OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
Two typo's on #endif comments fixed:
OPENSSL_NO_ECB fixed to OPENSSL_NO_OCB
OPENSSL_NO_HW_SureWare fixed to OPENSSL_NO_HW_SUREWARE
Reviewed-by: Richard Levitte <levitte@openssl.org>
BN_init and BN_RECP_CTX_init are deprecated and are not exported
from shared libraries on some platforms (e.g. Windows) convert
bntest to use BN_new and BN_RECP_CTX_new instead.
Reviewed-by: Matt Caswell <matt@openssl.org>
Update the X509v3 name parsing to allow multiple xn-- international
domain name indicators in a name. Previously, only allowed one at
the beginning of a name, which was wrong.
Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
This commit removes NCR, Tandem, Cray.
Regenerates TABLE.
Removes another missing BEOS fluff.
The last platform remaining on this ticket is WIN16.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Add INSTALLDIRS variable, list of directories where things get
installed. Change install_html_docs to use perl mkdir-p script.
Add uninstall, uninstall_sw, uninstall_docs, uninstall_html_docs
to Makefile.org. The actions of these targets were figured out
by "inverting" the install target.
Recurse into subdirs to do uninstall as needed. Added uninstall
targets whose actions were similarly figured out by "inverting"
the install target.
Also remove some 'space before tab' complaints in Makefile.org
Reviewed-by: Tim Hudson <tjh@openssl.org>
ssl3_setup_buffers or pqueue_insert fail. The former will fail if there is a
malloc failure, whilst the latter will fail if attempting to add a duplicate
record to the queue. This should never happen because duplicate records should
be detected and dropped before any attempt to add them to the queue.
Unfortunately records that arrive that are for the next epoch are not being
recorded correctly, and therefore replays are not being detected.
Additionally, these "should not happen" failures that can occur in
dtls1_buffer_record are not being treated as fatal and therefore an attacker
could exploit this by sending repeated replay records for the next epoch,
eventually causing a DoS through memory exhaustion.
Thanks to Chris Mueller for reporting this issue and providing initial
analysis and a patch. Further analysis and the final patch was performed by
Matt Caswell from the OpenSSL development team.
CVE-2015-0206
Reviewed-by: Dr Stephen Henson <steve@openssl.org>
Fix to prevent use of DH client certificates without sending
certificate verify message.
If we've used a client certificate to generate the premaster secret
ssl3_get_client_key_exchange returns 2 and ssl3_get_cert_verify is
never called.
We can only skip the certificate verify message in
ssl3_get_cert_verify if the client didn't send a certificate.
Thanks to Karthikeyan Bhargavan for reporting this issue.
CVE-2015-0205
Reviewed-by: Matt Caswell <matt@openssl.org>
of the crash due to p being NULL. Steve's fix prevents this situation from
occuring - however this is by no means obvious by looking at the code for
dtls1_get_record. This fix just makes things look a bit more sane.
Reviewed-by: Dr Stephen Henson <steve@openssl.org>
Show only the #define, not the values, in BIO_f_buffer. Data
abstraction and we can remove a "see also" entry.
Remove internal forward reference to NOTES in EVP_EncryptInit; just
say "see below" as we do in the other pages.
Add missing (3) in pem.pod so the L<> entry is consistent.
Fix entry to point to the "master" page, not the symlink'd one.
Reviewed-by: Matt Caswell <matt@openssl.org>
Fix memory leak by freeing up saved_message.data if it is not NULL.
PR#3489
Reviewed-by: Stephen Henson <steve@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
OpenSSL clients would tolerate temporary RSA keys in non-export
ciphersuites. It also had an option SSL_OP_EPHEMERAL_RSA which
enabled this server side. Remove both options as they are a
protocol violation.
Thanks to Karthikeyan Bhargavan for reporting this issue.
(CVE-2015-0204)
Reviewed-by: Matt Caswell <matt@openssl.org>
Fix bug where an OpenSSL client would accept a handshake using an
ephemeral ECDH ciphersuites with the server key exchange message omitted.
Thanks to Karthikeyan Bhargavan for reporting this issue.
CVE-2014-3572
Reviewed-by: Matt Caswell <matt@openssl.org>
By using non-DER or invalid encodings outside the signed portion of a
certificate the fingerprint can be changed without breaking the signature.
Although no details of the signed portion of the certificate can be changed
this can cause problems with some applications: e.g. those using the
certificate fingerprint for blacklists.
1. Reject signatures with non zero unused bits.
If the BIT STRING containing the signature has non zero unused bits reject
the signature. All current signature algorithms require zero unused bits.
2. Check certificate algorithm consistency.
Check the AlgorithmIdentifier inside TBS matches the one in the
certificate signature. NB: this will result in signature failure
errors for some broken certificates.
3. Check DSA/ECDSA signatures use DER.
Reencode DSA/ECDSA signatures and compare with the original received
signature. Return an error if there is a mismatch.
This will reject various cases including garbage after signature
(thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
(negative or with leading zeroes).
CVE-2014-8275
Reviewed-by: Emilia Käsper <emilia@openssl.org>
This change documents the world as-is, by turning all warnings on,
and then turning warnings that trigger off again.
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
This facilitates "universal" builds, ones that target multiple
architectures, e.g. ARMv5 through ARMv7. See commentary in
Configure for details.
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
MS Server gated cryptography is obsolete and dates from the time of export
restrictions on strong encryption and is only used by ancient versions of
MSIE.
Reviewed-by: Matt Caswell <matt@openssl.org>
When parsing ClientHello clear any existing extension state from
SRP login and SRTP profile.
Thanks to Karthikeyan Bhargavan for reporting this issue.
Reviewed-by: Matt Caswell <matt@openssl.org>
We need this for the freebsd kernel with glibc as used in the Debian kfreebsd
ports. There shouldn't be a problem defining this on systems not using glibc.
Reviewed-by: Richard Levitte <levitte@openssl.org>
This commit removes DG-UX.
It also flushes out some left-behinds in config.
And regenerates TABLE from Configure (hadn't been done in awhile).
Reviewed-by: Richard Levitte <levitte@openssl.org>
* adds links to various related documents.
* fixes a few typos.
* rewords a few sentences.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
Return an error code for I/O errors instead of an assertion failure.
PR#3470
Reviewed-by: Stephen Henson <steve@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Also introduce OPENSSL_USE_DEPRECATED. If OPENSSL_NO_DEPRECATED is
defined at config stage then OPENSSL_USE_DEPRECATED has no effect -
deprecated functions are not available.
If OPENSSL_NO_DEPRECATED is not defined at config stage then
applications must define OPENSSL_USE_DEPRECATED in order to access
deprecated functions.
Also introduce compiler warnings for gcc for applications using
deprecated functions
Reviewed-by: Rich Salz <rsalz@openssl.org>
According to X6.90 null, object identifier, boolean, integer and enumerated
types can only have primitive encodings: return an error if any of
these are received with a constructed encoding.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
Causes more problems than it fixes: even though error codes
are not part of the stable API, several users rely on the
specific error code, and the change breaks them. Conversely,
we don't have any concrete use-cases for constant-time behaviour here.
This reverts commit 4aac102f75.
Reviewed-by: Andy Polyakov <appro@openssl.org>
Various build fixes, mostly uncovered by clang's unused-const-variable
and unused-function errors.
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
(cherry picked from commit 0e1c318ece)
From BoringSSL
- Send an alert when the client key exchange isn't correctly formatted.
- Reject overly short RSA ciphertexts to avoid a (benign) out-of-bounds memory access.
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Invalid zero-padding in the divisor could cause a division by 0.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(cherry picked from commit a43bcd9e96)
The client_version needs to be preserved for the RSA key exchange.
This change also means that renegotiation will, like TLS, repeat the old
client_version rather than advertise only the final version. (Either way,
version change on renego is not allowed.) This is necessary in TLS to work
around an SChannel bug, but it's not strictly necessary in DTLS.
(From BoringSSL)
Reviewed-by: Emilia Käsper <emilia@openssl.org>
for dummytest if gost is compiled out, since the name of the test is not
standard (dummytest segfaults). Also the old name caused problems for git
because the executable was not in the .gitignore file
Reviewed-by: Emilia Käsper <emilia@openssl.org>
Don't remove c_rehash that wasn't created by make; this script
is created by configure.
