[ec] Match built-in curves on EC_GROUP_new_from_ecparameters
Description
-----------
Upon `EC_GROUP_new_from_ecparameters()` check if the parameters match any
of the built-in curves. If that is the case, return a new
`EC_GROUP_new_by_curve_name()` object instead of the explicit parameters
`EC_GROUP`.
This affects all users of `EC_GROUP_new_from_ecparameters()`:
- direct calls to `EC_GROUP_new_from_ecparameters()`
- direct calls to `EC_GROUP_new_from_ecpkparameters()` with an explicit
parameters argument
- ASN.1 parsing of explicit parameters keys (as it eventually
ends up calling `EC_GROUP_new_from_ecpkparameters()`)
A parsed explicit parameter key will still be marked with the
`OPENSSL_EC_EXPLICIT_CURVE` ASN.1 flag on load, so, unless
programmatically forced otherwise, if the key is eventually serialized
the output will still be encoded with explicit parameters, even if
internally it is treated as a named curve `EC_GROUP`.
Before this change, creating any `EC_GROUP` object using
`EC_GROUP_new_from_ecparameters()`, yielded an object associated with
the default generic `EC_METHOD`, but this was never guaranteed in the
documentation.
After this commit, users of the library that intentionally want to
create an `EC_GROUP` object using a specific `EC_METHOD` can still
explicitly call `EC_GROUP_new(foo_method)` and then manually set the
curve parameters using `EC_GROUP_set_*()`.
Motivation
----------
This has obvious performance benefits for the built-in curves with
specialized `EC_METHOD`s and subtle but important security benefits:
- the specialized methods have better security hardening than the
generic implementations
- optional fields in the parameter encoding, like the `cofactor`, cannot
be leveraged by an attacker to force execution of the less secure
code-paths for single point scalar multiplication
- in general, this leads to reducing the attack surface
Check the manuscript at https://arxiv.org/abs/1909.01785 for an in depth
analysis of the issues related to this commit.
It should be noted that `libssl` does not allow to negotiate explicit
parameters (as per RFC 8422), so it is not directly affected by the
consequences of using explicit parameters that this commit fixes.
On the other hand, we detected external applications and users in the
wild that use explicit parameters by default (and sometimes using 0 as
the cofactor value, which is technically not a valid value per the
specification, but is tolerated by parsers for wider compatibility given
that the field is optional).
These external users of `libcrypto` are exposed to these vulnerabilities
and their security will benefit from this commit.
Related commits
---------------
While this commit is beneficial for users using built-in curves and
explicit parameters encoding for serialized keys, commit b783beeadf6b80bc431e6f3230b5d5585c87ef87 (and its equivalents for the
1.0.2, 1.1.0 and 1.1.1 stable branches) fixes the consequences of the
invalid cofactor values more in general also for other curves
(CVE-2019-1547).
The following list covers commits in `master` that are related to the
vulnerabilities presented in the manuscript motivating this commit:
- d2baf88c43 [crypto/rsa] Set the constant-time flag in multi-prime RSA too
- 311e903d84 [crypto/asn1] Fix multiple SCA vulnerabilities during RSA key validation.
- b783beeadf [crypto/ec] for ECC parameters with NULL or zero cofactor, compute it
- 724339ff44 Fix SCA vulnerability when using PVK and MSBLOB key formats
Note that the PRs that contributed the listed commits also include other
commits providing related testing and documentation, in addition to
links to PRs and commits backporting the fixes to the 1.0.2, 1.1.0 and
1.1.1 branches.
Responsible Disclosure
----------------------
This and the other issues presented in https://arxiv.org/abs/1909.01785
were reported by Cesar Pereida García, Sohaib ul Hassan, Nicola Tuveri,
Iaroslav Gridin, Alejandro Cabrera Aldaya and Billy Bob Brumley from the
NISEC group at Tampere University, FINLAND.
The OpenSSL Security Team evaluated the security risk for this
vulnerability as low, and encouraged to propose fixes using public Pull
Requests.
Configure: clang: move -Wno-unknown-warning-option to the front
While gcc ignores unknown options of the type '-Wno-xxx', clang by default issues
a warning [-Wunknown-warning-option] (see [3]), which together with '-Werror'
causes the build to fail. This turned out to be a problem on the 1.0.2 stable branch
in the case of the '-Wextended-offsetof' option, which was removed in version 6.0.0,
but needs to be kept here in order to support older clang versions, too (see #9446).
Incidentally, master and 1.1.1 branch already contained the -Wno-unknown-warning-option
option. Due to its special role and its importance, this commit adds an explaining
commit message and moves the option to the front.
