In the definition of the latest revised LoongArch64 vector instruction manual,
it is clearly pointed out that the undefined upper three bits of each byte in
the control register of the vshuf.b instruction should not be used, otherwise
uncertain results may be obtained. Therefore, it is necessary to correct the
use of the vshuf.b instruction in the existing vpaes-loongarch64.pl code to
avoid erroneous calculation results in future LoongArch64 processors.
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21530)
(cherry picked from commit 780ce3849f)
This backports 7b508cd1e1 together with .pl fixes
Makes the SIGILL-based code easier to read, and doesn't use it on Apple Silicon
or where getauxval() is present, thereby improving stability when debugging on
Darwin (macOS/iOS/etc) and in multi-threaded programs (both Darwin and Linux).
Fixes#21541
Change-Id: I07912f0ddcbfe15bf3c1550533855a6583d21b67
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21583)
Also fixes a similar regression in X509_VERIFY_PARAM_add0_table().
Commit 38ebfc3 introduced a regression in 3.0.6 that changed the return
value of the two functions above from 1 on success to the number of entries
in the stack. If there are more than one entry then this is a change in
behaviour which should not have been introduced into a stable release.
This reverts the behaviour back to what it was prior to the change. The code
is slightly different to the original code in that we also handle a possible
-1 return value from the stack push function. This should never happen in
reality because we never pass a NULL stack as a parameter - but for the sake
of robustness we handle it anyway.
Note that the changed behaviour exists in all versions of 3.1 (it never had
the original version). But 3.1 should be fully backwards compatible with 3.0
so we should change it there too.
Fixes#21570
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
(Merged from https://github.com/openssl/openssl/pull/21576)
(cherry picked from commit e3d897d3fa)
Running LSX instructions requires both the hardware support and the
kernel support. The `cpucfg` instruction only tests the hardware
support, causing a SIGILL if the hardware supports LSX but the kernel
does not.
Use `getauxval(AT_HWCAP)` as the ["Software Development and Build
Convention for LoongArch Architectures"][1] manual suggests.
The LOONGARCH_HWCAP_LSX and LOONGARCH_HWCAP_LASX bits are copied from
the manual too. In Glibc 2.38 they'll be provided by <sys/auxv.h> as
well, but they are unavailable in earlier Glibc versions so we cannot
rely on it.
The getauxval syscall and Glibc wrapper are available since day one
(Linux-5.19 and Glibc-2.36) for LoongArch.
Fixes#21508.
[1]:https://github.com/loongson/la-softdev-convention/blob/master/la-softdev-convention.adoc#kernel-constraints
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21509)
(cherry picked from commit c612289b77)
Apply normal sentence case to db update message
CLA: trivial
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/21535)
(cherry picked from commit ccb2f3080d)
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/21550)
(cherry picked from commit 4b29762802)
This must fail with DH_CHECK_INVALID_Q_VALUE and
with DH_CHECK_Q_NOT_PRIME unset.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/21550)
(cherry picked from commit ad5d355726)
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: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/21550)
(cherry picked from commit 1c16253f3c)
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/21524)
(cherry picked from commit 81d10e61a4)
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21515)
(cherry picked from commit bc5d9cc871)
When an early error happens in the test_pipelining
function and idx == 5 the error handling would try
to call OPENSSL_free(msg), but msg is at that time
just a string constant in read-only memory, so a
crash would be the result. Fixed that by using
fragsize as an indication when to free msg.
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21512)
(cherry picked from commit 06a0d40322)
Function `X509at_add1_attr()` (crypto/x509/x509_att.c) rejects to add a duplicity into `*x` but it searches in a wrong stack.
Changed to search in `*x`.
CLA: trivial
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21505)
(cherry picked from commit 7551264186)
Running 'openssl speed eddsa' fails with
Doing 253 bits sign Ed25519 ops for 10s: EdDSA sign failure
000003FF9306C7D0:error:030000BC:digital envelope routines:EVP_DigestSign:
final error:crypto/evp/m_sigver.c:585:
-1 253 bits Ed25519 sign ops in 0.00s
Doing 253 bits verify Ed25519 ops for 10s: EdDSA verify failure
000003FF9306C7D0:error:030000BC:digital envelope routines:EVP_DigestVerify:
final error:crypto/evp/m_sigver.c:694:
-1 253 bits Ed25519 verify ops in 0.00s
This is because the EVP_DigestSign/Verify() calls in the EdDSA_sign/verify_loop()
fail because the context has already been finalized by the previous
EVP_DigestSign/Verify call during the EdDSA signature test done by speed_main().
