This is especially harmful since OPENSSL_cleanup() has already called
the RAND cleanup function
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3137)
If no default method was yet given, RAND_get_rand_method() will set it
up. Doing so just to clean it away seems pretty silly, so instead,
use the default_RAND_meth variable directly.
This also clears a possible race condition where this will try to init
things, such as ERR or ENGINE when in the middle of a OPENSSL_cleanup.
Fixes#3128
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3136)
This commit contains some optimizations in PKCS5_PBKDF2_HMAC() and
HMAC_CTX_copy() functions which together makes PBKDF2 computations
faster by 15-40% according to my measurements made on x64 Linux with
both asm optimized and no-asm versions of SHA1, SHA256 and SHA512.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1708)
Credit to OSS-Fuzz for finding this.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3088)
It seems to be problematic to probe processor capabilities with SIGILL
on MacOS X. The problem should be limited to cases when application code
is debugged, but crashes were reported even during normal execution...
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Originally there was dependency on BN configuration parameters, but
it stemmed from times when "long long" support was optional. Today
we require 64-bit support from compiler, and there is no reason to
have "greatest-width integer" depend on BN configuration.
Reviewed-by: Richard Levitte <levitte@openssl.org>
'j' is specified as modifier for "greatest-width integer type", which in
practice means 64 bits on both 32- and 64-bit platforms. Since we rely
on __attribute__((__format__(__printf__,...))) to sanitize BIO_print
format, we can use it to denote [u]int64_t-s in platform-neutral manner.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3083)
Add OPENSSL_SYS_UEFI to remove unused syslog and uid stuffs for
more clean UEFI build.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2961)
Fix some comments too
[skip ci]
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3069)
Even though Apple refers to Procedure Call Standard for ARM Architecture
(AAPCS), they apparently adhere to custom version that doesn't follow
stack alignment constraints in the said standard. [Why or why? If it's
vendor lock-in thing, then it would be like worst spot ever.] And since
bsaes-armv7 relied on standard alignment, it became problematic to
execute the code on iOS.
Reviewed-by: Rich Salz <rsalz@openssl.org>
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: Richard Levitte <levitte@openssl.org>
One could have fixed the problem by arranging 64-bit alignment of
EVP_AES_OCB_CTX.aad_buf in evp/e_aes.c, but CRYPTO_ocb128_aad
prototype doesn't imply alignment and we have to honour it.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2994)
Initial IV was disregarded on SHAEXT-capable processors. Amazingly
enough bulk AES128-SHA* talk-to-yourself tests were passing.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2992)
As hinted by its name new subroutine processes 8 input blocks in
parallel by loading data to 512-bit registers. It still needs more
work, as it needs to handle some specific input lengths better.
In this sense it's yet another intermediate step...
Reviewed-by: Rich Salz <rsalz@openssl.org>
These two functions do the same thing.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3001)
LONG and ZLONG items (which are OpenSSL private special cases of
ASN1_INTEGER) are encoded into DER with padding if the leading octet
has the high bit set, where the padding can be 0x00 (for positive
numbers) or 0xff (for negative ones).
When decoding DER to LONG or ZLONG, the padding wasn't taken in
account at all, which means that if the encoded size with padding
is one byte more than the size of long, decoding fails. This change
fixes that issue.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3000)
Fix a strict aliasing issue in ui_dup_method_data.
Add test coverage for CRYPTO_dup_ex_data, use OPENSSL_assert.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2988)
The documentation of this function states that the password parameter
can be NULL. However, the implementation returns an error in this case
due to the inner workings of the HMAC_Init_ex() function.
With this change, NULL password will be treated as an empty string and
PKCS5_PBKDF2_HMAC() no longer fails on this input.
I have also added two new test cases that tests the handling of the
special values NULL and -1 of the password and passlen parameters,
respectively.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1692)
Just as for DH, DSA and RSA, this gives the engine associated with the
key.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2960)
and RSA_verify_PKCS1_PSS_mgf1 with 512-bit RSA vs. sha-512.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2801)
As hinted by its name new subroutine processes 4 input blocks in
parallel. It still operates on 256-bit registers and is just
another step toward full-blown AVX512IFMA procedure.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Exteneded feature flags were not pulled on AMD processors, as result
a number of extensions were effectively masked on Ryzen. Original fix
for x86_64cpuid.pl addressed this problem, but messed up processor
vendor detection. This fix moves extended feature detection past
basic feature detection where it belongs. 32-bit counterpart is
harmonized too.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
conf has the ability to expand variables in config files. Repeatedly doing
this can lead to an exponential increase in the amount of memory required.
This places a limit on the length of a value that can result from an
expansion.
