As was done for ciphers, supported groups, and EC point formats in
https://github.com/openssl/openssl/pull/9162, only write the negotiated
SNI hostname value to the session object when not resuming, even for
TLS 1.3 resumptions. Otherwise, when using a stateful session cache
(as is done by default when 0-RTT data is enabled), we can have multiple
SSLs active using the same in-memory session object, which leads to
double-frees and similar race conditions in the SNI handler prior
to this commit.
Fortunately, since draft-ietf-tls-tls13-22, there is no requirement
that the SNI hostname be preserved across TLS 1.3 resumption, and thus
not a need to continually update the session object with the "current"
value (to be used when producing session tickets, so that the subsequent
resumption can be checked against the current value). So we can just
relax the logic and only write to the session object for initial handshakes.
This still leaves us in a somewhat inconsistent state, since if the SNI value
does change across handshakes, the session object will continue to record
the initial handshake's value, even if that bears no relation to the
current handshake. The current SSL_get_servername() implementation
prefers the value from the session if s->hit, but a more complete fix
for that and related issues is underway in
https://github.com/openssl/openssl/pull/10018; there is no need to wait
for the complete fix for SNI name handling in order to close the
race condition and avoid runtime crashes.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10441)
(cherry picked from commit 2a5385511051d33be8d2b20d7669d8b1862fe510)
Check for NULL and return error if so.
This can possibly be called from apps/ca.c with a NULL argument.
Reviewed-by: Paul Yang <kaishen.yy@antfin.com>
(Merged from https://github.com/openssl/openssl/pull/10474)
(cherry picked from commit ab5c77b4766e0992751d86560193ca42b49cf316)
Add NULL check for return from pkey_asn1_meths.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10473)
(cherry picked from commit 9bada854de16bcc1a9dc199b4b352b19ab6897fc)
Verifications are public, there is no need to clear the used storage before
freeing it.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10475)
(cherry picked from commit cff7d199e0dc51ae939de5fb7702aab2a9ef30fc)
We store a secondary frame pointer info for the debugger
in the red zone.
Fixes#8853
[extended tests]
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9624)
(cherry picked from commit 9ce91035bcf7d74fe15c94650f3bc1f89b7c0f07)
In addition to 67c81ec3 which introduced this behavior in CCM mode
docs but only implemented it for AES-CCM.
Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10331)
(cherry picked from commit f7382fbbd846dd3bdea6b8c03b6af22faf0ab94f)
Conflicts:
test/recipes/30-test_evp_data/evpciph.txt
Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10454)
(cherry picked from commit 97a986f78289fef71bf8778dc4763458e983750c)
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/10388)
(cherry picked from commit 2555285fa5e4248ad4a5a0bc14ae4606443856c2)
Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10446)
(cherry picked from commit 985412f8c14853b9936852bc7ef4d9438db27b88)
Avoid conflicts with some linkers.
Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10439)
(cherry picked from commit e74b5dcf16dfd7c91d9f9a7e69c447f00d778e17)
Conflicts:
test/build.info
Signed-off-by: Joerg Schmidbauer <jschmidb@de.ibm.com>
Reviewed-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10417)
(cherry picked from commit c31950b964a2f3f7b9e6ad98076954178ee1e77d)
Add more explicit documentation about the relation between
EC_POINT_point2oct(), EC_POINT_point2hex(), EC_POINT_point2bn() and
their reverse.
In particular highlight that EC_POINT_point2oct() and
EC_POINT_oct2point() conform to, respectively, Sec. 2.3.3 and Sec. 2.3.4
of the SECG SEC 1 standard (which is the normative reference for the
already mentioned RFC 5480), highlighting with a note how this affect
the encoding/decoding of the point at infinity (which in contrast with
any other valid generic point of a curve is assigned an exceptional
fixed octet string encoding, i.e., 0x00).
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10329)
(cherry picked from commit 3cc26f2eba8a8c16ac559e68c05c094d7ea6bd8b)
Adds tests for each curve to ensure that encodings obtained through
EC_POINT_hex2point() can be fed to EC_POINT_point2hex() yielding a point
identical to the one from which the encoding is generated.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10329)
(cherry picked from commit 35ed029b5a488924890fda2487c87f664361a33b)
EC_POINT_bn2point() rejected BIGNUMs with a zero value.
This behavior indirectly caused failures when converting a point
at infinity through EC_POINT_point2hex() and then back to a point with
EC_POINT_hex2point().