This fix brought to you by the letter "f" and
Reviewed-by: Emilia Kasper <emilia@openssl.org>
Doing 'config ; make clean' broke because clean removed
header files that normal build didn't create. So don't
remove those files. Hopefully will be better addressed by
Geoff's no-symlinks patch.
Reviewed-by: Matt Caswell <matt@openssl.org>
The temporary variable causes unused variable warnings in opt mode with clang,
because the subsequent assert is compiled out.
Reviewed-by: Rich Salz <rsalz@openssl.org>
In EVP_EncryptInit remove duplicate mention of EVP_idea_cbc()
In EVP_PKEY_CTX_ctrl.pod remove EVP_PKEY_get_default_digest_nid
since it is documented elsewhere.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Some Makefiles had actions for "dclean" that really belonged
to the "clean" target. This is wrong because clean ends up,
well, not really cleaning everything.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Exported headers shouldn't be included as "foo.h" by code from the same
module, it should only do so for module-internal headers. This is
because the symlinking of exported headers (from include/openssl/foo.h
to crypto/foo/foo.h) is being removed, and the exported headers are
being moved to the include/openssl/ directory instead.
Change-Id: I4c1d80849544713308ddc6999a549848afc25f94
Signed-off-by: Geoff Thorpe <geoff@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
OPENSSL_FIPSCANISTER is only set if the fips module is being built
(as opposed to being used). Since the fips module wont be built in
master this is redundant.
Reviewed-by: Tim Hudson <tjh@openssl.org>
The fips test utilities are only build if an FIPS module is being
built from source. As this isn't done in master these are redundant.
Reviewed-by: Tim Hudson <tjh@openssl.org>
Odd-length lists should be rejected everywhere upon parsing. Nevertheless,
be extra careful and add guards against off-by-one reads.
Also, drive-by replace inexplicable double-negation with an explicit comparison.
Reviewed-by: Matt Caswell <matt@openssl.org>
The Supported Elliptic Curves extension contains a vector of NamedCurves
of 2 bytes each, so the total length must be even. Accepting odd-length
lists was observed to lead to a non-exploitable one-byte out-of-bounds
read in the latest development branches (1.0.2 and master). Released
versions of OpenSSL are not affected.
Thanks to Felix Groebert of the Google Security Team for reporting this issue.
Reviewed-by: Matt Caswell <matt@openssl.org>
Always use goto err on failure and call ssl_cert_free() on the error path so all
fields and "ret" itself are freed
Signed-off-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Richard Levitte <levitte@openssl.org>
and UDP header) when setting an mtu. This constant is not always correct (e.g.
if using IPv6). Use the new DTLS_CTRL functions instead.
Reviewed-by: Tim Hudson <tjh@openssl.org>
at least the minimum or it will fail.
There were some instances in dtls1_query_mtu where the final mtu can end up
being less than the minimum, i.e. where the user has set an mtu manually. This
shouldn't be allowed. Also remove dtls1_guess_mtu that, despite having
logic for guessing an mtu, was actually only ever used to work out the minimum
mtu to use.
Reviewed-by: Tim Hudson <tjh@openssl.org>
and instead use the value provided by the underlying BIO. Also provide some
new DTLS_CTRLs so that the library user can set the mtu without needing to
know this constant. These new DTLS_CTRLs provide the capability to set the
link level mtu to be used (i.e. including this IP/UDP overhead). The previous
DTLS_CTRLs required the library user to subtract this overhead first.
Reviewed-by: Tim Hudson <tjh@openssl.org>
used with no explanation. Some of this was introduced as part of RT#1929. The
value 28 is the length of the IP header (20 bytes) plus the UDP header (8
bytes). However use of this constant is incorrect because there may be
instances where a different value is needed, e.g. an IPv4 header is 20 bytes
but an IPv6 header is 40. Similarly you may not be using UDP (e.g. SCTP).
This commit introduces a new BIO_CTRL that provides the value to be used for
this mtu "overhead". It will be used by subsequent commits.
Reviewed-by: Tim Hudson <tjh@openssl.org>
mtu that we have received is not less than the minimum. If its less it uses the
minimum instead. The second call to query the mtu does not do that, but
instead uses whatever comes back. We have seen an instance in RT#3592 where we
have got an unreasonably small mtu come back. This commit makes both query
checks consistent.
Reviewed-by: Tim Hudson <tjh@openssl.org>
automatically updated, and we should use the one provided instead.
Unfortunately there are a couple of locations where this is not respected.
Reviewed-by: Tim Hudson <tjh@openssl.org>
RT#3592 provides an instance where the OPENSSL_assert that this commit
replaces can be hit. I was able to recreate this issue by forcing the
underlying BIO to misbehave and come back with very small mtu values. This
happens the second time around the while loop after we have detected that the
MTU has been exceeded following the call to dtls1_write_bytes.
Reviewed-by: Tim Hudson <tjh@openssl.org>
Previously, state variant was not advanced, which resulted in state
being stuck in the st1 variant (usually "_A").
This broke certificate callback retry logic when accepting connections
that were using SSLv2 ClientHello (hence reusing the message), because
their state never advanced to SSL3_ST_SR_CLNT_HELLO_C variant required
for the retry code path.
Reported by Yichun Zhang (agentzh).
Signed-off-by: Piotr Sikora <piotr@cloudflare.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
The current documentation contains a bunch of spelling and grammar mistakes. I also
found it hard to understand some paragraphs, so here is my attempt to improve its
readability.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Workaround for NetWare CodeWarrior compiler which doesn't properly lookup
includes when in same directory as the C file which includes it.
PR#3569
Reviewed-by: Stephen Henson <steve@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
When using the -xcert option to test certificate validity print out
if we pass Suite B compliance. We print out "not tested" if we aren't
in Suite B mode.
Reviewed-by: Matt Caswell <matt@openssl.org>
In keygen, return KEY_SIZE_TOO_SMALL not INVALID_KEYBITS.
** I also increased the minimum from 256 to 512, which is now
documented in CHANGES file. **
Reviewed-by: Matt Caswell <matt@openssl.org>
Fix CONF_load_modules to CONF_modules_load.
Document that it calls exit.
Advise against using it now.
Add an error print to stderr.
Reviewed-by: Matt Caswell <matt@openssl.org>
the session's version (server).
See also BoringSSL's commit bdf5e72f50e25f0e45e825c156168766d8442dde.
Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
ECDH_compute_key is silently ignored and the KDF is run on duff data
Thanks to github user tomykaira for the suggested fix.
Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
once the ChangeCipherSpec message is received. Previously, the server would
set the flag once at SSL3_ST_SR_CERT_VRFY and again at SSL3_ST_SR_FINISHED.
This would allow a second CCS to arrive and would corrupt the server state.
(Because the first CCS would latch the correct keys and subsequent CCS
messages would have to be encrypted, a MitM attacker cannot exploit this,
though.)
Thanks to Joeri de Ruiter for reporting this issue.
Reviewed-by: Matt Caswell <matt@openssl.org>
The server must send a NewSessionTicket message if it advertised one
in the ServerHello, so make a missing ticket message an alert
in the client.
An equivalent change was independently made in BoringSSL, see commit
6444287806d801b9a45baf1f6f02a0e3a16e144c.
Reviewed-by: Matt Caswell <matt@openssl.org>
The client sends a session ID with the session ticket, and uses
the returned ID to detect resumption, so we do not need to peek
at handshake messages: s->hit tells us explicitly if we're resuming.
An equivalent change was independently made in BoringSSL, see commit
407886f589cf2dbaed82db0a44173036c3bc3317.
Reviewed-by: Matt Caswell <matt@openssl.org>
This ensures that it's zeroed even if the SSL object is reused
(as in ssltest.c). It also ensures that it applies to DTLS, too.
Reviewed-by: Matt Caswell <matt@openssl.org>
If no keyfile has been specified use the certificate file instead.
Fix typo: we need to check the chain is not NULL, not the chain file.
Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 786370b1b0)
When no-ssl3 is set only make SSLv3 disabled by default. Retain -ssl3
options for s_client/s_server/ssltest.
