The real cause for this change is that test/ec_internal_test.c
includes ec_lcl.h, and including curve448/curve448_lcl.h from there
doesn't work so well with compilers who always do inclusions relative
to the C file being compiled.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8334)
Thanks to David Benjamin who reported this, performed the analysis and
suggested the patch. I have incorporated some of his analysis in the
comments below.
This issue can cause an out-of-bounds read. It is believed that this was
not reachable until the recent "fixed top" changes. Analysis has so far
only identified one code path that can encounter this - although it is
possible that others may be found. The one code path only impacts 1.0.2 in
certain builds. The fuzzer found a path in RSA where iqmp is too large. If
the input is all zeros, the RSA CRT logic will multiply a padded zero by
iqmp. Two mitigating factors:
- Private keys which trip this are invalid (iqmp is not reduced mod p).
Only systems which take untrusted private keys care.
- In OpenSSL 1.1.x, there is a check which rejects the oversize iqmp,
so the bug is only reproducible in 1.0.2 so far.
Fortunately, the bug appears to be relatively harmless. The consequences of
bn_cmp_word's misbehavior are:
- OpenSSL may crash if the buffers are page-aligned and the previous page is
non-existent.
- OpenSSL will incorrectly treat two BN_ULONG buffers as not equal when they
are equal.
- Side channel concerns.
The first is indeed a concern and is a DoS bug. The second is fine in this
context. bn_cmp_word and bn_cmp_part_words are used to compute abs(a0 - a1)
in Karatsuba. If a0 = a1, it does not matter whether we use a0 - a1 or
a1 - a0. The third would be worth thinking about, but it is overshadowed
by the entire Karatsuba implementation not being constant time.
Due to the difficulty of tripping this and the low impact no CVE is felt
necessary for this issue.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8326)
(cherry picked from commit 576129cd72)
Ty Baen-Price explains:
> Problem and Resolution:
> The following lines of code make use of the Microsoft API ExitProcess:
>
> ```
> Apps\Speed.c line 335: ExitProcess(ret);
> Ms\uplink.c line 22: ExitProcess(1);
> ```
>
> These function calls are made after fatal errors are detected and
> program termination is desired. ExitProcess(), however causes
> _orderly_ shutdown of a process and all its threads, i.e. it unloads
> all dlls and runs all destructors. See MSDN for details of exactly
> what happens
> (https://msdn.microsoft.com/en-us/library/windows/desktop/ms682658(v=vs.85).aspx).
> The MSDN page states that ExitProcess should never be called unless
> it is _known to be safe_ to call it. These calls should simply be
> replaced with calls to TerminateProcess(), which is what should be
> called for _disorderly_ shutdown.
>
> An example of usage:
>
> ```
> TerminateProcess(GetCurrentProcess(), exitcode);
> ```
>
> Effect of Problem:
> Because of a compilation error (wrong c++ runtime), my program
> executed the uplink.c ExitProcess() call. This caused the single
> OpenSSL thread to start executing the destructors of all my dlls,
> and their objects. Unfortunately, about 30 other threads were
> happily using those objects at that time, eventually causing a
> 0xC0000005 ACCESS_VIOLATION. Obviously an ACCESS_VIOLATION is the
> best case scenario, as I'm sure you can imagine at the consequences
> of undiscovered memory corruption, even in a terminating process.
And on the subject of `TerminateProcess()` being asynchronous:
> That is technically true, but I think it's probably synchronous
> "enough" for your purposes, since a call to TerminateProcess
> suspends execution of all threads in the target process. This means
> it's really only asynchronous if you're calling TerminateProcess one
> some _other_ process. If you're calling TerminateProcess on your own
> process, you'll never return from the TerminateProcess call.
Fixes#2489
Was originally RT-4526
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8301)
(cherry picked from commit 9257959950)
Prior to this commit we were keeping a count of how many KeyUpdates we
have processed and failing if we had had too many. This simplistic approach
is not sufficient for long running connections. Since many KeyUpdates
would not be a particular good DoS route anyway, the simplest solution is
to simply remove the key update count.
Fixes#8068
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/8299)
(cherry picked from commit 3409a5ff8a)
Fixes#7950
It was reported that there might be a null pointer dereference in the
implementation of the dasync_aes_128_cbc_hmac_sha1() cipher, because
EVP_aes_128_cbc_hmac_sha1() can return a null pointer if AES-NI is
not available. It took some analysis to find out that this is not
an issue in practice, and these comments explain the reason to comfort
further NPD hunters.