With this change such BIGNUMs are treated like any other and exported to
an octet buffer filled with zero.
It is then EC_POINT_oct2point() (either the default implementation or
the custom one in group->meth->oct2point) to determine if such encoding
maps to a valid point (generally the point at infinity is encoded as
0x00).
Fixes#10258
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10329)
(cherry picked from commit d47c10875656790d146f62ac3c437db54c58dbf7)
This also removes the incorrect documentation comments by those
functions.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10403)
The resumption_label variable when CHARSET_EBCDIC was enabled, was misspelled.
Instead of evaluating to 'res binder' as expected, it evaluated to 'red binder'.
CLA: trivial
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/10396)
(cherry picked from commit 6ed12cec7216c3e81b58f5cafa41775e456feaee)
Many Windows-based GOST TLS implementations are unable to extend the
list of supported SignatureAlgorithms because of lack of the necessary
callback in Windows. So for TLS 1.2 it makes sense to imply the support
of GOST algorithms in case when the GOST ciphersuites are present.
This is a backport of #10377 to 1.1.1 branch
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10378)
Appease -Wstring-plus-int.
Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9608)
(cherry picked from commit e0249827b3)
'__builtin_strncpy' offset [275, 4095] from the object at
'direntry' is out of the bounds of referenced subobject 'd_name'
with type 'char[256]' at offset 19
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10343)
(cherry picked from commit db5cf86535b305378308c58c52596994e1ece1e6)
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/10361)
(cherry picked from commit d1c1fb2d41a627293483d832aaffcb6eca9075f9)
We also add this to our x86_64 builds on appveyor
(cherry picked from commit b4a7b4ec4acc712b1f22a83966ac986b510f25d8)
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10365)
If no connection could be made, addr_iter will eventually end up being
NULL, and if the user didn't check the returned error value, the
BIO_CONN_S_CONNECT code will be performed again and will crash.
So instead, we add a state BIO_CONN_S_CONNECT_ERROR that we enter into
when we run out of addresses to try. That state will just simply say
"error" back, until the user does something better with the BIO, such
as free it or reset it.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10333)
i2v_GENERAL_NAMES call i2v_GENERAL_NAME repeatedly as required. Each
time i2v_GENERAL_NAME gets called it allocates adds data to the passed in
stack and then returns a pointer to the stack, or NULL on failure. If
the passed in stack is itself NULL then it allocates one.
i2v_GENERAL_NAMES was not correctly handling the case where a NULL gets
returned from i2v_GENERAL_NAME. If a stack had already been allocated then
it just leaked it.
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10300)
(cherry picked from commit 45b244620a74248b46ebe1c85e86437b9641447a)
clang imposes some restrictions on the assembler code that
gcc does not.
Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10330)
(cherry picked from commit 6f93f06135cbbd36c3fe98d63717e8303a5d559b)
Conflicts:
crypto/perlasm/s390x.pm (non-existant)
crypto/s390xcpuid.pl (code to be changed non-existant)
This system services is based on FreeBSD 12's getentropy(), and is
therefore treated the same way as getentropy() with regards to amount
of entropy bits per data bit.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8926)
(cherry picked from commit 8b9896eb293a0861f0b8c191b7a278f176b729e6)
Signed-off-by: Christian Heimes <christian@python.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6553)
(cherry picked from commit 132b5facf8d681db5dfa45828d8b02f1bf5df64b)
Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10311)
(cherry picked from commit 351ba5bd27645d5b5a2bc643b2709bd30bcdf09c)
Free dukm in error handling of dh_cms_encrypt()
Fixes#10294
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Patrick Steuer <patrick.steuer@de.ibm.com>
(Merged from https://github.com/openssl/openssl/pull/10310)
(cherry picked from commit 6624e1f7b6a397948561e9cc2774f0c8af1d2c79)
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9823)
(cherry picked from commit 2aa28a1abc893fb16b99ba77e2fecb1cbc8769c7)
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10209)
(cherry picked from commit 305bf9c8668aff78e668131061f4eb088457be5f)
The `./pyca-cryptography/.travis/downstream.d` subdirectory that causes the `rm` command to fail (albeit harmlessly, but with a warning from `make` nonetheless).