When no-ssl3-method is set SSLv3_*method() is removed and all -ssl3
options.
We should document this somewhere, e.g. wiki, FAQ or manual page.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
Don't send or parse any extensions other than RI (which is needed
to handle secure renegotation) for SSLv3.
Reviewed-by: Matt Caswell <matt@openssl.org>
The supported signature algorithms extension needs to be processed before
the certificate to use is decided and before a cipher is selected (as the
set of shared signature algorithms supported may impact the choice).
Reviewed-by: Matt Caswell <matt@openssl.org>
(cherry picked from commit 56e8dc542b)
Conflicts:
ssl/ssl.h
ssl/ssl_err.c
Don't attempt to access msg structure if recvmsg returns an error.
PR#3483
Reviewed-by: Stephen Henson <steve@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
These correspond to targets of the same name in test/Makefile that clash when
using the single-makefile build method using GitConfigure and GitMake.
Change-Id: If7e900c75f4341b446608b6916a3d76f202026ea
Signed-off-by: Mike Bland <mbland@acm.org>
Signed-off-by: Geoff Thorpe <geoff@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Before this change, variables for which a '=' appeared in the assignment would
be parsed as the entire string up until the final '='. For example:
BUILD_CMD=shlib_target=; if [ -n "$(SHARED_LIBS)" ]; then \
would result in the variable name "BUILD_CMD=shlib_target". This doesn't
appear to harm the current generation of MINFO, but creates problems for other
Makefile-related work I'm attempting.
Change-Id: I1f3a606d67fd5464bb459e8f36c23b3e967b77e1
Signed-off-by: Mike Bland <mbland@acm.org>
Signed-off-by: Geoff Thorpe <geoff@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
This fixes the errors when trying to assemble .s files using GitMake on OS X.
Change-Id: I2221f558619302d22e0c57d7203173d634155678
Signed-off-by: Mike Bland <mbland@acm.org>
Signed-off-by: Geoff Thorpe <geoff@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Print out more details of the conection in ssltest specifically:
server certificate curve name for EC, server temporary key (if any)
and peer signing digest.
Reviewed-by: Matt Caswell <matt@openssl.org>
Add command line support for SSL_CONF: server side arguments are
prefixed by -s_ (e.g. -s_no_ssl3) and client side with -c_.
Reviewed-by: Matt Caswell <matt@openssl.org>
If the hash or public key algorithm is "undef" the signature type
will receive special handling and shouldn't be included in the
cross reference table.
Reviewed-by: Tim Hudson <tjh@openssl.org>
Out is the buffer which needs to contain at least inl + cipher_block_size - 1 bytes. Outl
is just an int*.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
This doesn't really fix the datarace but changes it so it can only happens
once. This isn't really a problem since we always just set it to the same
value. We now just stop writing it after the first time.
PR3584, https://bugs.debian.org/534534
Signed-off-by: Kurt Roeckx <kurt@roeckx.be>
Reviewed-by: Rich Salz <rsalz@openssl.org>
The trial division and probable prime with coprime tests are disabled
on WIN32 builds because they use internal functions not exported from
the WIN32 DLLs.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
Tighten client-side session ticket handling during renegotiation:
ensure that the client only accepts a session ticket if the server sends
the extension anew in the ServerHello. Previously, a TLS client would
reuse the old extension state and thus accept a session ticket if one was
announced in the initial ServerHello.
Reviewed-by: Bodo Moeller <bodo@openssl.org>
SSL_set_SSL_CTX is used to change the SSL_CTX for SNI, keep the
supported signature algorithms and raw cipherlist.
Reviewed-by: Tim Hudson <tjh@openssl.org>
Facilitate switch to custom scatter-gather routines. This modification
does not change algorithms, only makes it possible to implement
alternative. This is achieved by a) moving precompute table to assembly
(perlasm parses ecp_nistz256_table.c and is free to rearrange data to
match gathering algorithm); b) adhering to explicit scatter subroutine
(which for now is simply a memcpy). First implementations that will use
this option are 32-bit assembly implementations, ARMv4 and x86, where
equivalent of current read-whole-table-select-single-value algorithm
is too time-consuming. [On side note, switching to scatter-gather on
x86_64 would allow to improve server-side ECDSA performance by ~5%].
Reviewed-by: Bodo Moeller <bodo@openssl.org>
When we're configured with no-ssl3 and we receive an SSL v3 Client Hello, we set
the method to NULL. We didn't used to do that, and it breaks things. This is a
regression introduced in 62f45cc27d. Keep the old
method since the code is not able to deal with a NULL method at this time.
CVE-2014-3569, PR#3571
Reviewed-by: Emilia Käsper <emilia@openssl.org>
The different -I compiler parameters will take care of the rest...
Reviewed-by: Tim Hudson <tjh@openssl.org>
Conflicts:
crypto/evp/evp_enc.c
crypto/rsa/rsa_oaep.c
crypto/rsa/rsa_pk1.c
CVE-2014-3513
This issue was reported to OpenSSL on 26th September 2014, based on an original
issue and patch developed by the LibreSSL project. Further analysis of the issue
was performed by the OpenSSL team.
The fix was developed by the OpenSSL team.
Reviewed-by: Tim Hudson <tjh@openssl.org>
Patch supplied by Matthieu Patou <mat@matws.net>, and modified to also
remove duplicate definition of PKCS7_type_is_digest.
PR#3551
Reviewed-by: Rich Salz <rsalz@openssl.org>
If data is NULL, return the size needed to hold the
derived key. No other API to do this, so document
the behavior.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reencode DigestInto in DER and check against the original: this
will reject any improperly encoded DigestInfo structures.
Note: this is a precautionary measure, there is no known attack
which can exploit this.
Thanks to Brian Smith for reporting this issue.
Reviewed-by: Tim Hudson <tjh@openssl.org>
The following #ifdef tests were all removed:
__MWERKS__
MAC_OS_pre_X
MAC_OS_GUSI_SOURCE
MAC_OS_pre_X
OPENSSL_SYS_MACINTOSH_CLASSIC
OPENSSL_SYS_MACOSX_RHAPSODY
Reviewed-by: Andy Polyakov <appro@openssl.org>
Do the final padding check in EVP_DecryptFinal_ex in constant time to
avoid a timing leak from padding failure.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Also tweak s3_cbc.c to use new constant-time methods.
Also fix memory leaks from internal errors in RSA_padding_check_PKCS1_OAEP_mgf1
This patch is based on the original RT submission by Adam Langley <agl@chromium.org>,
as well as code from BoringSSL and OpenSSL.
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
that fixed PR#3450 where an existing cast masked an issue when i was changed
from int to long in that commit
Picked up on z/linux (s390) where sizeof(int)!=sizeof(long)
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
GetDIBits has been around since Windows2000 and
BitBitmapBits is an old Win16 compatibility function
that is much slower.
Reviewed-by: Tim Hudson <tjh@openssl.org>
If we don't find a signer in the internal list, then fall
through and look at the internal list; don't just return NULL.
Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
This is funny; Ben commented in the source, Matt opend a ticket,
and Rich is doing the submit. Need more code-review? :)
Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
When calling X509_set_version to set v1 certificate, that
should mean that the version number field is omitted.
Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
This is a more comprehensive fix. It changes all
keygen apps to use 2K keys. It also changes the
default to use SHA256 not SHA1. This is from
Kurt's upstream Debian changes.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
The EXT_BITSTRING and EXT_IA5STRING are defined in x509v3.h, but
the low-level functions are not public. They are useful, no need
to make them static. Note that BITSTRING already was exposed since
this RT was created, so now we just export IA5STRING functions.
Reviewed-by: Tim Hudson <tjh@openssl.org>
pod2man now complains when item tags are not sequential.
Also complains about missing =back and other tags.
Silence the warnings; most were already done.
Reviewed-by: Tim Hudson <tjh@openssl.org>
The original RT request included a patch. By the time
we got around to doing it, however, the callback scheme
had changed. So I wrote a new function RSA_check_key_ex()
that uses the BN_GENCB callback. But thanks very much
to Vinet Sharma <vineet.sharma@gmail.com> for the
initial implementation.
Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
In the current code, the check isn't redundant.
And in fact the REAL check was missing.
This avoids a NULL-deref crash.
Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
i2d_re_X509_tbs re-encodes the TBS portion of the certificate.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Dr Stephen Henson <steve@openssl.org>
Fix a bug in handling of 128 byte long PSK identity in
psk_client_callback.
OpenSSL supports PSK identities of up to (and including) 128 bytes in
length. PSK identity is obtained via the psk_client_callback,
implementors of which are expected to provide a NULL-terminated
identity. However, the callback is invoked with only 128 bytes of
storage thus making it impossible to return a 128 byte long identity and
the required additional NULL byte.
This CL fixes the issue by passing in a 129 byte long buffer into the
psk_client_callback. As a safety precaution, this CL also zeroes out the
buffer before passing it into the callback, uses strnlen for obtaining
the length of the identity returned by the callback, and aborts the
handshake if the identity (without the NULL terminator) is longer than
128 bytes.
(Original patch amended to achieve strnlen in a different way.)
Reviewed-by: Rich Salz <rsalz@openssl.org>
"inline" without static is not correct as the compiler may choose to ignore it
and will then either emit an external definition, or expect one.
Reviewed-by: Geoff Thorpe <geoff@openssl.org>
RT842, closed back in 2004, changed the default serial number
to be a random number rather than zero. Finally time to update
the doc
Reviewed-by: Tim Hudson <tjh@openssl.org>
Add .crt/.cer/.crl to the filenames parsed.
I also updated the podpage (since it didn't exist when
this ticket was first created, nor when it was re-created
seven years later).
Reviewed-by: Tim Hudson <tjh@openssl.org>
Re-order algorithm list.
Be consistent in command synopsis.
Add content about signing.
Add EXAMPLE section
Add some missing options: -r, -fips-fingerprint -non-fips-allow
Various other fixes.
Reviewed-by: Andy Polyakov <appro@openssl.org>
Andy found an additional typo "can be can be".
Now I have that silly "Que sera sera" song stuck in my head.
Reviewed-by: Andy Polyakov <appro@openssl.org>
Add Darwin to list of case-insensitive filenames when
installing manapges. When doing this, I noticed that
we weren't setting "filecase" for the HTML doc install.
Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
RT1665: aes documentation.
Paul Green wrote a nice aes.pod file.
But we now encourage the EVP interface.
So I took his RT item and used it as impetus to add
the AES modes to EVP_EncryptInit.pod
I also noticed that rc4.pod has spurious references to some other
cipher pages, so I removed them.
RT2300: Clean up MD history (merged into RT1665)
Put HISTORY section only in EVP_DigestInit.pod. Also add words
to discourage use of older cipher-specific API, and remove SEE ALSO
links that point to them.
Make sure digest pages have a NOTE that says use EVP_DigestInit.
Review feedback:
More cleanup in EVP_EncryptInit.pod
Fixed SEE ALSO links in ripemd160.pod, sha.pod, mdc2.pod, blowfish.pod,
rc4.d, and des.pod. Re-order sections in des.pod for consistency
Reviewed-by: Matt Caswell <matt@openssl.org>
Problem with #ifdef in the BIO_CTRL_DGRAM_MTU_DISCOVER case that
is different from the BIO_CTRL_DGRAM_QUERY_MTU one which seems
correct.
Reviewed-by: Matt Caswell <matt@openssl.org>
Fix comments in ssltest.c: return value of 0 now means extension is
omitted and add_cb is not called for servers if the corresponding
extension is absent in ClientHello.
Test add_cb is not called if extension is not received.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
Instead of SSL_CTX_set_custom_cli_ext and SSL_CTX_set_custom_srv_ext
use SSL_CTX_add_client_custom_ext and SSL_CTX_add_server_custom_ext.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
Support separate parse and add callback arguments.
Add new callback so an application can free extension data.
Change return value for send functions so < 0 is an error 0
omits extension and > 0 includes it. This is more consistent
with the behaviour of other functions in OpenSSL.
Modify parse_cb handling so <= 0 is an error.
Make SSL_CTX_set_custom_cli_ext and SSL_CTX_set_custom_cli_ext argument
order consistent.
NOTE: these changes WILL break existing code.
Remove (now inaccurate) in line documentation.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
Use "parse" and "add" for function and callback names instead of
"first" and "second".
Change arguments to callback so the extension type is unsigned int
and the buffer length is size_t. Note: this *will* break existing code.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
Since sanity checks are performed for all custom extensions the
serverinfo checks are no longer needed.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
Reject attempts to use extensions handled internally.
Add flags to each extension structure to indicate if an extension
has been sent or received. Enforce RFC5246 compliance by rejecting
duplicate extensions and unsolicited extensions and only send a
server extension if we have sent the corresponding client extension.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
Use the same structure for client and server custom extensions.
Add utility functions in new file t1_ext.c.
Use new utility functions to handle custom server and client extensions
and remove a lot of code duplication.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
Add the wrapper to all public header files (Configure
generates one). Don't bother for those that are just
lists of #define's that do renaming.
Reviewed-by: Tim Hudson <tjh@openssl.org>
The old code implicitly relies on the ASN.1 code returning a \0-prefixed buffer
when the buffer length is 0. Change this to verify explicitly that the ASN.1 string
has positive length.
Reviewed-by: Dr Stephen Henson <steve@openssl.org>
When d2i_ECPrivateKey reads a private key with a missing (optional) public key,
generate one automatically from the group and private key.
Reviewed-by: Dr Stephen Henson <steve@openssl.org>
This change saves several EC routines from crashing when an EC_KEY is
missing a public key. The public key is optional in the EC private key
format and, without this patch, running the following through `openssl
ec` causes a crash:
-----BEGIN EC PRIVATE KEY-----
MBkCAQEECAECAwQFBgcIoAoGCCqGSM49AwEH
-----END EC PRIVATE KEY-----
Reviewed-by: Dr Stephen Henson <steve@openssl.org>
The description of when the server creates a DH key is
confusing. This cleans it up.
(rsalz: also removed trailing whitespace.)
Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
The description of when the server creates a DH key is
confusing. This cleans it up.
(rsalz: also removed trailing whitespace.)
Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
The EXAMPLE that used FILE and RC2 doesn't compile due to a
few minor errors. Tweak to use IDEA and AES-128. Remove
examples about RC2 and RC5.
Reviewed-by: Emilia Kasper <emilia@openssl.org>
Use existing error code SSL_R_RECORD_TOO_SMALL for too many empty records.
For ease of backporting the patch to release branches.
Reviewed-by: Bodo Moeller <bodo@openssl.org>
Add a dozen more const declarations where appropriate.
These are from Justin; while adding his patch, I noticed
ASN1_BIT_STRING_check could be fixed, too.
Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
While RFC6367 focuses on Camellia-GCM cipher suites, it also adds a few
cipher suites that use SHA-2 based HMAC that can be very easily
added.
Tested against gnutls 3.3.5
PR#3443
Reviewed-by: Tim Hudson <tjh@openssl.org>
Add a declaration for get_issuer_sk() so that other
functions in x509_vf.c could use it. (Planned work
around cross-certification chains.)
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
"Another machine, another version of gcc, another batch
of compiler warnings." Add "=NULL" to some local variable
declarations that are set by passing thier address into a
utility function; confuses GCC it might not be set.
Reviewed-by: Emilia Ksper <emilia@silkandcyanide.net>
Move custom extension structures from SSL_CTX to CERT structure.
This change means the form can be revised in future without binary
compatibility issues. Also since CERT is part of SSL structures
so per-SSL custom extensions could be supported in future as well as
per SSL_CTX.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Emilia Käsper <emilia@openssl.org>
In Makefile, when build manpages, put the current directory
at the start of the podpath so that cross-refs find the
local directory first.
Reviewed-by: Tim Hudson <tjh@cryptosoft.com>
In two OpenSSL manual pages, in the NAME section, the last word of the
name list is followed by a stray trailing comma. While this may seem
minor, it is worth fixing because it may confuse some makewhatis(8)
implementations.
While here, also add the missing word "size" to the one line
description in SSL_CTX_set_max_cert_list(3).
Reviewed by: Dr Stephen Henson <shenson@drh-consultancy.co.uk>
Update the dgst.pod page to include SHA224...512 algorithms.
