Prepare for release of 1.1.1w

Reviewed-by: Matt Caswell <matt@openssl.org>
Release: yes

Copyright year updates

Reviewed-by: Matt Caswell <matt@openssl.org>
Release: yes

Add CVE-2023-4807 fix to CHANGES and NEWS

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22034)

remove unused Appveyour config

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22000)

(cherry picked from commit 8ac32e1e1b1a786366333acf897d332339610e6b)

Fix error handling in CMS_EncryptedData_encrypt

That caused several memory leaks in case of error.
Also when the CMS object that is created by CMS_EncryptedData_encrypt
is not used in the normal way, but instead just deleted
by CMS_ContentInfo_free some memory was lost.

Fixes #21985

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22008)

Fix a possbile memleak in rsa_pub_encode

That seems to be only an issue for RSA-PSS with parameters.
Spotted by code review, so it looks like there is no test coverage for this.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22007)

Fix a possible memleak in eckey_priv_encode

Additionally use OPENSSL_clear_free on the private
key data in case of error.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22007)

Bump actions/checkout from 2 to 4

Bumps [actions/checkout](https://github.com/actions/checkout) from 2 to 4.
- [Release notes](https://github.com/actions/checkout/releases)
- [Changelog](https://github.com/actions/checkout/blob/main/CHANGELOG.md)
- [Commits](https://github.com/actions/checkout/compare/v2...v4)

---
updated-dependencies:
- dependency-name: actions/checkout
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
CLA: trivial

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21953)

(cherry picked from commit d4231af60a8d04196b3b873c2fa8638daff36173)

Avoid clobbering non-volatile XMM registers

This affects some Poly1305 assembler functions
which are only used for certain CPU types.

Remove those functions for Windows targets,
as a simple interim solution.

Fixes #21522

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21808)

(cherry picked from commit 7b8e27bc2e02238986d89ef0ece067ec1b48e165)

Fix EVP_PKEY_asn1_copy

Add the copy of the omitted ASN1 public key method and
other members.

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/21125)

Fix typo in function name

Signed-off-by: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21608)

(cherry picked from commit 38c70a161cc6f96682bd77c8a935c5767355438c)

A null pointer dereference occurs when memory allocation fails

Fixes #21605

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21606)

Prepare for 1.1.1w

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Release: yes

Prepare for release of 1.1.1v

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Release: yes

Copyright year updates

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Release: yes

Add CHANGES/NEWS for CVE-2023-3817

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21551)

dhtest.c: Add test of DH_check() with q = p + 1

This must fail with DH_CHECK_INVALID_Q_VALUE and
with DH_CHECK_Q_NOT_PRIME unset.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21551)

DH_check(): Do not try checking q properties if it is obviously invalid

If  |q| >= |p| then the q value is obviously wrong as q
is supposed to be a prime divisor of p-1.

We check if p is overly large so this added test implies that
q is not large either when performing subsequent tests using that
q value.

Otherwise if it is too large these additional checks of the q value
such as the primality test can then trigger DoS by doing overly long
computations.

Fixes CVE-2023-3817

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21551)

Make DH_check set some error bits in recently added error

The pre-existing error cases where DH_check returned zero
are not related to the dh params in any way, but are only
triggered by out-of-memory errors, therefore having *ret
set to zero feels right, but since the new error case is
triggered by too large p values that is something different.
On the other hand some callers of this function might not
be prepared to handle the return value correctly but only
rely on *ret. Therefore we set some error bits in *ret as
additional safety measure.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21533)

Update CHANGES/NEWS for CVE-2023-3446

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21452)

Add a test for CVE-2023-3446

Confirm that the only errors DH_check() finds with DH parameters with an
excessively long modulus is that the modulus is too large. We should not
be performing time consuming checks using that modulus.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21452)

Fix DH_check() excessive time with over sized modulus

The DH_check() function checks numerous aspects of the key or parameters
that have been supplied. Some of those checks use the supplied modulus
value even if it is excessively large.