[extended tests]
Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9447)
Billy Brumley [Fri, 6 Sep 2019 14:26:40 +0000 (17:26 +0300)]
CHANGES entry: for ECC parameters with NULL or zero cofactor, compute it
This is a forward port from https://github.com/openssl/openssl/pull/9781
of the CHANGES entry for the functionality added in
https://github.com/openssl/openssl/pull/9727
Billy Brumley [Fri, 6 Sep 2019 14:26:08 +0000 (17:26 +0300)]
[test] computing ECC cofactors: regression test
This is a forward port from
https://github.com/openssl/openssl/pull/9781
for the test logic introduced by
https://github.com/openssl/openssl/pull/9727
As @mattcaswell commented
(https://github.com/openssl/openssl/pull/9781#discussion_r321621541):
> These `TEST_true` calls should be `!TEST_false` because we are
> *expecting* a failure.
> The difference is that the test framework will print error details if
> the test doesn't give the expected answer.
> So by using `TEST_true` instead of `!TEST_false` we'll get error
> details printed, but the test will succeed anyway.
Nicola Tuveri [Thu, 1 Aug 2019 23:08:34 +0000 (02:08 +0300)]
Make BN_num_bits() consttime upon BN_FLG_CONSTTIME
This issue was partially addressed by commit 972c87dfc7e765bd28a4964519c362f0d3a58ca4, which hardened its callee
BN_num_bits_word() to avoid leaking the most-significant word of its
argument via branching and memory access pattern.
The commit message also reported:
> There are a few places where BN_num_bits is called on an input where
> the bit length is also secret. This does *not* fully resolve those
> cases as we still only look at the top word.
BN_num_bits() is called directly or indirectly (e.g., through
BN_num_bytes() or BN_bn2binpad() ) in various parts of the `crypto/ec`
code, notably in all the currently supported implementations of scalar
multiplication (in the generic path through ec_scalar_mul_ladder() as
well as in dedicated methods like ecp_nistp{224,256,521}.c and
ecp_nistz256.c).
Under the right conditions, a motivated SCA attacker could retrieve the
secret bitlength of a secret nonce through this vulnerability,
potentially leading, ultimately, to recover a long-term secret key.
With this commit, exclusively for BIGNUMs that are flagged with
BN_FLG_CONSTTIME, instead of accessing only bn->top, all the limbs of
the BIGNUM are accessed up to bn->dmax and bitwise masking is used to
avoid branching.
Memory access pattern still leaks bn->dmax, the size of the lazily
allocated buffer for representing the BIGNUM, which is inevitable with
the current BIGNUM architecture: reading past bn->dmax would be an
out-of-bound read.
As such, it's the caller responsibility to ensure that bn->dmax does not
leak secret information, by explicitly expanding the internal BIGNUM
buffer to a public value sufficient to avoid any lazy reallocation
while manipulating it: this should be already done at the top level
alongside setting the BN_FLG_CONSTTIME.
Thanks to David Schrammel and Samuel Weiser for reporting this issue
through responsible disclosure.
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/9511)
Nicola Tuveri [Thu, 1 Aug 2019 22:33:05 +0000 (01:33 +0300)]
Fix a SCA leak using BN_bn2bin()
BN_bn2bin() is not constant-time and leaks the number of bits in the
processed BIGNUM.
The specialized methods in ecp_nistp224.c, ecp_nistp256.c and
ecp_nistp521.c internally used BN_bn2bin() to convert scalars into the
internal fixed length representation.
This can leak during ECDSA/ECDH key generation or handling the nonce
while generating an ECDSA signature, when using these implementations.
The amount and risk of leaked information useful for a SCA attack
varies for each of the three curves, as it depends mainly on the
ratio between the bitlength of the curve subgroup order (governing the
size of the secret nonce/key) and the limb size for the internal BIGNUM
representation (which depends on the compilation target architecture).
To fix this, we replace BN_bn2bin() with BN_bn2binpad(), bounding the
output length to the width of the internal representation buffer: this
length is public.
Internally the final implementation of both BN_bn2binpad() and
BN_bn2bin() already has masking in place to avoid leaking bn->top
through memory access patterns.
Memory access pattern still leaks bn->dmax, the size of the lazily
allocated buffer for representing the BIGNUM, which is inevitable with
the current BIGNUM architecture: reading past bn->dmax would be an
out-of-bound read.
As such, it's the caller responsibility to ensure that bn->dmax does not
leak secret information, by explicitly expanding the internal BIGNUM
buffer to a public value sufficient to avoid any lazy reallocation
while manipulating it: this is already done at the top level alongside
setting the BN_FLG_CONSTTIME.
Finally, the internal implementation of BN_bn2binpad() indirectly calls
BN_num_bits() via BN_num_bytes(): the current implementation of
BN_num_bits() can leak information to a SCA attacker, and is addressed
in the next commit.