Update apps/progs.pl to add them to the digest command table.
Reviewed-by: Tim Hudson <tjh@cryptosoft.com>
The addition of SRP authentication needs to be checked in various places
to work properly. Specifically:
A certificate is not sent.
A certificate request must not be sent.
Server key exchange message must not contain a signature.
If appropriate SRP authentication ciphersuites should be chosen.
Reviewed-by: Matt Caswell <matt@openssl.org>
Invalid parameters passed to the SRP code can be overrun an internal
buffer. Add sanity check that g, A, B < N to SRP code.
Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
Group for reporting this issue.
If a client attempted to use an SRP ciphersuite and it had not been
set up correctly it would crash with a null pointer read. A malicious
server could exploit this in a DoS attack.
Thanks to Joonas Kuorilehto and Riku Hietamäki from Codenomicon
for reporting this issue.
CVE-2014-2970
Reviewed-by: Tim Hudson <tjh@openssl.org>
- Upon parsing, reject OIDs with invalid base-128 encoding.
- Always NUL-terminate the destination buffer in OBJ_obj2txt printing function.
CVE-2014-3508
Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
In a couple of functions, a sequence number would be calculated twice.
Additionally, in |dtls1_process_out_of_seq_message|, we know that
|frag_len| <= |msg_hdr->msg_len| so the later tests for |frag_len <
msg_hdr->msg_len| can be more clearly written as |frag_len !=
msg_hdr->msg_len|, since that's the only remaining case.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Emilia Käsper <emilia@openssl.org>
Previously, a truncated DTLS fragment in
|dtls1_process_out_of_seq_message| would cause *ok to be cleared, but
the return value would still be the number of bytes read. This would
cause |dtls1_get_message| not to consider it an error and it would
continue processing as normal until the calling function noticed that
*ok was zero.
I can't see an exploit here because |dtls1_get_message| uses
|s->init_num| as the length, which will always be zero from what I can
see.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Emilia Käsper <emilia@openssl.org>
The |pqueue_insert| function can fail if one attempts to insert a
duplicate sequence number. When handling a fragment of an out of
sequence message, |dtls1_process_out_of_seq_message| would not call
|dtls1_reassemble_fragment| if the fragment's length was zero. It would
then allocate a fresh fragment and attempt to insert it, but ignore the
return value, leaking the fragment.
This allows an attacker to exhaust the memory of a DTLS peer.
Fixes CVE-2014-3507
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Emilia Käsper <emilia@openssl.org>
In |dtls1_reassemble_fragment|, the value of
|msg_hdr->frag_off+frag_len| was being checked against the maximum
handshake message size, but then |msg_len| bytes were allocated for the
fragment buffer. This means that so long as the fragment was within the
allowed size, the pending handshake message could consume 16MB + 2MB
(for the reassembly bitmap). Approx 10 outstanding handshake messages
are allowed, meaning that an attacker could consume ~180MB per DTLS
connection.
In the non-fragmented path (in |dtls1_process_out_of_seq_message|), no
check was applied.
Fixes CVE-2014-3506
Wholly based on patch by Adam Langley with one minor amendment.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
The |item| variable, in both of these cases, may contain a pointer to a
|pitem| structure within |s->d1->buffered_messages|. It was being freed
in the error case while still being in |buffered_messages|. When the
error later caused the |SSL*| to be destroyed, the item would be double
freed.
Thanks to Wah-Teh Chang for spotting that the fix in 1632ef74 was
inconsistent with the other error paths (but correct).
Fixes CVE-2014-3505
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Emilia Käsper <emilia@openssl.org>
Don't call internal functions directly call them through
SSL_test_functions(). This also makes unit testing work on
Windows and platforms that don't export internal functions
from shared libraries.
By default unit testing is not enabled: it requires the compile
time option "enable-unit-test".
Reviewed-by: Geoff Thorpe <geoff@openssl.org>
The call to setenv in gost2814789t.c is not portable and may
not reflect the location of the GOST ENGINE on all platforms anyway.
Instead set OPENSSL_ENGINES in test/Makefile
Reviewed-by: Geoff Thorpe <geoff@openssl.org>
This has been unmaintained for a long time. If it's still of interest
to anyone, it can be obtained easily enough by reverting this commit.
(It could join other demo code in some other repository, perhaps.) In
any case we don't want it taking up space in the baseline source
package, so <snip>.
Signed-off-by: Geoff Thorpe <geoff@openssl.org>
ssl/ssl_locl.h now comes first to ensure that it will compile standalone.
test/testutil.h is considered to be in the same directory as the test file,
since the test file will be linked into test/ and built there.
Reviewed-by: Tim Hudson <tjh@openssl.org>
Add support for RFC5649 key wrapping with padding.
Add RFC5649 tests to evptests.txt
Based on PR#3434 contribution by Petr Spacek <pspacek@redhat.com>.
EVP support and minor changes added by Stephen Henson.
Doxygen comment block updates by Tim Hudson.
Reviewed-by: Tim Hudson <tjh@openssl.org>
Add description of the option to advertise support of
Next Protocol Negotiation extension (-nextprotoneg) to
man pages of s_client and s_server.
PR#3444
This is actually ok for this function, but initialised to zero anyway if
PURIFY defined.
This does have the impact of masking any *real* unitialised data reads in bn though.
Patch based on approach suggested by Rich Salz.
PR#3415
Reduces number of silly casts in OpenSSL code and likely most
applications. Consistent with (char *) for "peername" value from
X509_check_host() and X509_VERIFY_PARAM_get0_peername().
Internal pointers in CCM, GCM and XTS contexts should either be
NULL or set to point to the appropriate key schedule. This needs
to be adjusted when copying contexts.
IN parameter.
Under the old docs, the only thing stated was "at most
EVP_PKEY_size(pkey) bytes will be written". It was kind of misleading
since it appears EVP_PKEY_size(pkey) WILL be written regardless of the
signature's buffer size.
If CSR verify fails in ca utility print out error messages.
Otherwise some errors give misleading output: for example
if the key size exceeds the library limit.
PR#2875
In the ssl_cipher_get_evp() function, fix off-by-one errors in index validation before accessing arrays.
Bug discovered and fixed by Miod Vallat from the OpenBSD team.
PR#3375
cms, ocsp, s_client, s_server and smime tools also use args_verify()
for parsing options, that makes them most of the same options
verify tool does. Add those options to man pages and reference
their explanation in the verify man page.
The options related to policy used for verification, verification
of subject names in certificate and certificate chain handling
were missing in the verify(1) man page. This fixes this issue.
Allow CCS after finished has been sent by client: at this point
keys have been correctly set up so it is OK to accept CCS from
server. Without this renegotiation can sometimes fail.
PR#3400
A client reference identity of ".example.com" matches a server
certificate presented identity that is any sub-domain of "example.com"
(e.g. "www.sub.example.com).
With the X509_CHECK_FLAG_SINGLE_LABEL_SUBDOMAINS flag, it matches
only direct child sub-domains (e.g. "www.sub.example.com").
* Make a clear distinction between DH and ECDH key exchange.
* Group all key exchange cipher suite identifiers, first DH then ECDH
* add descriptions for all supported *DH* identifiers
* add ECDSA authentication descriptions
* add example showing how to disable all suites that offer no
authentication or encryption
Defines SETUP_TEST_FIXTURE and EXECUTE_TEST, and updates ssl/heartbeat_test.c
using these macros. SETUP_TEST_FIXTURE makes use of the new TEST_CASE_NAME
macro, defined to use __func__ or __FUNCTION__ on platforms that support those
symbols, or to use the file name and line number otherwise. This should fix
several reported build problems related to lack of C99 support.
Because of a missing include <fcntl.h> we don't have O_CREATE and don't create
the file with open() using mode 0600 but fall back to using fopen() with the
default umask followed by a chmod().
Problem found by Jakub Wilk <jwilk@debian.org>.
A buffer overrun attack can be triggered by sending invalid DTLS fragments
to an OpenSSL DTLS client or server. This is potentially exploitable to
run arbitrary code on a vulnerable client or server.
Fixed by adding consistency check for DTLS fragments.
Thanks to Jüri Aedla for reporting this issue.