There is already a maximum DH modulus size (10,000 bits) over which
OpenSSL will not generate or derive keys. DH_check() will however still
perform various tests for validity on such a large modulus. We introduce a
new maximum (32,768) over which DH_check() will just fail.

An application that calls DH_check() and supplies a key or parameters
obtained from an untrusted source could be vulnerable to a Denial of
Service attack.

The function DH_check() is itself called by a number of other OpenSSL
functions. An application calling any of those other functions may
similarly be affected. The other functions affected by this are
DH_check_ex() and EVP_PKEY_param_check().

CVE-2023-3446

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21452)

Prepare for 1.1.1v

Reviewed-by: Richard Levitte <levitte@openssl.org>
Release: yes

Prepare for release of 1.1.1u

Reviewed-by: Richard Levitte <levitte@openssl.org>
Release: yes

Copyright year updates

Reviewed-by: Richard Levitte <levitte@openssl.org>
Release: yes

Restrict the size of OBJECT IDENTIFIERs that OBJ_obj2txt will translate

OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
numeric text form.  For gigantic sub-identifiers, this would take a very
long time, the time complexity being O(n^2) where n is the size of that
sub-identifier.

To mitigate this, a restriction on the size that OBJ_obj2txt() will
translate to canonical numeric text form is added, based on RFC 2578
(STD 58), which says this:

> 3.5. OBJECT IDENTIFIER values
>
> An OBJECT IDENTIFIER value is an ordered list of non-negative numbers.
> For the SMIv2, each number in the list is referred to as a sub-identifier,
> there are at most 128 sub-identifiers in a value, and each sub-identifier
> has a maximum value of 2^32-1 (4294967295 decimal).

Fixes otc/security#96
Fixes CVE-2023-2650

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>

x509: Handle ossl_policy_level_add_node errors

The invocation of ossl_policy_level_add_node in tree_calculate_user_set
did not have any error handling. Add it to prevent a memory leak for the
allocated extra policy data.

Also add error handling to sk_X509_POLICY_NODE_push to ensure that if
a new node was allocated, but could not be added to the stack, it is
freed correctly.

Fix error handling if tree->user_policies cannot be allocated by
returning 0, indicating failure, rather than 1.

Signed-off-by: Clemens Lang <cllang@redhat.com>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21066)

x509: Fix possible use-after-free when OOM

ossl_policy_level_add_node() first adds the new node to the level->nodes
stack, and then attempts to add extra data if extra_data is true. If
memory allocation or adding the extra data to tree->extra_data fails,
the allocated node (that has already been added to the level->nodes
stack) is freed using ossl_policy_node_free(), which leads to
a potential use after free.

Additionally, the tree's node count and the parent's child count would
not be updated, despite the new node being added.

Fix this by either performing the function's purpose completely, or not
at all by reverting the changes on error.

Signed-off-by: Clemens Lang <cllang@redhat.com>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21066)

util/mkdef.pl: Take shlib_variant into account

For platforms using import libraries, the lack of this causes a
disjoint between the name of the DLL that's produced, and the
corresponding import library.

Fixes #20942 (follows up #20732)

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20968)

Add OpenSSL OMC key to list of key fingerprints

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21030)

Fix the padlock engine

... after it was broken for almost 5 years,
since the first 1.1.1 release.
Note: The last working version was 1.1.0l release.

Fixes #20073

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/20147)

Fix stack corruption in ui_read

This is an alternative to #20893

Additionally this fixes also a possible issue in UI_UTIL_read_pw:

When UI_new returns NULL, the result code would still be zero
as if UI_UTIL_read_pw succeeded, but the password buffer is left
uninitialized, with subsequent possible stack corruption or worse.

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20957)

(cherry picked from commit a64c48cff88e032cf9513578493c4536df725a22)

Add negative integer check when using ASN1_BIT_STRING

The negative integer check is done to prevent potential overflow.
Fixes #20719.

CLA: trivial

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20862)

(cherry picked from commit 1258a8e4361320cd3cfaf9ede692492ce01034c8)

Drop invalid ERR_raise() call from incorrect cherry pick

Fixes #20849

Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20852)

Have Windows and VMS build files use shlib_variant

This is an omission, it should have been in place a long time ago.

