TLSv1.3 includes an extra byte after the payload for the content type.
We should incorporate that in the calculation of the default buffer length.
Fixes#23015
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/23021)
If you decrypt a random input using RSAES-PKCS-v1_5, then there is a
non-negligible chance that the result will look like a valid plaintext
(that is why RSAES-PKCS-v1_5 shouldn't be used anymore). This was the
cause of an intermittent failure in a test that did a cms-encrypt
operation targetting multiple recipients.
The failure happened during key-only decrypt. The recipient decrypts
every RSA ciphertext -- only one is supposed to decrypt successfully,
which would reveal the right content-key. Occassionally, more than
one decrypted successfully.
Update the test by specifying the recipient cert in the decrypt op
(this avoids looping over all RSA ciphertexts).
Add a new test to get coverage for key-only decrypt, but use RSA-OAEP
during the encrypt op.
Fixes https://github.com/openssl/project/issues/380
Testing:
$ make TESTS='test_cms' test
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/23055)
In commit d7c0fc5b1a we removed L_ENDIAN
definition for guessed linux64-loongarch64 as it had caused an
inconsistency between configurations with and without explicit
specifying linux64-loongarch64. Now add it back to the proper location.
Unlike MIPS or RISC-V, LoongArch is always little-endian [1].
By the way, change "LOONGARCH" to "LoongArch" in a comment as LOONGARCH
should only appear in the identifiers of macros, constants, etc.
[1]:https://loongson.github.io/LoongArch-Documentation/LoongArch-Vol1-EN.html#endian
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/23064)
In the command `nmake TEST='foo' test`, on Windows the runner
will look for test `'foo'` and complain about the test not being found
(due to the extraneous single quotes), whereas with `nmake TEST="foo" test`,
the test `foo` will be correctly found.
CLA: trivial
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/23059)
EVP_MAC_CTX_get_mac_size() cannot be called on older
unfixed versions before EVP_MAC_init().
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/23006)
gh_gen_type_common_set_params looks up a dsa contexts gen_type using
name2id, but if it returns error, we inadvertently set gctx->gen_type to
-1, which is an invalid value, which may lead to improper behavior in
future calls, in the event that said future calls preform an operation
of the form;
if (gen_type == <VALID VALUE>) {
do_stuff
else {
do_other_stuff
}
Technically it is not correct to continue with the operations on the
gen context after failed parameters setting but this makes it more
predictable.
Fix it by assigning the result of a lookup to a stack variable, and only
update gctx->gen_value if the lookup returns a non-failing value
In leiu of testing this specific case, also add an ossl_assert in dsa_gen
to validate the gen_val input prior to continuing, should other code
points attempt to do the same thing
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22991)
gh_gen_type_common_set_params looks up a dh contexts gen_type using
name2id, but if it returns error, we set gctx->gen_type to -1, which
is an invalid value, which may lead to undefined behavior in
future calls, in the event that said future calls preform an operation
of the form;
if (gen_type == <VALID VALUE>) {
do_stuff
else {
do_other_stuff
}
Technically it is not correct to continue with the operations on the
gen context after failed parameters setting but this makes it more
predictable.
Fix it by assigning the result of a lookup to a stack variable, and only
update gctx->gen_value if the lookup returns a non-failing value
In leiu of testing this specific case, also add an ossl_assert in dh_gen
to validate the gen_val input prior to continuing, should other code
points attempt to do the same thing
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22991)
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22911)
In the event that a config file contains this sequence:
=======
openssl_conf = openssl_init
config_diagnostics = 1
[openssl_init]
oid_section = oids
[oids]
testoid1 = 1.2.3.4.1
testoid2 = A Very Long OID Name, 1.2.3.4.2
testoid3 = ,1.2.3.4.3
======
The leading comma in testoid3 can cause a heap buffer overflow, as the
parsing code will move the string pointer back 1 character, thereby
pointing to an invalid memory space
correct the parser to detect this condition and handle it by treating it
as if the comma doesn't exist (i.e. an empty long oid name)
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22957)
At some point the asn1parse applet was changed to default the inform to
PEM, and defalt input file to stdin. Doing so broke the -genstr|conf options,
in that, before we attempt to generate an ASN1 block from the provided
genstr string, we attempt to read a PEM input from stdin. As a result,
this command:
openssl asn1parse -genstr OID:1.2.3.4
hangs because we are attempting a blocking read on stdin, waiting for
data that never arrives
Fix it by giving priority to genstr|genconf, such that, if set, will just run
do_generate on that string and exit
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22957)
"=over 1" is too small. Use "=over 2" so that list items are
displayed correctly in the generated man-page.
You can check the man-page using the following command:
cd doc && pod2man man3/OSSL_PARAM_int.pod | man /dev/stdin
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/22974)
This typedef is already created in aes_local.h as `typedef uint64_t u64;`.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22969)
In param_build.c, the functions OSSL_PARAM_BLD_push_utf8_string() and
OSSL_PARAM_BLD_push_utf8_ptr() use strlen() to compute the length of
the string when bsize is zero. However, the size_t returned by
strlen() might be too large (it is stored in an intermediate "int"),
so check for that.
