Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6538)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6499)

848113a3 added mitigation for a
side-channel attack. This commit extends approach to all code
paths for consistency.

[It also removes redundant white spaces introduced in last commit.]
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6480)

This module is used only with odd input lengths, i.e. not used in normal
PKI cases, on contemporary processors. The problem was "illuminated" by
fuzzing tests.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6440)

The One&Done attack, which is described in a paper to appear in the
USENIX Security'18 conference, uses EM emanations to recover the values
of the bits that are obtained using BN_is_bit_set while constructing
the value of the window in BN_mod_exp_consttime. The EM signal changes
slightly depending on the value of the bit, and since the lookup of a
bit is surrounded by highly regular execution (constant-time Montgomery
multiplications) the attack is able to isolate the (very brief) part of
the signal that changes depending on the bit. Although the change is
slight, the attack recovers it successfully >90% of the time on several
phones and IoT devices (all with ARM processors with clock rates around
1GHz), so after only one RSA decryption more than 90% of the bits in
d_p and d_q are recovered correctly, which enables rapid recovery of
the full RSA key using an algorithm (also described in the paper) that
modifies the branch-and-prune approach for a situation in which the
exponents' bits are recovered with errors, i.e. where we do not know
a priori which bits are correctly recovered.

The mitigation for the attack is relatively simple - all the bits of
the window are obtained at once, along with other bits so that an
entire integer's worth of bits are obtained together using masking and
shifts, without unnecessarily considering each bit in isolation. This
improves performance somewhat (one call to bn_get_bits is faster than
several calls to BN_is_bit_set), so the attacker now gets one signal
snippet per window (rather than one per bit) in which the signal is
affected by all bits in the integer (rather than just the one bit).
Reviewed-by: NAndy Polyakov <appro@openssl.org>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6276)

Thanks to Guido Vranken and OSSFuzz for finding this issue.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6355)

Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6371)

Experiments have shown that the lookup table used by BN_GF2m_mod_arr
introduces sufficient timing signal to recover the private key for an
attacker with access to cache timing information on the victim's host.
This only affects binary curves (which are less frequently used).

No CVE is considered necessary for this issue.

The fix is to replace the lookup table with an on-the-fly calculation of
the value from the table instead, which can be performed in constant time.

Thanks to Youngjoo Shin for reporting this issue.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6270)

Reviewed-by: NRich Salz <rsalz@openssl.org>
GH: #6208

Reviewed-by: NRichard Levitte <levitte@openssl.org>
Reviewed-by: NAndy Polyakov <appro@openssl.org>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6070)

Calculating BN_mod_inverse where n is 1 (or -1) doesn't make sense. We
should return an error in that case. Instead we were returning a valid
result with value 0.

Fixes #6004
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6119)

Montgomery multiplication post-conditions in some of code paths were
formally non-constant time. Cache access pattern was result-neutral,
but a little bit asymmetric, which might have produced a signal [if
processor reordered load and stores at run-time].
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6141)

Reviewed-by: NRich Salz <rsalz@openssl.org>
Reviewed-by: NMatt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6099)

Reviewed-by: NBernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6085)

Reviewed-by: NAndy Polyakov <appro@openssl.org>
Reviewed-by: NMatt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6009)

Co-authored-by: NNicola Tuveri <nic.tuv@gmail.com>
Co-authored-by: NCesar Pereida Garcia <cesar.pereidagarcia@tut.fi>
Co-authored-by: NSohaib ul Hassan <soh.19.hassan@gmail.com>
Reviewed-by: NAndy Polyakov <appro@openssl.org>
Reviewed-by: NMatt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6009)

Old code replaced in favor of a clearer implementation.
Performances are not penalized.

Updated the copyright end date to 2018.
Reviewed-by: NDavid Benjamin <davidben@google.com>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5963)

Reviewed-by: NMatt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5990)

Almost all *alloc failures now set an error code.
Reviewed-by: NMatthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/5842)

Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5851)

Reviewed-by: NDr. Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
Reviewed-by: NRich Salz <rsalz@openssl.org>

Fixes: #4641
GH: #4665

Thanks to Darovskikh Andrei for for reporting this issue.

Fixes: #5785
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5793)

Based on the description in https://github.com/openssl/openssl/pull/5757,
this re-implements the "allow NULL to be passed" behavior of a number of
xxx_free routines.  I also fixed up some egregious formatting errors
that were nearby.
Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5761)

This removes some code because we cannot trace the original contributor
to get their agreement for the licence change (original commit e03ddfae).

After this change there will be numerous failures in the test cases until
someone rewrites the missing code.