>rm -f `find . -name '*.d' \! -name '.*' -print`
>rm: cannot remove './pyca-cryptography/.travis/downstream.d': Is a directory
>make: [Makefile:1910: clean] Error 1 (ignored)
Exclude directories from being matched by the `find` commands.
CLA: trivial
Reviewed-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10264)
(cherry picked from commit 38b71bd4704ee1746e862f5a7a4e170fd84a5eb0)
... if the fixed-size buffer is too small.
Fixes#9732
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Patrick Steuer <patrick.steuer@de.ibm.com>
(Merged from https://github.com/openssl/openssl/pull/10276)
(cherry picked from commit 7c2d95d47ccb3797f0da6bd4446747c6eee07b87)
This was fixed in #8321 right after the 1.1.1 was released but never
back ported to 1.1.1. Now fix it.
Issue reported from lua-openssl project.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10278)
Running s_server in WWW mode on Windows can allow a client to read files
outside the s_server directory by including backslashes in the name, e.g.
GET /..\myfile.txt HTTP/1.0
There exists a check for this for Unix paths but it is not sufficient
for Windows.
Since s_server is a test tool no CVE is assigned.
Thanks to Jobert Abma for reporting this.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10215)
(cherry picked from commit 0a4d6c67480a4d2fce514e08d3efe571f2ee99c9)
The introductory paragraph for the TLSv1.3 server side PSK documentation
is a copy & paste of the client side documentation which has not been
updated with the server side equivalent information.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10245)
(cherry picked from commit c549cb46e0d3cb4e611acafae5f919b4a8df4007)
RSA-PSS keys use the same internal structure as RSA keys but do not
allow accessing it through EVP_PKEY_get0_RSA. This commit changes that
behavior.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10217)
(cherry picked from commit 465a58b117d5a85623f3998d6fbf2fe8712a5604)
The hardcoded code points for TLSv1.3 cipher suites are used in the TLS
PSK server callback. However, they seem to have been refactored a while
ago to use tls13_aes128gcmsha256_id, so these defines are not necessary
within the s_server code anymore.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10243)
(cherry picked from commit aed8c47cbcc8a289bea433ead2effea035187260)
PR https://github.com/openssl/openssl/pull/10122 introduced changes to
the BN_gcd function and the control logic inside it accessed `g->d[0]`
irrespective of `g->top`.
When BN_add is called, in case the result is zero, `BN_zero` is called.
The latter behaves differently depending on the API compatibility level
flag: normally `g->d[0]` is cleared but in `no-deprecated` builds only
`g->top` is set to zero.
This commit uses bitwise logic to ensure that `g` is treated as zero if
`g->top` is zero, irrespective of `g->d[0]`.
Co-authored-by: Nicola Tuveri <nic.tuv@gmail.com>
(cherry picked from commit 8aca4bfe8213402c80abc06fe25121461f79128d)
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10232)
- Use `()` to qualify function names, consistently
- Limit line width to 80 chars
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10235)
for the following functions.
EC_GROUP_get_order
EC_GROUP_get_cofactor
EC_GROUP_get_curve_name
EC_GROUP_get_asn1_flag
EC_GROUP_get_point_conversion_form
EC_GROUP_get_degree
(cherry picked from commit df3d1e84b3802acffeec11d6224e8a0e33d0aa83)
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9664)
This commit adds testing and Known Answer Tests (KATs) to OpenSSL for
the `BN_gcd` function.
(cherry picked from commit b75d6310857bc44ef2851bde68a1979c18bb4807)
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10122)
This commit replaces the current `BN_gcd` function with a constant-time
GCD implementation.
(cherry picked from commit f3c4adfc7eb13e9eff514039b4c60b457bdba433)
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10122)
This commit aims at refactoring the `BN_rshift` by making it a wrapper
around `bn_rshift_fixed_top`, in order to match the current design of
`BN_lshift`, as suggested in the discussion at
https://github.com/openssl/openssl/pull/10122#discussion_r332474277 .
As described in the code, by refactoring this function, `BN_rshift`
provides a constant-time behavior for sufficiently[!] zero-padded inputs
under the following assumptions: `|n < BN_BITS2|` or `|n / BN_BITS2|`
being non-secret.
Notice that `BN_rshift` returns a canonical representation of the
BIGNUM, if a `fixed_top` representation is required, the caller should
call `bn_rshift_fixed_top` instead.
(cherry picked from commit 8eba6de59e2b06f23c214344423a5a618d1c9ffd)
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10196)