Thanks to David Schrammel and Samuel Weiser for reporting this issue
through responsible disclosure.
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/9511)
[crypto/asn1] Fix multiple SCA vulnerabilities during RSA key validation.
This commit addresses multiple side-channel vulnerabilities present
during RSA key validation.
Private key parameters are re-computed using variable-time functions.
This issue was discovered and reported by the NISEC group at TAU Finland.
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9779)
Richard Levitte [Fri, 30 Aug 2019 14:54:47 +0000 (16:54 +0200)]
Deal with BUF_MEM_grow ambiguity
BUF_MEM_grow() returns the passed length, but also zero on error. If
the passed length was zero, an extra check to see if a returned zero
was an error or not is needed.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9662)
Richard Levitte [Fri, 30 Aug 2019 13:36:20 +0000 (15:36 +0200)]
More KDF cleanup
The EVP_KDF_ definitions are no longer needed, and neither is
EVP_get_kdfbyname()
test/evp_kdf_test.c tried to use a EVP_get_kdfbyname() that was rewritten
to use EVP_KDF_fetch() without ever freeing the resulting KDF method.
It's better to refactor the test to use EVP_KDF_fetch directly.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9662)
Richard Levitte [Fri, 30 Aug 2019 12:32:55 +0000 (14:32 +0200)]
crypto/evp/pkey_kdf.c: Redo parameter processing
Undo the caching scheme, pass through most controls as parameters, except
for SEED and INFO, where we keep supporting adding data through additional
ctrl calls by collecting the data, and only passing it to the EVP_KDF
before calling its derive function.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9662)
Matt Caswell [Thu, 5 Sep 2019 15:43:57 +0000 (16:43 +0100)]
Don't send a status_request extension in a CertificateRequest message
If a TLSv1.3 server configured to respond to the status_request extension
also attempted to send a CertificateRequest then it was incorrectly
inserting a non zero length status_request extension into that message.
The TLSv1.3 RFC does allow that extension in that message but it must
always be zero length.
In fact we should not be sending the extension at all in that message
because we don't support it.
Fixes #9767
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9780)
Viktor Dukhovni [Sun, 9 Dec 2018 22:10:29 +0000 (17:10 -0500)]
Undeprecate OpenSSL_version_num and OPENSSL_VERSION_NUMBER
The OpenSSL_version_num() function returns at runtime the
OPENSSL_VERSION_NUMBER of the compiled OpenSSL library. This is a
used and useful interface, and should not (at least yet) be
deprecated, we just introduced the new versioning schema, it seems
too early to deprecate the old.
Reviewed-by: Tim Hudson <tjh@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/7853)
Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9776)
Billy Brumley [Mon, 2 Sep 2019 12:02:30 +0000 (15:02 +0300)]
[crypto/ec] for ECC parameters with NULL or zero cofactor, compute it
The cofactor argument to EC_GROUP_set_generator is optional, and SCA mitigations for ECC currently use it. So the library currently falls back to very old SCA-vulnerable code if the cofactor is not present.
This PR allows EC_GROUP_set_generator to compute the cofactor for all curves of cryptographic interest. Steering scalar multiplication to more SCA-robust code.
This issue affects persisted private keys in explicit parameter form, where the (optional) cofactor field is zero or absent.
It also affects curves not built-in to the library, but constructed programatically with explicit parameters, then calling EC_GROUP_set_generator with a nonsensical value (NULL, zero).
The very old scalar multiplication code is known to be vulnerable to local uarch attacks, outside of the OpenSSL threat model. New results suggest the code path is also vulnerable to traditional wall clock timing attacks.
CVE-2019-1547
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9827)
Rich Salz [Sun, 18 Aug 2019 13:04:17 +0000 (09:04 -0400)]
Make failed messages easier to find
Now that we use travis_terminate, we can make the status messages
simpler to find, and we don't need the "OK" output.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9707)
Bernd Edlinger [Mon, 19 Aug 2019 15:12:22 +0000 (17:12 +0200)]
Fix error handling in x509_lu.c
Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9639)
Rich Salz [Thu, 29 Aug 2019 16:12:17 +0000 (12:12 -0400)]
Handle the renamed command POD files in find-doc-nits
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9729)
Billy Brawner [Wed, 28 Aug 2019 00:07:17 +0000 (17:07 -0700)]
Suppress 'No server certificate CA names sent' message
Fixes #9080
Signed-off-by: Billy Brawner <billy@wbrawner.com> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9710)
Richard Levitte [Tue, 3 Sep 2019 16:11:49 +0000 (18:11 +0200)]
New function EVP_CIPHER_free()
This function re-implements EVP_CIPHER_meth_free(), but has a name that
isn't encumbered by legacy EVP_CIPHER construction functionality.