All *_free functions should accept a NULL parameter. After this change
the following *_free functions will fail if a NULL parameter is passed:

BIO_ACCEPT_free()
BIO_CONNECT_free()
BN_BLINDING_free()
BN_CTX_free()
BN_MONT_CTX_free()
BN_RECP_CTX_free()
BUF_MEM_free()
COMP_CTX_free()
ERR_STATE_free()
TXT_DB_free()
X509_STORE_free()
ssl3_free()
ssl_cert_free()
SSL_SESSION_free()
SSL_free()

[skip ci]
Reviewed-by: NBernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5757)

Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5689)

- d2i_PKC8PrivateKey -> d2i_PKCS8PrivateKey
- bechmark -> benchmark
- ciperhsuite -> ciphersuite
- EncyptedPreMasterSecret -> EncryptedPreMasterSecret
Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5570)

It was calling the BN_rand() when it should have call BN_priv_rand()
Reviewed-by: NTim Hudson <tjh@openssl.org>
GH: #5514

Reviewed-by: NRichard Levitte <levitte@openssl.org>

On VMS, 'abort' is really 'decc$abort'
Reviewed-by: NAndy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5357)

The make variables LIB_CFLAGS, DSO_CFLAGS and so on were used in
addition to CFLAGS and so on.  This works without problem on Unix and
Windows, where options with different purposes (such as -D and -I) can
appear anywhere on the command line and get accumulated as they come.
This is not necessarely so on VMS.  For example, macros must all be
collected and given through one /DEFINE, and the same goes for
inclusion directories (/INCLUDE).

So, to harmonize all platforms, we repurpose make variables starting
with LIB_, DSO_ and BIN_ to be all encompassing variables that
collects the corresponding values from CFLAGS, CPPFLAGS, DEFINES,
INCLUDES and so on together with possible config target values
specific for libraries DSOs and programs, and use them instead of the
general ones everywhere.

This will, for example, allow VMS to use the exact same generators for
generated files that go through cpp as all other platforms, something
that has been impossible to do safely before now.
Reviewed-by: NAndy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5357)

Output copyright year depends on any input file(s) and the script.
This is not perfect, but better than what we had.
Also run 'make update'
Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5350)

Reviewed-by: NRichard Levitte <levitte@openssl.org>

Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5267)

BN_from_montgomery_word doesn't have a constant memory access pattern.
Replace the pointer trick with a constant-time select. There is, of
course, still the bn_correct_top leak pervasive in BIGNUM itself.

See also https://boringssl-review.googlesource.com/22904 from BoringSSL.
Reviewed-by: NAndy Polyakov <appro@openssl.org>
Reviewed-by: NKurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/5228)

The exponent here is one of d, dmp1, or dmq1 for RSA. This value and its
bit length are both secret. The only public upper bound is the bit width
of the corresponding modulus (RSA n, p, and q, respectively).

Although BN_num_bits is constant-time (sort of; see bn_correct_top notes
in preceding patch), this does not fix the root problem, which is that
the windows are based on the minimal bit width, not the upper bound. We
could use BN_num_bits(m), but BN_mod_exp_mont_consttime is public API
and may be called with larger exponents. Instead, use all top*BN_BITS2
bits in the BIGNUM. This is still sensitive to the long-standing
bn_correct_top leak, but we need to fix that regardless.

This may cause us to do a handful of extra multiplications for RSA keys
which are just above a whole number of words, but that is not a standard
RSA key size.
Reviewed-by: NPaul Dale <paul.dale@oracle.com>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5154)

(This patch was written by Andy Polyakov. I only wrote the commit
message. Mistakes in the analysis are my fault.)

BN_num_bits, by way of BN_num_bits_word, currently leaks the
most-significant word of its argument via branching and memory access
pattern.

BN_num_bits is called on RSA prime factors in various places. These have
public bit lengths, but all bits beyond the high bit are secret. This
fully resolves those cases.

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. Today, that is guaranteed to be
non-zero, but only because of the long-standing bn_correct_top timing
leak. Once that is fixed, a constant-time BN_num_bits on such inputs
must count bits on each word.

Instead, those cases should not call BN_num_bits at all. In particular,
BN_mod_exp_mont_consttime uses the exponent bit width to pick windows,
but it should be using the maximum bit width. The next patch will fix
this.

Thanks to Dinghao Wu, Danfeng Zhang, Shuai Wang, Pei Wang, and Xiao Liu
for reporting this issue.
Reviewed-by: NPaul Dale <paul.dale@oracle.com>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5154)

C preprocessor flags get separated from C flags, which has the
advantage that we don't get loads of macro definitions and inclusion
directory specs when linking shared libraries, DSOs and programs.

This is a step to add support for "make variables" when configuring.
Reviewed-by: NTim Hudson <tjh@openssl.org>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5177)

Reviewed-by: NPaul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/5110)

Commit 9f944291 changed the semantics of BN_copy() to additionally
copy the BN_FLG_CONSTTIME flag if it is set. This turns out to be
ill advised as it has unintended consequences. For example calling
BN_mod_inverse_no_branch() can sometimes return a result with the flag
set and sometimes not as a result. This can lead to later failures if we
go down code branches that do not support constant time, but check for
the presence of the flag.

The original commit was made due to an issue in BN_MOD_CTX_set(). The
original PR fixed the problem in that function, but it was changed in
review to fix it in BN_copy() instead. The solution seems to be to revert
the BN_copy() change and go back to the originally proposed way.
Reviewed-by: NPaul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/5080)

Credit to OSS-Fuzz for finding this.

CVE-2017-3738
Reviewed-by: NRich Salz <rsalz@openssl.org>