There are analogous functions in params.c, but they do not use an
intermediate "int" to store the size_t returned by strlen(). So there
is some inconsistency between the implementations.
Credit to Viktor D and Tomas M for spotting these missing checks.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22967)
The ASN1_OBJECT otmp was leaked if X509_VERIFY_PARAM_add0_policy fails.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22922)
The OPENSSL_DIR_end was missing in case of error.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22920)
Ensure we test the case where the port value is empty in the URL.
Reviewed-by: Todd Short <todd.short@me.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/22961)
sscanf can return -1 on an empty input string. We need to appropriately
handle such an invalid case.
The instance in OSSL_HTTP_parse_url could cause an uninitialised read of
sizeof(unsigned int) bytes (typically 4). In many cases this uninit read
will immediately fail on the following check (i.e. if the read value
>65535).
If the top 2 bytes of a 4 byte unsigned int are zero then the value will
be <=65535 and the uninitialised value will be returned to the caller and
could represent arbitrary data on the application stack.
The OpenSSL security team has assessed this issue and consider it to be
a bug only (i.e. not a CVE).
Reviewed-by: Todd Short <todd.short@me.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/22961)
Test that input value of 1 for p is treated as an error
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/22960)
If p is set to 1 when calling BN_GF2m_mod_inv then an infinite loop will
result. Calling this function set 1 when applications call this directly
is a non-sensical value - so this would be considered a bug in the caller.
It does not seem possible to cause OpenSSL internal callers of
BN_GF2m_mod_inv to call it with a value of 1.
So, for the above reasons, this is not considered a security issue.
Reported by Bing Shi.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/22960)
We remove a function that was left behind and is no longer called after the
record layer refactor
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/22864)
If an ENGINE has been loaded after the SSL_CTX has been created then
the cipher we have cached might be provider based, but the cipher we
actually end up using might not be. Don't try to set provider params on
a cipher that is actually ENGINE based.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/22864)
Confirm that using an ENGINE works as expected with TLS even if it is
loaded late (after construction of the SSL_CTX).
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/22864)
This fix removes explicit support for the SPT threading model in configurations.
This also reverts commit f63e1b48ac that were
required for SPT but broke other models.
Fixes: #22798
Signed-off-by: Randall S. Becker <randall.becker@nexbridge.ca>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22807)
Since the gen->type will not be set in a2i_GENERAL_NAME
the gen->d.otherName will not be automatically
cleaned up by GENERAL_NAME_free.
Also fixed a similar leak in a2i_GENERAL_NAME,
where ASN1_STRING_set may fail but gen->d.ia5
will not be automatically cleaned up.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22996)
The EVP interface explicitly allows in-place encryption/decryption,
but this fact is just 'partially' documented in `EVP_EncryptUpdate(3)`
(pun intended): the manual page mentions only operation failure in
case of 'partial' overlaps. This is not even correct, because
the check for partially overlapping buffers is only implemented
in legacy code paths.
Currently, in-place encryption/decryption is only documented for
RSA (`RSA_public_encrypt(3)`) and DES (`DES_ecb_encrypt(3)`), as
well as in the provider interface (`provider-cipher(7)`).
This commit amends `EVP_EncryptUpdate(3)` and `provider-cipher(7)`
to make the front-end and back-end documentation consistent.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
(Merged from https://github.com/openssl/openssl/pull/22875)
Test the next arithmetic operation to safely determine if adding the
next digit in the passed property string will overflow
Also, noted a bug in the parse_hex code. When parsing non-digit
characters (i.e. a-f and A-F), we do a tolower conversion (which is
fine), and then subtract 'a' to get the hex value from the ascii (which
is definately wrong). We should subtract 'W' to convert tolower
converted hex digits in the range a-f to their hex value counterparts
Add tests to test_property_parse_error to ensure overflow checks work
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
(Merged from https://github.com/openssl/openssl/pull/22874)
This also adds the configuration options 'enable-quic'.
Fixes#22907
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/22908)
For some reason, it was added to the Unix and VMS build templates, but
Windows was forgotten.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/22947)
Like in #17345, evp_extra_test links libcrypto statically, but also has
a dynamic/shared load via the legacy provider, which leads to ambiguous
behavior in evp_extra_test on some platforms, usually a crash (SIGSEGV)
on exit via the atexit handlers. Statically link the legacy provider to
avoid this.
Fixes#22819
Helped-by: Neil Horman <nhorman@openssl.org>
Helped-by: Tomas Mraz <tomas@openssl.org>
Signed-off-by: Randall S. Becker <randall.becker@nexbridge.ca>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22905)
This function can be called during OPENSSL_cleanup() when
the cache was already flushed and deallocated.
Fixes#22939
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22941)
