Some (non-AEAD) ciphers require an IV to be used. Always pass a (dummy) IV
when setting the key. It is ignored by ciphers that do not use an IV.
Commit 607a46d003 corrected the use of AEAD
ciphers, but removed the IV from being passed to EVP_CipherInit_ex() for
non-AEAD ciphers.
Fixes: 607a46d003
Signed-off-by: Ingo Franzki <ifranzki@linux.ibm.com>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Hugo Landau <hlandau@devever.net>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26346)
Once lcov is updated to 2.2 version or later, it could be dropped.
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
(Merged from https://github.com/openssl/openssl/pull/26381)
Simplify some decoder/encoder internals to facilitate upcoming support
for ML-KEM and ML-DSA.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26355)
Builds may be configured with CC or CFLAGS containing space and
double quotes. In particular on Windows, this may lead to passing
more than two arguments into mkbuildinf.pl.
In addition, backslashes must be escaped for constructing the C string.
Fixes#26253.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26315)
It was noticed recently that the enum for QUIC encryption levels doesn't
match the ordering that is outlined in the RFC. RFC 9000 s. 12.2 and
RFC 9002 s 14.4.1 indicate that encryption level ordering is
INITIAL/0RTT/HANDSHAKE/1RTT, but our enum is in the order
INITAL/HANDSHAKE/0RTT/1RTT.
Our enum isn't a direct wire translation, so as long as the wire->enum
mapping done in ossl_quic_pkt_type_to_enc_level is done consistently it
ideally wouldn't matter, but because we do coalescing in
ossl_quic_tx_packetiser_generate by iterating through all the values in
the enum, its possible we may coalesce in the wrong order when we do
start implementing 0RTT support.
Fix it by adjusting the enum properly to match the RFC order. This also
necessitates and adjustment to the archetypes array, which is a two
dimensional array indexed by encryption level and frame archetype
(PROBE/NORMAL/ACK ONLY). Moving the 0RTT enc level to index 1 requires
moving the (formerly) index 2 0RTT array row to be at index 1.
Fixes#26324
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Saša Nedvědický <sashan@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26334)
When default_context_inited is set to false we return NULL instead of
the global default context.
Fixes#25442
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26319)
Signed-off-by: Герман Семенов <GermanAizek@yandex.ru>
Signed-off-by: Herman Semenov <GermanAizek@yandex.ru>
Reviewed-by: Hugo Landau <hlandau@devever.net>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/23377)
UBSan complains about functions being called with incorrect signatures.
Relates to #22896
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26318)
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26321)
p8 is allocated using EVP_PKEY2PKCS8(), but when PKCS8_add_keyusage()
fails this memory is not freed. Fix this by adding a call to
PKCS8_PRIV_KEY_INFO_free().
Reviewed-by: Frederik Wedel-Heinen <fwh.openssl@gmail.com>
Reviewed-by: Hugo Landau <hlandau@devever.net>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25818)
in test/wpackettest.c:593:18: runtime error: load of misaligned address
for type 'uint64_t', which requires 8 byte alignment.
Reviewed-by: Saša Nedvědický <sashan@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26269)
recent gcc versions can optimize the memory leak away,
avoid that by declaring the lost variable to be volatile.
Reviewed-by: Saša Nedvědický <sashan@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26269)
Here the undefined value "npa" passed to a function
WPACKET_sub_memcpy_u16(pkt, npa, npalen).
However the value is not really used, because "npalen" is zero,
but the call statememt itself is considered an invalid operation
by the new sanitizer.