Detected by GitHub user @wurongxin1987 using the Sourcebrella Pinpoint
static analyzer.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8305)
(cherry picked from commit a4a0a1eb43)
The aes128_cbc_hmac_sha1 cipher in the dasync engine is broken. Probably
by commit e38c2e8535 which removed use of the "enc" variable...but not
completely.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/8291)
(cherry picked from commit 695dd3a332)
The option is a flag for Options, not a standalone setting.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8292)
(cherry picked from commit 4ac5e43da6)
This commit adds a simple unit test to make sure that the constant-time
flag does not "leak" among BN_CTX frames:
- test_ctx_consttime_flag() initializes (and later frees before
returning) a BN_CTX object, then it calls in sequence
test_ctx_set_ct_flag() and test_ctx_check_ct_flag() using the same
BN_CTX object. The process is run twice, once with a "normal"
BN_CTX_new() object, then with a BN_CTX_secure_new() one.
- test_ctx_set_ct_flag() starts a frame in the given BN_CTX and sets the
BN_FLG_CONSTTIME flag on some of the BIGNUMs obtained from the frame
before ending it.
- test_ctx_check_ct_flag() then starts a new frame and gets a number of
BIGNUMs from it. In absence of leaks, none of the BIGNUMs in the new
frame should have BN_FLG_CONSTTIME set.
In actual BN_CTX usage inside libcrypto the leak could happen at any
depth level in the BN_CTX stack, with varying results depending on the
patterns of sibling trees of nested function calls sharing the same
BN_CTX object, and the effect of unintended BN_FLG_CONSTTIME on the
called BN_* functions.
This simple unit test abstracts away this complexity and verifies that
the leak does not happen between two sibling functions sharing the same
BN_CTX object at the same level of nesting.
(cherry picked from commit fe16ae5f95)
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8253)
This commit adds a dedicated function in `EC_METHOD` to access a modular
field inversion implementation suitable for the specifics of the
implemented curve, featuring SCA countermeasures.
The new pointer is defined as:
`int (*field_inv)(const EC_GROUP*, BIGNUM *r, const BIGNUM *a, BN_CTX*)`
and computes the multiplicative inverse of `a` in the underlying field,
storing the result in `r`.
Three implementations are included, each including specific SCA
countermeasures:
- `ec_GFp_simple_field_inv()`, featuring SCA hardening through
blinding.
- `ec_GFp_mont_field_inv()`, featuring SCA hardening through Fermat's
Little Theorem (FLT) inversion.
- `ec_GF2m_simple_field_inv()`, that uses `BN_GF2m_mod_inv()` which
already features SCA hardening through blinding.
From a security point of view, this also helps addressing a leakage
previously affecting conversions from projective to affine coordinates.
This commit also adds a new error reason code (i.e.,
`EC_R_CANNOT_INVERT`) to improve consistency between the three
implementations as all of them could fail for the same reason but
through different code paths resulting in inconsistent error stack
states.
Co-authored-by: Nicola Tuveri <nic.tuv@gmail.com>
(cherry picked from commit e0033efc30)
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/8262)
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8175)
(cherry picked from commit 8e981051ce)
The AIX binder needs to be instructed that the output will have no entry
point (see AIX' ld manual: -e in the Flags section; autoexp and noentry
in the Binder section).
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8282)
(cherry picked from commit c1b3846242)
Cygwin binaries should not enforce text mode these days, just
use text mode if the underlying mount point requests it
Signed-off-by: Corinna Vinschen <vinschen@redhat.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8248)
(cherry picked from commit 9b57e4a1ef)
The add/double shortcut in ecp_nistz256-x86_64.pl left one instruction
point that did not unwind, and the "slow" path in AES_cbc_encrypt was
not annotated correctly. For the latter, add
.cfi_{remember,restore}_state support to perlasm.
Next, fill in a bunch of functions that are missing no-op .cfi_startproc
and .cfi_endproc blocks. libunwind cannot unwind those stack frames
otherwise.
Finally, work around a bug in libunwind by not encoding rflags. (rflags
isn't a callee-saved register, so there's not much need to annotate it
anyway.)
These were found as part of ABI testing work in BoringSSL.
Reviewed-by: Richard Levitte <levitte@openssl.org>
GH: #8109
(cherry picked from commit c0e8e5007b)
safestack.h, lhash.h and sparse_array.h all define macros to generate
a full API for the containers as static inline functions. This
potentially generates unused code, which some compilers may complain
about.
We therefore need to mark those generated functions as unused, so the
compiler knows that we know, and stops complaining about it.
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/8246)
(cherry picked from commit 48fe4ce104)
Otherwise this can result in an incorrect calculation of the maximum
encoded integer length, meaning an insufficient buffer size is allocated.
Thanks to Billy Brumley for helping to track this down.
Fixes#8209
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8237)
(cherry picked from commit 9fc8f18f59)
The "verify_return_error" option in s_client is documented as:
Return verification errors instead of continuing. This will typically
abort the handshake with a fatal error.
