This is about a timing leak in the topmost limb
of the internal result of RSA_private_decrypt,
before the padding check.
There are in fact at least three bugs together that
caused the timing leak:
First and probably most important is the fact that
the blinding did not use the constant time code path
at all when the RSA object was used for a private
decrypt, due to the fact that the Montgomery context
rsa->_method_mod_n was not set up early enough in
rsa_ossl_private_decrypt, when BN_BLINDING_create_param
needed it, and that was persisted as blinding->m_ctx,
although the RSA object creates the Montgomery context
just a bit later.
Then the infamous bn_correct_top was used on the
secret value right after the blinding was removed.
And finally the function BN_bn2binpad did not use
the constant-time code path since the BN_FLG_CONSTTIME
was not set on the secret value.
In order to address the first problem, this patch
makes sure that the rsa->_method_mod_n is initialized
right before the blinding context.
And to fix the second problem, we add a new utility
function bn_correct_top_consttime, a const-time
variant of bn_correct_top.
Together with the fact, that BN_bn2binpad is already
constant time if the flag BN_FLG_CONSTTIME is set,
this should eliminate the timing oracle completely.
In addition the no-asm variant may also have
branches that depend on secret values, because the last
invocation of bn_sub_words in bn_from_montgomery_word
had branches when the function is compiled by certain
gcc compiler versions, due to the clumsy coding style.
So additionally this patch stream-lined the no-asm
C-code in order to avoid branches where possible and
improve the resulting code quality.
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20281)
This reverts commit b1892d21f8.
Except for the moving derive_kdk to a separate function.
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20281)
Function EVP_PKEY_get_bn_param() uses temporary buffer (on stack or
heap allocated) to store serialized bignum, but after deserializing it
into BIGNUM*, the buffer is not erased and may contain sensitive data.
This change makes sure the buffer is erased if it was successfully
filled before. Unfortunately, it does not distinguish between public and
private key components, and will always erase the buffer.
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20639)
Calling OPENSSL_init_crypto from inside a RUN_ONCE seems like a bad idea.
This is especially bad if OPENSSL_init_crypto can recursively end up
attempting to call the RUN_ONCE that we're already inside.
The initialisation in OPENSSL_init_crypto is already "run once" protected.
There is no need to protect it "twice".
Fixes#20653
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20662)
(cherry picked from commit a9745427cd)
We also test SSL_is_dtls(), SSL_is_tls() and SSL_is_quic().
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20650)
We have the existing functions SSL_version(), SSL_get_version() and
SSL_is_dtls(). We extend the first two to return something sensible when
using QUIC. We additionally provide the new functions SSL_is_tls() and
SSL_is_quic() to provide a mechanism to figure out what protocol we are
using.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20650)
Discovered during coverage testing.
Remove unneccesary check when using ossl_dh_get0_params() and
ossl_dsa_get0_params(). These point to addresses and can not fail
for any existing calls.
Make dsa keygen tests only available in the FIPS module - as they are
not used in the default provider.
Change ossl_ffc_set_digest() to return void as it cannot fail.
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20359)
Fixes a "Incompatible function pointer types" error.
In prior versions, this was a warning, but is now an error.
CLA: trivial
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20654)
Related to #20268
DSA_sign() assumes that the signature passed in is related to DSA_size().
If q is negative then DSA_size() actually fails and returns 0.
A test that tries to allocate the signature buffer using DSA_size() and then
pass it to DSA_sign() will then either.
(1) Have a signature buffer of NULL. In this case it was leaking data
returned via i2d_DSA_SIG.
(2) Cause a seg fault because we created a buffer that was not large
enough to hold the signature. As it already checked zero we also now
check for negative values also.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/20553)
CLA: trivial
There is an incorrect null pointer check and this ccommit resolves it.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/20646)
Similiar to the issue found in PR #20553 for DSA_sign().
ECDSA_sign() leaked memory if the signature was NULL
when i2d_ECDSA_SIG was called.
Note that this does not affect the higher level EVP
functions as they correctly handle NULL.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/20554)
On NetBSD the device node is essentially always present, but these
days it's mostly not useful except for testing the kernel crypto
stack since the cost of CPU crypto has gone down much faster than the
cost of I/O to external crypto engines. So on most systems it is
disabled in a way that makes open fail with ENXIO. NetBSD has had
this warning for ENXIO patched away for years.
CLA: trivial
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20636)
We had a test for a handshake record appearing before epoch change, and
a test for an app data record appearing before Finished - but not one for
the app data record appearing before epoch change.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20628)
It is possible that DTLS records are received out of order such that
records from the next epoch arrive before we have finished processing the
current epoch. We are supposed to buffer such records but for some reason
we only did that for handshake and alert records. This is incorrect since
it is perfectly possible for app data records to arrive early too.
Fixes#20597
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20628)
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20629)
The s_client command mode just looks at the first letter on a line
and if it is one that maps to a "command" then that command is executed
and everything else on the line is ignored.
This means it is impossible to send anything over s_client that starts with
one of the "magic" letters unless you switch off command processing
entirely.
With the forthcoming quic support that will be added to s_client we expect
a significant new number of commands being added (for example to send fin
on a stream, or to list or swap between different streams, etc). This will
use up more "letters" and reduces the usability of s_client.
This PR adds a new "advanced" command processing mode. Commands are enclosed
in braces {}, and can appear anywhere in the text typed into s_client. A
brace can be escaped by sending a double {{. This gives much more
flexibility without reducing the usability of s_client. Commands can also
have "arguments" which also extends the scope of what they can be used for.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20566)