The original sanitizer error report was:
==49175==WARNING: MemorySanitizer: use-of-uninitialized-value
#0 0x55a276b29d6f in tls_construct_stoc_next_proto_neg /home/runner/work/openssl/openssl/ssl/statem/extensions_srvr.c:1518:21
#1 0x55a276b15d7d in tls_construct_extensions /home/runner/work/openssl/openssl/ssl/statem/extensions.c:909:15
#2 0x55a276b513dc in tls_construct_server_hello /home/runner/work/openssl/openssl/ssl/statem/statem_srvr.c:2471:10
#3 0x55a276b2e160 in write_state_machine /home/runner/work/openssl/openssl/ssl/statem/statem.c:896:26
#4 0x55a276b2e160 in state_machine /home/runner/work/openssl/openssl/ssl/statem/statem.c:490:21
#5 0x55a276b2f562 in ossl_statem_accept /home/runner/work/openssl/openssl/ssl/statem/statem.c:309:12
#6 0x55a276a9f867 in SSL_do_handshake /home/runner/work/openssl/openssl/ssl/ssl_lib.c:4890:19
#7 0x55a276a9f605 in SSL_accept /home/runner/work/openssl/openssl/ssl/ssl_lib.c:2169:12
#8 0x55a276a3d4db in create_bare_ssl_connection /home/runner/work/openssl/openssl/test/helpers/ssltestlib.c:1281:24
#9 0x55a276a3d7cb in create_ssl_connection /home/runner/work/openssl/openssl/test/helpers/ssltestlib.c:1350:10
#10 0x55a276a64c0b in test_npn /home/runner/work/openssl/openssl/test/sslapitest.c:12266:14
#11 0x55a276b9fc20 in run_tests /home/runner/work/openssl/openssl/test/testutil/driver.c:377:21
#12 0x55a276ba0b10 in main /home/runner/work/openssl/openssl/test/testutil/main.c:31:15
Reviewed-by: Saša Nedvědický <sashan@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26269)
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Neil Horman <nhorman@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26322)
OpenSSL currently does not support encryption with originator flag so it
should fail nicely instead of segfaulting.
Reviewed-by: Hugo Landau <hlandau@devever.net>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26014)
Reviewed-by: Hugo Landau <hlandau@devever.net>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26014)
As suggested in https://github.com/animetosho/md5-optimisation?tab=readme-ov-file#dependency-shortcut-in-g-function,
we can delay the dependency on 'x' by recognizing that ((x & z) | (y & ~z))
is equivalent to ((x & z) + (y + ~z)) in this scenario, and we can perform
those additions independently, leaving our dependency on x to the final
addition. This speeds it up around 5% on both platforms.
Signed-off-by: Oli Gillespie <ogillesp@amazon.com>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Hugo Landau <hlandau@devever.net>
(Merged from https://github.com/openssl/openssl/pull/25737)
For details about this file format see:
https://floss.fund/funding-manifest/
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Hugo Landau <hlandau@devever.net>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26247)
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26295)
In ossl_property_merge() we can drop the realloc because it just makes
the allocation smaller.
In quic-hq-interop.c we check the realloc result.
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26244)
This changes the alert according to RFC 8446.
Fixes: #25402
Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25547)
When BIO_parse_hostserv() fails it may still have allocated memory, yet
this memory is not freed. Fix it by jumping to the err label.
Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25817)
Also avoid leak if stack push fails.
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/26230)
Recent test additions have increased the number of jobs spawned by the
interop runner test which exceeds the maximum allowed.
This occurs because the matrix expands to:
7 server elements
6 client elements
7 tests
2 test steps (client interop and server interop
Because of how github ci does matrix expansion, this results in
2 * 7 * 7 * 6 = 588
But most of those are invalid because each of the 2 steps only considers
either the client or server elements, and so get rerun multiple times
Alter the steps to be individual jobs, each with their own reduced
matrix to only run each relevant test once, limiting our job count to
at most 49 jobs.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
Reviewed-by: Saša Nedvědický <sashan@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26275)
This makes `ikmlen` have a length of at least `Nsk`.
Closes#26213
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26254)
Add a test to check that if the user reduces the default TLS security level
at configure time, then the tests still pass.
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26256)
The compile time default TLS security level can be changed if the user
sets `-DOPENSSL_TLS_SECURITY_LEVEL=x` at configure time (where "x" is some
number, typically 0 or 1).
Since OpenSSL 3.4 tests are failing if the default security level is 0. We
fix the tests for this case.
Fixes#26255
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26256)