This happens since commit 3fc2b7d6b8 where the
EVP_DigestSign/Verify() functions have been changed to set a flag that the
context has been finalized.
Fix this by re-initializing the context using EVP_DigestSign/Verify() in the
EdDSA_sign/verify_loop().
Signed-off-by: Ingo Franzki <ifranzki@linux.ibm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21491)
(cherry picked from commit 0c85bcbaea)
Free the signature stack after iterating over all found signatures.
Free the kem and signature stacks at the end of speed_main() if not
NULL.
Signed-off-by: Ingo Franzki <ifranzki@linux.ibm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21491)
(cherry picked from commit cc7e2b20de)
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/21451)
(cherry picked from commit 4ec53ad6e1)
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/21451)
(cherry picked from commit ede782b4c8)
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/21451)
(cherry picked from commit 9e0094e2aa)
Only modify doc/man* in the openssl-3.1 branch.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21465)
Signed-off-by: lan1120 <lanming@huawei.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21170)
(cherry picked from commit 4c4fefa5c7)
Reviewed-by: Tomas Mraz <tomas@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/21401)
(cherry picked from commit f1b7243cda)
A BIO is documented to return -1 on write retry - but sometimes they return
0. ssl3_do_write() was incorrectly handling a 0 response.
Fixes#21422
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21435)
Test various scenarios for a write retry occuring during a handshake.
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21435)
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21459)
(cherry picked from commit 3d2f96e2c8)
Reviewed-by: Tomas Mraz <tomas@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/21401)
(cherry picked from commit 64b1d2fb06)
The AES-SIV mode allows for multiple associated data items
authenticated separately with any of these being 0 length.
The provided implementation ignores such empty associated data
which is incorrect in regards to the RFC 5297 and is also
a security issue because such empty associated data then become
unauthenticated if an application expects to authenticate them.
Fixes CVE-2023-2975
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21384)
(cherry picked from commit c426c281cf)
This change is intended to provide some safety for uninitialized stack failures
that have appeared in 80-test_cmp_http on NonStop x86 when run in a complex
CI/CD Jenkins environment. This change also adds init_pint() to handle the
initialization of a pointer to int value.
Fixes: #21083
Signed-off-by: Randall S. Becker <randall.becker@nexbridge.ca>
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/21109)
If you are building the latest release source code with enable-fips configured
then the FIPS provider you are using is not likely to be FIPS compliant.
This update demonstrates how to build a FIPS provider that is compliant
and use it with the latest source code.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/20907)
(cherry picked from commit 2b42290f08)
rhotates tables are placed to .text section which confuses tools such as BOLT.
Move them to rodata to unbreak and avoid polluting icache/iTLB with data.
CLA: trivial
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Paul Yang <kaishen.yy@antfin.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21440)
(cherry picked from commit 2fd82c2283)
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/21386)
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/21397)
(cherry picked from commit 5be8233d2b)
FIX: https://github.com/openssl/openssl/issues/21299
ret in ossl_config_int() only used to check return value of
CONF_modules_load_file(), should set it to 1 if in UEFI system.
Signed-off-by: Yi Li <yi1.li@intel.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21300)
(cherry picked from commit 500e479db1)
CLA: trivial
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21319)
(cherry picked from commit 063cdca61a)
The implementation is not usable there at all.
Fixes#21301
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21312)
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21295)
In TLSv1.2 we should not attempt to use a supported_group value that is
intended for use with TLSv1.3 - even if both the server and the client
support it, e.g. the ffdhe groups are supported by OpenSSL for TLSv1.3 but
not for TLSv1.2.
Fixes#21081
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21295)
zhuchen
Tom Cosgrove
Matt Caswell
Xi Ruoyao
Fatih Arslan Tugay
Tomas Mraz
Bernd Edlinger
atishkov
Adam Šulc
Ingo Franzki
Samuel Lee
Dr. David von Oheimb
Dimitri Papadopoulos
lan1120
ljuzwiuk
Randall S. Becker
dependabot[bot]
slontis
Amir Ayupov
Pauli
Yi Li
Duncan Thomson
Jakub Wilk