Fixes #20732

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20734)

Re-add BN_F_OSSL_BN_RSA_DO_UNBLIND which was incorrectly removed

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20784)

Alternative fix for CVE-2022-4304

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/20284)

Revert "Fix Timing Oracle in RSA decryption"

This reverts commit 43d8f88511991533f53680a751e9326999a6a31f.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20284)

Fix documentation of X509_VERIFY_PARAM_add0_policy()

The function was incorrectly documented as enabling policy checking.

Fixes: CVE-2023-0466
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20564)

Updated CHANGES and NEWS for CVE-2023-0465

Also updated the entries for CVE-2023-0464

Related-to: CVE-2023-0465
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20588)

Add a Certificate Policies Test

Test that a valid certificate policy is accepted and that an invalid
certificate policy is rejected. Specifically we are checking that a
leaf certificate with an invalid policy is detected.

Related-to: CVE-2023-0465
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20588)

Ensure that EXFLAG_INVALID_POLICY is checked even in leaf certs

Even though we check the leaf cert to confirm it is valid, we
later ignored the invalid flag and did not notice that the leaf
cert was bad.

Fixes: CVE-2023-0465
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20588)

Generate some certificates with the certificatePolicies extension

Related-to: CVE-2023-0465
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20588)

Disable the policy tree exponential growth test conditionally

If there is no EC specified, the test won't pass.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/20572)

(cherry picked from commit f5935fcf8e4bc2191ac4a32e5b7ec32817642f1e)

changes: note about policy tree size limits and circumvention

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/20569)

test: add test cases for the policy resource overuse

These trees have pathological properties with respect to building.  The small
tree stays within the imposed limit, the large tree doesn't.

The large tree would consume over 150Gb of RAM to process.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/20569)

x509: excessive resource use verifying policy constraints

A security vulnerability has been identified in all supported versions
of OpenSSL related to the verification of X.509 certificate chains
that include policy constraints.  Attackers may be able to exploit this
vulnerability by creating a malicious certificate chain that triggers
exponential use of computational resources, leading to a denial-of-service
(DoS) attack on affected systems.

Fixes CVE-2023-0464

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/20569)

Fix Configure test for -mips in CFLAGS

We want to add -mips2 or -mips3 only if the user hasn't already
specified a mips version in CFLAGS. The existing test was a
double-negative.

CLA: trivial

Fixes: https://github.com/openssl/openssl/issues/20214
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20536)

Fix copyright disclaimer.

The mention of the GPL shouldn't have been there.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20517)

(cherry picked from commit c879f8ac56170a5cf929fab8067beb2a5902be2b)

Fix copyright header

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20473)

crypto/conf: gcc build warning fix

Fix the gcc build warning from conf_sap.c:
variable flags set but not used [-Wunused-but-set-variable]
variable appname set but not used [-Wunused-but-set-variable]
variable filename set but not used [-Wunused-but-set-variable]

CLA: trivial

Signed-off-by: Gang Chen <gang.c.chen@intel.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20347)

Fix failing cms test when no-des is used

The test tries to use DES but that may not be available.
But for the purpose of regression testing CVE-2023-0215
the cipher is not relevant, so we use AES-128 instead.

Fixes #20249

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/20276)

(cherry picked from commit c400a1fe477b44a5eacbad2be8d50f2eaa92925c)

rsa: add msvc intrinsic for non x64 platforms

_umul128() is x86_64 (x64) only, while __umulh() works everywhere, but
doesn't generate optimal code on x64

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20244)

(cherry picked from commit 075652f224479dad2e64b92e791b296177af8705)

Prepare for 1.1.1u-dev

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Release: yes

Prepare for 1.1.1t release

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Release: yes

Update copyright year

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Release: yes

Add CHANGES and NEWS entries for the 1.1.1t release

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>

CVE-2023-0286: Fix GENERAL_NAME_cmp for x400Address (1.1.1)

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>

Check CMS failure during BIO setup with -stream is handled correctly

Test for the issue fixed in the previous commit

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>

Fix a UAF resulting from a bug in BIO_new_NDEF

If the aux->asn1_cb() call fails in BIO_new_NDEF then the "out" BIO will
be part of an invalid BIO chain. This causes a "use after free" when the
BIO is eventually freed.