(cherry picked from commit 1632ef7448)
Only accept change cipher spec when it is expected instead of at any
time. This prevents premature setting of session keys before the master
secret is determined which an attacker could use as a MITM attack.
Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for reporting this issue
and providing the initial fix this patch is based on.
(cherry picked from commit bc8923b1ec)
Unnecessary recursion when receiving a DTLS hello request can be used to
crash a DTLS client. Fixed by handling DTLS hello request without recursion.
Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
(cherry picked from commit d3152655d5)
Add certificates if -nocerts and -certfile specified when signing
in smime application. This can be used this to specify the
order certificates appear in the PKCS#7 structure: some broken
applications require a certain ordering.
PR#3316
Add TLS padding extension to SSL_OP_ALL so it is used with other
"bugs" options and can be turned off.
This replaces SSL_OP_SSLREF2_REUSE_CERT_TYPE_BUG which is an ancient
option referring to SSLv2 and SSLREF.
PR#3336
the verify app man page didn't describe the usage of attime option
even though it was listed as a valid option in the -help message.
This patch fixes this omission.
When looking for the issuer of a certificate, if current candidate is
expired, continue looking. Only return an expired certificate if no valid
certificates are found.
PR#3359
subjectAltName field. The Name Contraint example in x509v3_config(5) even use IP
as an example:
nameConstraints=permitted;IP:192.168.0.0/255.255.0.0
However, until now, the verify code for IP name contraints did not exist. Any
check with a IP Address Name Constraint results in a "unsupported name constraint
type" error.
This patch implements support for IP Address Name Constraint (v4 and v6). This code
validaded correcly certificates with multiple IPv4/IPv6 address checking against
a CA certificate with these constraints:
permitted;IP.1=10.9.0.0/255.255.0.0
permitted;IP.2=10.48.0.0/255.255.0.0
permitted;IP.3=10.148.0.0/255.255.0.0
permitted;IP.4=fdc8:123f:e31f::/ffff:ffff:ffff::
Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com>
The previous calls to memset() were added to tear_down() when I noticed the
test spuriously failing in opt mode, with different results each time. This
appeared to be because the allocator zeros out memory in debug mode, but not
in opt mode. Since the heartbeat functions silently drop the request on error
without modifying the contents of the write buffer, whatever random contents
were in memory before being reallocated to the write buffer used in the test
would cause nondeterministic test failures in the Heartbleed regression cases.
Adding these calls allowed the test to pass in both debug and opt modes.
Ben Laurie notified me offline that the test was aborting in
debug-ben-debug-64-clang mode, configured with GitConfigure and built with
GitMake. Looking into this, I realized the first memset() call was zeroing out
a reference count used by SSL_free() that was checked in
debug-ben-debug-64-clang mode but not in the normal debug mode.
Removing the memset() calls from tear_down() and adding a memset() for the
write buffer in set_up() addresses the issue and allows the test to
successfully execute in debug, opt, and debug-ben-debug-64-clang modes.
Replace manual ASN.1 decoder with ASN1_get object. This
will decode the tag and length properly and check against
it does not exceed the supplied buffer length.
PR#3335
"Teaser" means that it's not integrated yet and purpose of this
commit is primarily informational, to exhibit design choices,
such as how to handle alignment and endianness. In other words
it's proof-of-concept code that EVP module will build upon.
If the key type does not match any CMS recipient type return
an error instead of using a random key (MMA mitigation). This
does not leak any useful information to an attacker.
PR#3348
The "-unix <path>" argument allows s_server and s_client to use a unix
domain socket in the filesystem instead of IPv4 ("-connect", "-port",
"-accept", etc). If s_server exits gracefully, such as when "-naccept"
is used and the requested number of SSL/TLS connections have occurred,
then the domain socket file is removed. On ctrl-C, it is likely that
the stale socket file will be left over, such that s_server would
normally fail to restart with the same arguments. For this reason,
s_server also supports an "-unlink" option, which will clean up any
stale socket file before starting.
If you have any reason to want encrypted IPC within an O/S instance,
this concept might come in handy. Otherwise it just demonstrates that
there is nothing about SSL/TLS that limits it to TCP/IP in any way.
(There might also be benchmarking and profiling use in this path, as
unix domain sockets are much lower overhead than connecting over local
IP addresses).
Signed-off-by: Geoff Thorpe <geoff@openssl.org>
This patch resolves RT ticket #2608.
Thanks to Robert Dugal for originally spotting this, and to David
Ramos for noticing that the ball had been dropped.
Signed-off-by: Geoff Thorpe <geoff@openssl.org>
The lazy-initialisation of BN_MONT_CTX was serialising all threads, as
noted by Daniel Sands and co at Sandia. This was to handle the case that
2 or more threads race to lazy-init the same context, but stunted all
scalability in the case where 2 or more threads are doing unrelated
things! We favour the latter case by punishing the former. The init work
gets done by each thread that finds the context to be uninitialised, and
we then lock the "set" logic after that work is done - the winning
thread's work gets used, the losing threads throw away what they've done.
Signed-off-by: Geoff Thorpe <geoff@openssl.org>
Even though the meat of dso_vms.c is compiled out on non-VMS builds,
the (pre-)compiler still traverses some of the macro handling. This
trips up at least one non-VMS build configuration, so this commit
makes the skip-VMS case more robust.
Signed-off-by: Geoff Thorpe <geoff@openssl.org>
It's not clear whether this inconsistency could lead to an actual
computation error, but it involved a BIGNUM being passed around the
montgomery logic in an inconsistent state. This was found using flags
-DBN_DEBUG -DBN_DEBUG_RAND, and working backwards from this assertion
in 'ectest';
ectest: bn_mul.c:960: BN_mul: Assertion `(_bnum2->top == 0) ||
(_bnum2->d[_bnum2->top - 1] != 0)' failed
Signed-off-by: Geoff Thorpe <geoff@openssl.org>
There are certainly many more constifiable strings in the various
interfaces, which I hope to get to eventually.
Signed-off-by: Geoff Thorpe <geoff@openssl.org>
Gets rid of this;
defined(@array) is deprecated at ../util/mkerr.pl line 792.
(Maybe you should just omit the defined()?)
defined(@array) is deprecated at ../util/mkerr.pl line 800.
(Maybe you should just omit the defined()?)
Signed-off-by: Geoff Thorpe <geoff@openssl.org>
o_time.h was removed in commit ff49a94, which breaks "make update"
unless mkdir.pl is updated accordingly.
Signed-off-by: Geoff Thorpe <geoff@openssl.org>
Treat a zero length passed to ssleay_rand_add a no op: the existing logic
zeroes the md value which is very bad. OpenSSL itself never does this
internally and the actual call doesn't make sense as it would be passing
zero bytes of entropy.
Thanks to Marcus Meissner <meissner@suse.de> for reporting this bug.
(cherry picked from commit 5be1ae28ef)
A missing bounds check in the handling of the TLS heartbeat extension
can be used to reveal up to 64k of memory to a connected client or
server.
Thanks for Neel Mehta of Google Security for discovering this bug and to
Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
preparing the fix (CVE-2014-0160)
(cherry picked from commit 96db9023b8)
Use bufsiz - 1 not BUFSIZ - 1 when prompting for a password in
the openssl utility.
Thanks to Rob Mackinnon, Leviathan Security for reporting this issue.
Security callback: selects which parameters are permitted including
sensible defaults based on bits of security.
The "parameters" which can be selected include: ciphersuites,
curves, key sizes, certificate signature algorithms, supported
signature algorithms, DH parameters, SSL/TLS version, session tickets
and compression.
In some cases prohibiting the use of a parameters will mean they are
not advertised to the peer: for example cipher suites and ECC curves.
In other cases it will abort the handshake: e.g DH parameters or the
peer key size.
Documentation to follow...
New function ssl_cipher_disabled.
Check for disabled client ciphers using ssl_cipher_disabled.
New function to return only supported ciphers.
New option to ciphers utility to print only supported ciphers.
Add auto DH parameter support. This is roughly equivalent to the
ECDH auto curve selection but for DH. An application can just call
SSL_CTX_set_auto_dh(ctx, 1);
and appropriate DH parameters will be used based on the size of the
server key.