In practice this option was ignored unless also accompanied with the
"-verify" option. It's unclear what the original intention was. One fix
could have been to change the documentation to match the actual behaviour.
However it seems unecessarily complex and unexpected that you should need
to have both options. Instead the fix implemented here is make the option
match the documentation so that "-verify" is not also required.
Note that s_server has a similar option where "-verify" (or "-Verify") is
still required. This makes more sense because those options additionally
request a certificate from the client. Without a certificate there is no
possibility of a verification failing, and so "-verify_return_error" doing
nothing seems ok.
Fixes#8079
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/8080)
(cherry picked from commit 78021171db)
The original 1.1.1 design was to use SSL_CB_HANDSHAKE_START and
SSL_CB_HANDSHAKE_DONE to signal start/end of a post-handshake message
exchange in TLSv1.3. Unfortunately experience has shown that this confuses
some applications who mistake it for a TLSv1.2 renegotiation. This means
that KeyUpdate messages are not handled properly.
This commit removes the use of SSL_CB_HANDSHAKE_START and
SSL_CB_HANDSHAKE_DONE to signal the start/end of a post-handshake
message exchange. Individual post-handshake messages are still signalled in
the normal way.
This is a potentially breaking change if there are any applications already
written that expect to see these TLSv1.3 events. However, without it,
KeyUpdate is not currently usable for many applications.
Fixes#8069
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8096)
(cherry picked from commit 4af5836b55)
set_cipher_list() sets TLSv1.2 (and below) ciphers, and its success or
failure should not depend on whether set_ciphersuites() has been used to
setup TLSv1.3 ciphers.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/7759)
(cherry picked from commit 3c83c5ba4f)
There are times when one might want to use something like
DEFINE_STACK_OF in a .c file, because it defines a stack for a type
defined in that .c file. Unfortunately, when configuring with
`--strict-warnings`, clang aggressively warn about unused functions in
such cases, which forces the use of such DEFINE macros to header
files.
We therefore disable this warning from the `--strict-warnings`
definition for clang.
(note for the curious: `-Wunused-function` is enabled via `-Wall`)
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8234)
(cherry picked from commit f11ffa505f)
CLA: trivial
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Paul Yang <yang.yang@baishancloud.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8137)
(cherry picked from commit b754a8a159)
ARMv8.3 adds pointer authentication extension, which in this case allows
to ensure that, when offloaded to stack, return address is same at return
as at entry to the subroutine. The new instructions are nops on processors
that don't implement the extension, so that the vetification is backward
compatible.
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8205)
(cherry picked from commit 9a18aae5f2)
This allows the user to override our defaults if needed, and in a
consistent manner.
Partial fix for #7607
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/7624)
(cherry picked from commit ca811248d8)
If the old openssl versions not supporting the .include directive
load a config file with it, they will bail out with error.
This change allows using the .include = <filename> syntax which
is interpreted as variable assignment by the old openssl
config file parser.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8141)
(cherry picked from commit 9d5560331d)
o2i_ECPublicKey() requires an EC_KEY structure filled with an EC_GROUP.
o2i_ECPublicKey() is called by d2i_PublicKey(). In order to fulfill the
o2i_ECPublicKey()'s requirement, d2i_PublicKey() needs to be called with
an EVP_PKEY with an EC_KEY containing an EC_GROUP.
However, the call to EVP_PKEY_set_type() frees any existing key structure
inside the EVP_PKEY, thus freeing the EC_KEY with the EC_GROUP that
o2i_ECPublicKey() needs.
This means you can't d2i_PublicKey() for an EC key...
The fix is to check to see if the type is already set appropriately, and
if so, not call EVP_PKEY_set_type().
Reviewed-by: Paul Yang <yang.yang@baishancloud.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8168)
(cherry picked from commit 2aa2beb06c)
The manual says this in its notes:
... and therefore applications using static linking should also call
OPENSSL_thread_stop() on each thread. ...
Fixes#8171
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/8173)
(cherry picked from commit 03cdfe1efa)
Making this a no-op removes a potential infinite loop than can occur in
some situations.
Fixes#2865
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8167)
(cherry picked from commit ef45aa14c5)
Trim trailing whitespace. It doesn't match OpenSSL coding standards,
AFAICT, and it can cause problems with git tooling.
Trailing whitespace remains in test data and external source.
Backport-of: https://github.com/openssl/openssl/pull/8092
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8134)
When computing the end-point shared secret, don't take the
terminating NULL character into account.
Please note that this fix breaks interoperability with older
versions of OpenSSL, which are not fixed.
Fixes#7956
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/7957)
(cherry picked from commit 09d62b336d)
VMS doesn't currently support unloading of shared object, and we need
to reflect that. Without this, the shlibload test fails
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8131)
(cherry picked from commit d1dd5d6f4c)