Based on an original patch by Viktor Dukhovni and an idea from Theo
Buehler.

Thanks to Octavio Galland for reporting this issue.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>

Add a test for CVE-2022-4450

Call PEM_read_bio_ex() and expect a failure. There should be no dangling
ptrs and therefore there should be no double free if we free the ptrs on
error.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>

Avoid dangling ptrs in header and data params for PEM_read_bio_ex

In the event of a failure in PEM_read_bio_ex() we free the buffers we
allocated for the header and data buffers. However we were not clearing
the ptrs stored in *header and *data. Since, on success, the caller is
responsible for freeing these ptrs this can potentially lead to a double
free if the caller frees them even on failure.

Thanks to Dawei Wang for reporting this issue.

Based on a proposed patch by Kurt Roeckx.

CVE-2022-4450

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>

Fix Timing Oracle in RSA decryption

A timing based side channel exists in the OpenSSL RSA Decryption
implementation which could be sufficient to recover a plaintext across
a network in a Bleichenbacher style attack. To achieve a successful
decryption an attacker would have to be able to send a very large number
of trial messages for decryption. The vulnerability affects all RSA
padding modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.

Patch written by Dmitry Belyavsky and Hubert Kario

CVE-2022-4304

Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>

Use $config{build_file} instead of $target{build_file}

If the user specifies an alternative build file than the default, this
alternative is recorded in $config{build_file}, not $target{build_file}.
Therefore, the former should be used, leaving the latter as a mere default.

This is a bug.

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/20174)

crypto/bio: drop float formating for UEFI

Using floating point is not supported in UEFI and can cause build
problems, for example due to SSE being disabled and x64 calling
convention passing floats in SSE registers.

Avoid those problems by not compiling the formating code for floating
point numbers.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19738)

Fix corruption when searching for CRLs in hashed directories (1.1)

The by_dir certificate/CRL lookup code uses an OPENSSL_STACK to track
how many sequentially numbered CRL files have been loaded for a given
X509_NAME hash which is being requested. This avoids loading already
loaded CRL files and repeated stat() calls.

This OPENSSL_STACK is searched using sk_find, however this mutates
the OPENSSL_STACK unless it is known to be sorted. This operation
therefore requires a write lock, which was not taken.

Fix this issue by sorting the OPENSSL_STACK whenever it is mutated. This
guarantees no mutation will occur during sk_find. This is chosen over
taking a write lock during sk_find as retrieving a CRL by X509_NAME is
assumed to be a hotter path than the case where a new CRL is installed.

Also optimise the code by avoiding creating the structure to track the
last CRL file sequence number in the circumstance where it would match
the initial value, namely where no CRL with the given hash is installed.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20127)

Add DTLS support to the large app data test

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20087)

Ensure our buffer allocation allows for the Explicit IV

Some ciphers/protocol versions have an explicit IV. We need to make sure we
have sufficient room for it in the underlying buffer.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20087)

Add a test for large app data

Test that sending large app data records works correctly.

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20087)

Padlock: fix byte swapping assembly for AES-192 and 256

Byte swapping code incorrectly uses the number of AES rounds to swap expanded
AES key, while swapping only a single dword in a loop, resulting in swapped
key and partially swapped expanded keys, breaking AES encryption and
decryption on VIA Padlock hardware.

This commit correctly sets the number of swapping loops to be done.

Fixes #20073

CLA: trivial

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20077)

(cherry picked from commit 7331e7ef79fe4499d81cc92249e9c97e9ff9291a)

Fix CI failures with ubuntu-22.04

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19930)

Fix various clang-14/15 compiler warnings

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19930)

Limit size of modulus for bn_mul_mont and BN_mod_exp_mont_consttime

Otherwise the alloca can cause an exception.