Unlike ECDH there is no way a peer can indicate the range of DH parameters
it supports. Some peers cannot handle DH keys larger that 1024 bits for
example. In this case if you call:
SSL_CTX_set_auto_dh(ctx, 2);
Only 1024 bit DH parameters will be used.
If the server key is 7680 bits or more in size then 8192 bit DH parameters
will be used: these will be *very* slow.
The old export ciphersuites aren't supported but those are very
insecure anyway.
Don't clear verification errors from the error queue unless
SSL_BUILD_CHAIN_FLAG_CLEAR_ERROR is set.
If errors occur during verification and SSL_BUILD_CHAIN_FLAG_IGNORE_ERROR
is set return 2 so applications can issue warnings.
(cherry picked from commit 2dd6976f6d)
Some CMS SignedData structure use a signature algorithm OID such
as SHA1WithRSA instead of the RSA algorithm OID. Workaround this
case by tolerating the signature if we recognise the OID.
Fix for the attack described in the paper "Recovering OpenSSL
ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
by Yuval Yarom and Naomi Benger. Details can be obtained from:
http://eprint.iacr.org/2014/140
Thanks to Yuval Yarom and Naomi Benger for discovering this
flaw and to Yuval Yarom for supplying a fix.
(cherry picked from commit 2198be3483)
Conflicts:
CHANGES
The problem is that OpenSSH calls EVP_Cipher, which is not as
protective as EVP_CipherUpdate. Formally speaking we ought to
do more checks in *_cipher methods, including rejecting
lengths not divisible by block size (unless ciphertext stealing
is in place). But for now I implement check for zero length in
low-level based on precedent.
PR: 3087, 2775
Add option to set an alternative to the default hmacWithSHA1 PRF
for PKCS#8 private key encryptions. This is used automatically
by PKCS8_encrypt if the nid specified is a PRF.
Add option to pkcs8 utility.
Update docs.
(cherry picked from commit b60272b01f)
Although the memory allocated by compression methods is fixed and
cannot grow over time it can cause warnings in some leak checking
tools. The function SSL_COMP_free_compression_methods() will free
and zero the list of supported compression methods. This should
*only* be called in a single threaded context when an application
is shutting down to avoid interfering with existing contexts
attempting to look up compression methods.
(cherry picked from commit 976c58302b)
Windows 8 SDKs complain that GetVersion() is deprecated.
We only use GetVersion like this:
(GetVersion() < 0x80000000)
which checks if the Windows version is NT based. Use a macro check_winnt()
which uses GetVersion() on older SDK versions and true otherwise.
New flags to build certificate chains. The can be used to rearrange
the chain so all an application needs to do is add all certificates
in arbitrary order and then build the chain to check and correct them.
Add verify error code when building chain.
Update docs.
The flag SSL_OP_MSIE_SSLV2_RSA_PADDING hasn't done anything since OpenSSL
0.9.7h but deleting it will break source compatibility with any software
that references it. Restore it but #define to zero.
(cherry picked from commit b17d6b8d1d)
If you use "-newkey rsa" it's supposed to read the default number of bits from the
config file. However the value isn't used to generate the key, but it does
print it's generating such a key. The set_keygen_ctx() doesn't call
EVP_PKEY_CTX_set_rsa_keygen_bits() and you end up with the default set in
pkey_rsa_init() (1024). Afterwards the number of bits gets read from the config
file, but nothing is done with that anymore.
We now read the config first and use the value from the config file when no size
is given.
PR: 2592
apps/pkcs12.c accepts -password as an argument. The document author
almost certainly meant to write "-password, -passin".
However, that is not correct, either. Actually the code treats
-password as equivalent to -passin, EXCEPT when -export is also
specified, in which case -password as equivalent to -passout.
When a chain is complete and ends in a trusted root checks are also
performed on the TA and the callback notified with ok==1. For
consistency do the same for chains where the TA is not self signed.
If multiple TLS extensions are expected but not received, the TLS extension and supplemental data 'generate' callbacks are the only chance for the receive-side to trigger a specific TLS alert during the handshake.
Removed logic which no-op'd TLS extension generate callbacks (as the generate callbacks need to always be called in order to trigger alerts), and updated the serverinfo-specific custom TLS extension callbacks to track which custom TLS extensions were received by the client, where no-ops for 'generate' callbacks are appropriate.
If an application calls the macro SSL_CTX_get_extra_chain_certs
return either the old "shared" extra certificates or those associated
with the current certificate.
This means applications which call SSL_CTX_use_certificate_chain_file
and retrieve the additional chain using SSL_CTX_get_extra_chain_certs
will still work. An application which only wants to check the shared
extra certificates can call the new macro
SSL_CTX_get_extra_chain_certs_only
This allows to process multiple fragmets of maximum fragment size,
as opposite to chopping maximum-sized fragments to multiple smaller
ones. This approach relies on dynamic allocation of larger buffers,
which we trade for performance improvement, for several *times* in
some situations.
New ctrl sets current certificate based on certain criteria. Currently
two options: set the first valid certificate as current and set the
next valid certificate as current. Using these an application can
iterate over all certificates in an SSL_CTX or SSL structure.
Remove reference to ERR_TXT_MALLOCED in the error library as that is
only used internally. Indicate that returned error data must not be
freed.
(cherry picked from commit f2d678e6e8)
Always add a dynamically loaded ENGINE to list. Otherwise it can cause
problems when multiply loaded, especially if it adds new public key methods.
For all current engines we only want a single implementation anyway.
Replace the full ciphersuites with "EDH-" in their labels with "DHE-"
so that all DHE ciphersuites are referred to in the same way.
Leave backward-compatible aliases for the ciphersuites in question so
that configurations which specify these explicitly will continue
working.
This change normalizes the SSL_CK_DHE_ #defines to use the common term
"DHE", while permitting older code that uses the more uncommon "EDH"
constants to compile properly.
DHE is the standard term used by the RFCs and by other TLS
implementations. It's useful to have the internal variables use the
standard terminology.
This patch leaves a synonym SSL_kEDH in place, though, so that older
code can still be built against it, since that has been the
traditional API. SSL_kEDH should probably be deprecated at some
point, though.
other parts of packet tracing emit the standard "DHE" label instead of
"edh". This change brings the output of ssl_print_client_keyex() and
ssl_print_server_keyex() into accordance with the standard term.
The standard terminology in https://tools.ietf.org/html/rfc5426 is
"DHE". "openssl ciphers" outputs "DHE" (for the most part). But
users of the library currently cannot specify "DHE", they must
currently specify "EDH".
This change allows users to specify the common term in cipher suite
strings without breaking backward compatibility.
ECDHE is the standard term used by the RFCs and by other TLS
implementations. It's useful to have the internal variables use the
standard terminology.
This patch leaves a synonym SSL_kEECDH in place, though, so that older
code can still be built against it, since that has been the
traditional API. SSL_kEECDH should probably be deprecated at some
point, though.
other parts of packet tracing emit the standard "ECDHE" label instead
of "EECDH". This change brings the output of ssl_print_client_keyex()
and ssl_print_server_keyex() into accordance with the standard term.
The standard terminology in https://tools.ietf.org/html/rfc4492 is
ECDHE. "openssl ciphers" outputs ECDHE. But users of the library
currently cannot specify ECDHE, they must specify EECDH.
This change allows users to specify the common term in cipher suite
strings without breaking backward compatibility.
When sending an invalid version number alert don't change the
version number to the client version if a session is already
established.
Thanks to Marek Majkowski for additional analysis of this issue.
PR#3191
If content is detached and not binary mode translate the input to
CRLF format. Before this change the input was verified verbatim
which lead to a discrepancy between sign and verify.
For DTLS we might need to retransmit messages from the previous session
so keep a copy of write context in DTLS retransmission buffers instead
of replacing it after sending CCS. CVE-2013-6450.
(cherry picked from commit 34628967f1)
SHA512_Transform was initially added rather as tribute to tradition
than for practucal reasons. But use was recently found in ssl/s3_cbc.c
and it turned to be problematic on platforms that don't tolerate
misasligned references to memory and lack assembly subroutine.
Move the IP, email and host checking fields from the public
X509_VERIFY_PARAM structure into an opaque X509_VERIFY_PARAM_ID
structure. By doing this the structure can be modified in future
without risk of breaking any applications.
