a NULL pointer dereference if
the certificate key is invalid. This function is rarely used in practice.
This issue was discovered by Brian Carpenter.
[Stephen Henson]
*) Removed the export ciphers from the DEFAULT ciphers
[Kurt Roeckx]
Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
*) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
ARMv5 through ARMv8, as opposite to "locking" it to single one.
So far those who have to target multiple plaforms would compromise
and argue that binary targeting say ARMv5 would still execute on
ARMv8. "Universal" build resolves this compromise by providing
near-optimal performance even on newer platforms.
[Andy Polyakov]
*) Accelerated NIST P-256 elliptic curve implementation for x86_64
(other platforms pending).
[Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
*) Add support for the SignedCertificateTimestampList certificate and
OCSP response extensions from RFC6962.
[Rob Stradling]
*) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
for corner cases. (Certain input points at infinity could lead to
bogus results, with non-infinity inputs mapped to infinity too.)
[Bodo Moeller]
*) Initial support for PowerISA 2.0.7, first implemented in POWER8.
This covers AES, SHA256/512 and GHASH. "Initial" means that most
common cases are optimized and there still is room for further
improvements. Vector Permutation AES for Altivec is also added.
[Andy Polyakov]
*) Add support for little-endian ppc64 Linux target.
[Marcelo Cerri (IBM)]
*) Initial support for AMRv8 ISA crypto extensions. This covers AES,
SHA1, SHA256 and GHASH. "Initial" means that most common cases
are optimized and there still is room for further improvements.
Both 32- and 64-bit modes are supported.
[Andy Polyakov, Ard Biesheuvel (Linaro)]
*) Improved ARMv7 NEON support.
[Andy Polyakov]
*) Support for SPARC Architecture 2011 crypto extensions, first
implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
SHA256/512, MD5, GHASH and modular exponentiation.
[Andy Polyakov, David Miller]
*) Accelerated modular exponentiation for Intel processors, a.k.a.