Issue reported by Jiayi Lin.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19735)

Ensure ossl_cms_EncryptedContent_init_bio() reports an error on no OID

If the cipher being used in ossl_cms_EncryptedContent_init_bio() has no
associated OID then we should report an error rather than continuing on
regardless. Continuing on still ends up failing - but later on and with a
more cryptic error message.

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19920)

Fix BIO_f_asn1() to properly report some errors

Some things that may go wrong in asn1_bio_write() are serious errors
that should be reported as -1, rather than 0 (which just means "we wrote
no data").

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19920)

Fix SMIME_crlf_copy() to properly report an error

If the BIO unexpectedly fails to flush then SMIME_crlf_copy() was not
correctly reporting the error. We modify it to properly propagate the
error condition.

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19920)

Fix BIO_f_cipher() flushing

If an error occurs during a flush on a BIO_f_cipher() then in some cases
we could get into an infinite loop. We add a check to make sure we are
making progress during flush and exit if not.

This issue was reported by Octavio Galland who also demonstrated an
infinite loop in CMS encryption as a result of this bug.

The security team has assessed this issue as not a CVE. This occurs on
*encryption* only which is typically processing trusted data. We are not
aware of a way to trigger this with untrusted data.

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19920)

Fix a logic flaw in test_mod_exp_zero

Due to the logic flaw, possible test failures
in this test case might be ignored.

Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19929)

(cherry picked from commit 42061268ee8f9ae0555d522870740fc91b744f4f)

Fix a memory leak in rsa_priv_encode

If PKCS8_pkey_set0 fails, the memory in rk need to be clear freed
otherwise it is owned by the PKCS8_PRIV_KEY_INFO.

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19694)

Fix the ceiling on how much encryption growth we can have

Stitched ciphersuites can grow by more during encryption than the code
allowed for. We fix the calculation and add an assert to check we go it
right.

Also if we are adding the MAC independently of the cipher algorithm then
the encryption growth will not include that MAC so we should remove it
from the amount of bytes that we reserve for that growth. Otherwise we
might exceed our buffer size and the WPACKET_reserve operation will
fail.

Note that this is not a security issue. Even though we can overflow the
amount of bytes reserved in the WPACKET for the encryption, the underlying
buffer is still big enough.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19585)

Resign test/certs/rootCA.pem to expire in 100 years

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19654)

(cherry picked from commit 43086b1bd48958ce95fadba8459ad88675da4fdf)

Update the validity period of ed25519 cerificates

Note: The private key is test/certs/root-ed25519.privkey.pem

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19654)

(cherry picked from commit 42f917432999b34ad8618ae03a5f199738a2b5ba)

Update GitHub actions as suggested by dependabot

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19581)

(cherry picked from commit ec33ed712665ca65cabcd87d446e5db79a64379e)

Prepare for 1.1.1t-dev

Reviewed-by: Richard Levitte <levitte@openssl.org>
Release: yes

Prepare for 1.1.1s release

Reviewed-by: Richard Levitte <levitte@openssl.org>
Release: yes

Update copyright year

Reviewed-by: Richard Levitte <levitte@openssl.org>
Release: yes
(Merged from https://github.com/openssl/openssl/pull/19560)

Update CHANGES and NEWS for new release

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19558)

Make openVMS seeding less dependent of OpenVMS version

SYS$GETTIM_PREC is a very new function, only available on OpenVMS v8.4.
OpenSSL binaries built on OpenVMS v8.4 become unusable on older OpenVM
versions, but building for the older CRTL version will make the high
precision time functions unavailable.

Tests have shown that on Alpha and Itanium, the time update granularity
between SYS$GETTIM and SYS$GETTIM_PREC is marginal, so the former plus
a sequence number turns out to be better to guarantee a unique nonce.