We also refactor most of EVP_CIPHER_meth_new() into an internal
evp_cipher_new() that's used when creating fetched methods.
EVP_CIPHER_meth_new() and EVP_CIPHER_meth_free() are rewritten in terms of
evp_cipher_new() and EVP_CIPHER_free(). This means that at any time, we can
deprecate all the EVP_CIPHER_meth_ functions with no harmful consequence.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9758)
Richard Levitte [Tue, 3 Sep 2019 15:47:13 +0000 (17:47 +0200)]
New function EVP_MD_free()
This function re-implements EVP_MD_meth_free(), but has a name that
isn't encumbered by legacy EVP_MD construction functionality.
We also refactor most of EVP_MD_meth_new() into an internal
evp_md_new() that's used when creating fetched methods.
EVP_MD_meth_new() and EVP_MD_meth_free() are rewritten in terms of
evp_md_new() and EVP_MD_free(). This means that at any time, we can
deprecate all the EVP_MD_meth_ functions with no harmful consequence.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9758)
Richard Levitte [Tue, 3 Sep 2019 13:10:43 +0000 (15:10 +0200)]
test/errtest.c: more conditions for checking __FILE__ and __LINE__
When at least one of OPENSSL_NO_ERR or OPENSSL_NO_FILENAMES is
defined, __FILE__ and __LINE__ are not saved with the error record.
This test only checked OPENSSL_NO_FILENAMES. Now fixed.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9755)
Richard Levitte [Fri, 23 Aug 2019 12:03:28 +0000 (14:03 +0200)]
Refactor how KEYMGMT methods get associated with other methods
KEYMGMT methods were attached to other methods after those were fully
created and registered, thereby creating a potential data race, if two
threads tried to create the exact same method at the same time.
Instead of this, we change the method creating function to take an
extra data parameter, passed all the way from the public fetching
function. In the case of EVP_KEYEXCH, we pass all the necessary data
that evp_keyexch_from_dispatch() needs to be able to fetch the
appropriate KEYMGMT method on the fly.
Fixes #9592
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9678)
Richard Levitte [Sun, 1 Sep 2019 08:58:19 +0000 (10:58 +0200)]
Diverse ERR fixes
1. There are still references to the removed ERR_put_func_error().
2. ERR_put_error() is deprecated as off version 3.0, so should
ERR_PUT_error().
3. 'no-err' didn't affect what was passed to ERR_set_debug().
Fixes #9522
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9747)
Richard Levitte [Wed, 28 Aug 2019 07:13:21 +0000 (09:13 +0200)]
Rework the documentation of our individual MAC implementations
We now describe parameters instead of controls.
Also, since macros like EVP_MAC_CMAC do not exist any more, we rename
the pod files from EVP_MAC_{algo}.pod to EVP_MAC-{algo}.pod. This
allows getting the documentation like this:
man EVP_MAC CMAC
[skip ci]
Fixes #9709
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9713)
Bernd Edlinger [Sat, 24 Aug 2019 09:28:19 +0000 (11:28 +0200)]
Remove extern declarations of OPENSSL_ia32cap_P
Use the header file internal/cryptlib.h instead.
Remove checks for OPENSSL_NO_ASM and I386_ONLY
in cryptlib.c, to match the checks in other
places where OPENSSL_ia32cap_P is used and
assumed to be initialized.
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/9688)
Matt Caswell [Thu, 29 Aug 2019 16:15:16 +0000 (17:15 +0100)]
Fix pkeyutl -verifyrecover
When performing a pkeyutl -verifyrecover operation the input file is not
a hash - it is the signature itself. Therefore don't do the check to make
sure it looks like a hash.
Fixes #9658
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9731)
Pauli [Thu, 29 Aug 2019 21:38:58 +0000 (07:38 +1000)]
Don't include the DEVRANDOM being seeded logic on Android.
It lacks exposure of the `shm*` functions and should prefer the GETRANDOM
source.
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/9735)
Matt Caswell [Wed, 28 Aug 2019 13:31:39 +0000 (14:31 +0100)]
Fix no-ec
A test in sslapitest.c was failing in a no-ec build because we were using
an EC based ciphersuite. That particular test doesn't require EC
specifically, so we swap to a non EC based ciphersuite.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9716)
Matt Caswell [Wed, 28 Aug 2019 16:30:14 +0000 (17:30 +0100)]
Update fuzz README.md
Building with enable-fuzz-afl has always required no-shared. We now also
need no-module for a successful build. Therefore update the README
accordingly.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9721)
Matt Caswell [Wed, 14 Aug 2019 17:09:28 +0000 (18:09 +0100)]
Fix data races in EVP_CIPHER_fetch and EVP_MD_fetch
Don't modify the cipher/md we just fetched - it could be shared by multiple
threads.
Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9590)
projects / openssl / log