Credit to OSS-Fuzz for finding this problem.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2894)
It shouldn't try to return an action description for UIT_PROMPT type
UI strings.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2904)
Exteneded feature flags were not pulled on AMD processors, as result a
number of extensions were effectively masked on Ryzen. It should have
been reported for Excavator since it implements AVX2 extension, but
apparently nobody noticed or cared...
Reviewed-by: Rich Salz <rsalz@openssl.org>
Commit d5aa14dd simplified the bn_expand_internal() and BN_copy() functions.
Unfortunately it also removed some checks which are still required,
otherwise we call memcpy passing in NULL which is not allowed.
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2836)
Private hstrerror was introduced to address linking problem on HP-UX,
but truth be told conemporary systems, HP-UX included, wouldn't come
to that call, they would use getaddrinfo and gai_strerror, while
gethostbyname and h_errno are there to serve legacy systems. Since
legacy systems are naturally disappearing breed, we can as well just
let user interpret number.
GH#2816
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
With VMS C, the second parameter takes a 32-bit pointer. When
building with 64-bit pointer size default, we must compensate.
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2811)
- FLAT_INC
- PKCS1_CHECK (the SSL_OP_PKCS1_CHECK options have been
no-oped)
- PKCS_TESTVECT (debugging leftovers)
- SSL_AD_MISSING_SRP_USERNAME (unfinished feature)
- DTLS_AD_MISSING_HANDSHAKE_MESSAGE (unfinished feature)
- USE_OBJ_MAC (note this removes a define from the public header but
very unlikely someone would be depending on it)
- SSL_FORBID_ENULL
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Stephen Henson <steve@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
It's undocumented and unused in the tree. The idea seems to have
never gained much traction, and can be removed without breaking
ABI compatibility.
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2800)
Cleanse instead, and free in the free routine.
Seems to have been introduced in commit
846ec07d90 when EVP_CIPHER_CTX was made
opaque.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2798)
This is a bogus, undocumented format that was intended for testing; I
don't think anyone is using it.
Reviewed-by: Richard Levitte <levitte@openssl.org>
It is never built and the code is duplicated in bf_enc.c.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2778)
There are a number of symbols in bn which are internal only and never used
by anything. They should be removed.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2766)
This removes the fips configure option. This option is broken as the
required FIPS code is not available.
FIPS_mode() and FIPS_mode_set() are retained for compatibility, but
FIPS_mode() always returns 0, and FIPS_mode_set() can only be used to
turn FIPS mode off.
Reviewed-by: Stephen Henson <steve@openssl.org>
This reimplementation was necessary before VMS C V7.1. Since that's
the minimum version we support in this OpenSSL version, the
reimplementation is no longer needed.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2762)
There are cases when, if you pass a NULL UI_METHOD, the called
function will use an internal default. This is well and good, but
there may be cases when this is undesirable and one would rather send
in a UI that does absolutely nothing (sort of a /dev/null). UI_null()
is the UI_METHOD for this purpose.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2744)
On pre-Skylake best optimization strategy was balancing port-specific
instructions, while on Skylake minimizing the sheer amount appears
more sensible.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Avoid a -Wundef warning in refcount.h
Avoid a -Wundef warning in o_str.c
Avoid a -Wundef warning in testutil.h
Include internal/cryptlib.h before openssl/stack.h
to avoid use of undefined symbol OPENSSL_API_COMPAT.
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2712)
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2721)
If ret is allocated, it may be leaked on error.
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2666)
Prevent that memory beyond the last element is accessed if every element
of group->poly[] is non-zero
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2689)
opendir(), readdir() and closedir() have been available on VMS since
version 7.0.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2707)
Change size comparison from > (GT) to >= (GTE) to ensure an additional
byte of output buffer, to prevent OOB reads/writes later in the function
Reject input strings larger than 2GB
Detect invalid output buffer size and return early
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2672)
The sh_add_to_list function will overwrite subsequent slots in the free list
for small allocations. This causes a segmentation fault if the writes goes
off the end of the secure memory. I've not investigated if this problem
can overwrite memory without the segmentation fault, but it seems likely.
This fix limits the minsize to the sizeof of the SH_LIST structure (which
also has a side effect of properly aligning the pointers).
The alternative would be to return an error if minsize is too small.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2657)
This implementation is written in endian agnostic C code. No attempt
at providing machine specific assembly code has been made. This
implementation expands the evptests by including the test cases from
RFC 5794 and ARIA official site rather than providing an individual
test case. Support for ARIA has been integrated into the command line
applications, but not TLS. Implemented modes are CBC, CFB1, CFB8,
CFB128, CTR, ECB and OFB128.
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2337)
Prevent undefined behavior in CRYPTO_cbc128_encrypt: calling this function
with the 'len' parameter being 0 would result in a memcpy where the source
and destination parameters are the same, which is undefined behavior.
Do same for AES_ige_encrypt.
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2671)