Fixes #18727

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/18730)

ssl_cipher_process_rulestr: don't read outside rule_str buffer

If rule_str ended in a "-", "l" was incremented one byte past the
end of the buffer.  This resulted in an out-of-bounds read when "l"
is dereferenced at the end of the loop.  It is safest to just return
early in this case since the condition occurs inside a nested loop.

CLA: trivial

Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19166)

(cherry picked from commit 428511ca66670e169a0e1b12e7540714b0be4cf8)

doc: fix copy/paste error

Fixes #19460

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/19461)

(cherry picked from commit 5b9480fc1e814bf8fa2dce0dbbede147f04d477c)

Fix many inconsistencies in doc of CMS_verify() and PKC7_verify() etc.

Also change B< to I< in {CMS,PKCS7}_verify.pod, PKCS7_sign{,_add_signer}.pod

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/19108)

(cherry picked from commit 312a6b3a0327a986344c85aa6bc43e135d70bc6c)

Fix no longer implicitly refresh the cached TBSCertificate

This reverts commit 748df1874f0488ce0c86b6d2d083921abb34b1e3.
Fixes #19388

Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19392)

Prepare for 1.1.1s-dev

Reviewed-by: Richard Levitte <levitte@openssl.org>
Release: yes

Prepare for 1.1.1r release

Reviewed-by: Richard Levitte <levitte@openssl.org>
Release: yes

Update copyright year

Reviewed-by: Richard Levitte <levitte@openssl.org>
Release: yes
(Merged from https://github.com/openssl/openssl/pull/19383)

Updated CHANGES and NEWS for new release

Reviewed-by: Richard Levitte <levitte@openssl.org>
Release: yes
(Merged from https://github.com/openssl/openssl/pull/19381)

Configurations: mips64*-linux-*abin32 needs bn_ops SIXTY_FOUR_BIT

The IRIX mips64-cpu, n32-abi configurations include SIXTY_FOUR_BIT in bn_ops,
but it is missing from mips64*-linux-*abin32 (which OpenSSL calls
"linux-mips64").  This causes heap corruption when verifying TLS certificates
(which tend to be RSA-signed) with openssl 1.1.1q:

```
nix@oak:~$ /nix/store/4k04dh6a1zs6hxiacwcg4a4nvxvgli2j-openssl-mips64el-unknown-linux-gnuabin32-1.1.1q-bin/bin/openssl s_client -host www.google.com -port 443free(): invalid pointer
Aborted
```

and a slightly different failure with current HEAD:

```
nix@oak:~$ /nix/store/9bqxharxajsl9fid0c8ls6fb9wxp8kdc-openssl-mips64el-unknown-linux-gnuabin32-1.1.1q-bin/bin/openssl s_client -host www.google.com -port 443
Connecting to 142.250.180.4
CONNECTED(00000003)
Fatal glibc error: malloc assertion failure in sysmalloc: (old_top == initial_top (av) && old_size == 0) || ((unsigned long) (old_size) >= MINSIZE && prev_inuse (old_top) && ((unsigned long) old_end & (pagesize - 1)) == 0)
Aborted
```

Applying this patch and recompiling produces the expected output instead of a
crash.

Note that Gentoo (and to my knowledge all other other distributions which
support mips64n32) use the `linux-generic32` configuration, which uses only
32-bit arithmetic (rather than full 64-bit arithmetic) and lacks assembler
implementations for the SHA hash functions:

  https://gitweb.gentoo.org/repo/gentoo.git/tree/dev-libs/openssl/files/gentoo.config-1.0.2#n102

For support in nixpkgs we would like to use the full 64-bit integer registers
and perlasm routines, so I'm submitting this upstream as well.

Fixes #19319

CLA: trivial

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19320)

(cherry picked from commit d250e8563fa400fd3d9b93cff609c7503149b908)

X509 x_all.c: Set 'modified' flag when ASN1_item_sign{,_ctx} call was successful

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/18879)

X509 x509_req.c: Set 'modified' flag when X509_req_info_st member data updated

We need to reencode X509_req_info_st if member data updated.

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/18879)

Convert serverinfo in SSL_CTX_use_serverinfo() to v2.

Fixes #18183.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19081)