(cherry picked from commit adc6bd73e3)
Conflicts:
crypto/x509/x509_vpm.c
For consistency with other cases if we are performing
partial chain verification with just one certificate
notify the callback with ok==1.
(cherry picked from commit 852553d9005e13aed7feb986a5d71cb885b994c7)
New functions to retrieve internal pointers to X509_VERIFY_PARAM
for SSL_CTX and SSL structures.
(cherry picked from commit be0c9270690ed9c1799900643cab91de146de857)
Some functions such as EVP_VerifyFinal only finalise a copy of the passed
context in case an application wants to digest more data. Doing this when
it is not needed is inefficient and many applications don't require it.
For compatibility the default is to still finalise a copy unless the
flag EVP_MD_CTX_FLAG_FINALISE is set in which case the passed
context is finalised an *no* further data can be digested after
finalisation.
If pointer comparison for current certificate fails check
to see if a match using X509_cmp succeeds for the current
certificate: this is useful for cases where the certificate
pointer is not available.
PR#3169
This patch, which currently applies successfully against master and
1_0_2, adds the following functions:
SSL_[CTX_]select_current_cert() - set the current certificate without
disturbing the existing structure.
SSL_[CTX_]get0_chain_certs() - get the current certificate's chain.
SSL_[CTX_]clear_chain_certs() - clear the current certificate's chain.
The patch also adds these functions to, and fixes some existing errors
in, SSL_CTX_add1_chain_cert.pod.
If the oid parameter is set to NULL in X509_add1_trust_object
create an empty list of trusted purposes corresponding to
"no purpose" if trust is checked.
Original definition depended on __LONG_MAX__ that is not guaranteed to
be present. As we don't support platforms with int narrower that 32 bits
it's appropriate to make defition inconditional.
PR: 3165
Based on a suggested workaround for the "TLS hang bug" (see FAQ and PR#2771):
if the TLS Client Hello record length value would otherwise be > 255 and less
that 512 pad with a dummy extension containing zeroes so it is at least 512.
To enable it use an unused extension number (for example 0x4242) using
e.g. -DTLSEXT_TYPE_wtf=0x4242
WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
PR: 2809
DTLS/SCTP requires DATA and FORWARD-TSN chunks to be protected with
SCTP-AUTH. It is checked if this has been activated successfully for
the local and remote peer. Due to a bug, however, the
gauth_number_of_chunks field of the authchunks struct is missing on
FreeBSD, and was therefore not considered in the OpenSSL implementation.
This patch sets the corresponding pointer for the check correctly
whether or not this bug is present.
(cherry picked from commit f596e3c491)
PR: 2808
With DTLS/SCTP the SCTP extension SCTP-AUTH is used to protect DATA and
FORWARD-TSN chunks. The key for this extension is derived from the
master secret and changed with the next ChangeCipherSpec, whenever a new
key has been negotiated. The following Finished then already uses the
new key. Unfortunately, the ChangeCipherSpec and Finished are part of
the same flight as the ClientKeyExchange, which is necessary for the
computation of the new secret. Hence, these messages are sent
immediately following each other, leaving the server very little time to
compute the new secret and pass it to SCTP before the finished arrives.
So the Finished is likely to be discarded by SCTP and a retransmission
becomes necessary. To prevent this issue, the Finished of the client is
still sent with the old key.
(cherry picked from commit 9fb523adce)
Don't require a public key in tls1_set_ec_id if compression status is
not needed. This fixes a bug where SSL_OP_SINGLE_ECDH_USE wouldn't work.
(cherry picked from commit 5ff68e8f6d)
This fixes problems in POD list formatting: extra or missing =back
sequences.
doc/ssl/SSL_CTX_set1_curves.pod around line 90: =back without =over
doc/ssl/SSL_CTX_set1_verify_cert_store.pod around line 73: =back without =over
doc/ssl/SSL_CTX_add1_chain_cert.pod around line 82: =back without =over
doc/crypto/evp.pod around line 40: '=item' outside of any '=over'
crypto/des/des.pod around line 184: You forgot a '=back' before '=head1'
PR#3147
Newer pod2man considers =item [1-9] part of a numbered list, while =item
0 starts an unnumbered list. Add a zero effect formatting mark to override
this.
doc/apps/smime.pod around line 315: Expected text after =item, not a
number
...
PR#3146
Instead, send random bytes, unless SSL_SEND_{CLIENT,SERVER}RANDOM_MODE
is set.
This is a forward-port of commits:
4af793036ff4c93b46ed3da721dac92583270191
While the gmt_unix_time record was added in an ostensible attempt to
mitigate the dangers of a bad RNG, its presence leaks the host's view
of the current time in the clear. This minor leak can help
fingerprint TLS instances across networks and protocols... and what's
worse, it's doubtful thet the gmt_unix_time record does any good at
all for its intended purpose, since:
* It's quite possible to open two TLS connections in one second.
* If the PRNG output is prone to repeat itself, ephemeral
handshakes (and who knows what else besides) are broken.
Removing RSA+MD5 from the default signature algorithm list
prevents its use by default.
If a broken implementation attempts to use RSA+MD5 anyway the sanity
checking of signature algorithms will cause a fatal alert.
Excessive fragmentation put additional burden (of addtional MAC
calculations) on the other size and limiting fragments it to 1KB
limits the overhead to ~6%.
Make DTLS behave like TLS when negotiating version: record layer has
DTLS 1.0, message version is 1.2.
Tolerate different version numbers if version hasn't been negotiated
yet.
- EC_GROUP_cmp shouldn't consider curves equal just because
the curve name is the same. (They really *should* be the same
in this case, but there's an EC_GROUP_set_curve_name API,
which could be misused.)
- EC_POINT_cmp shouldn't return 0 for ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED
or EC_R_INCOMPATIBLE_OBJECTS errors because in a cmp API, 0 indicates
equality (not an error).
Reported by: king cope
Experimental support for encrypt then mac from
draft-gutmann-tls-encrypt-then-mac-02.txt
To enable it set the appropriate extension number (0x10 for the test server)
using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x10
For non-compliant peers (i.e. just about everything) this should have no
effect.
Removed prior audit proof logic - audit proof support was implemented using the generic TLS extension API
Tests exercising the new supplemental data registration and callback api can be found in ssltest.c.
Implemented changes to s_server and s_client to exercise supplemental data callbacks via the -auth argument, as well as additional flags to exercise supplemental data being sent only during renegotiation.
* Many XMPP servers are configured with multiple domains (virtual hosts)
* In order to establish successfully the TLS connection you have to specify
which virtual host you are trying to connect.
* Test this, for example with ::
* Fail:
openssl s_client -connect talk.google.com:5222 -starttls xmpp
* Works:
openssl s_client -connect talk.google.com:5222 -starttls xmpp -xmpphost gmail.com
* When the host used in "-connect" is not what the remote XMPP server expects
the server will return an error like this:
<stream:error>
<host-unknown xmlns='urn:ietf:params:xml:ns:xmpp-streams'/>
</stream:error>
* But the actual code will stay on the loop forever because the stop condition
"/stream:features>" will never happen,
* Make this more robust: The stop condition should be that BIO_read failed
* Test if for example with ::
openssl s_client -connect random.jabb3r.net:5222 -starttls xmpp
* Some XMPP Servers (OpenFire) use double quotes.
* This makes s_client starttls work with this servers.
* Tested with OpenFire servers from http://xmpp.net/ ::
openssl s_client -connect coderollers.com:5222 -starttls xmpp
This fix ensures that
* A HelloRequest is retransmitted if not responded by a ClientHello
* The HelloRequest "consumes" the sequence number 0. The subsequent
ServerHello uses the sequence number 1.
* The client also expects the sequence number of the ServerHello to
be 1 if a HelloRequest was received earlier.
This patch fixes the RFC violation.
Add X9.42 DH KDF. Move sharedinfo generation code to CMS library as the
same structure is used by DH and ECDH.
Move ASN1_OBJECT typedef to ossl_typ.h so it can be picked up by dh headers
without the need to use ASN1.
This change adds support for ALPN[1] in OpenSSL. ALPN is the IETF
blessed version of NPN and we'll be supporting both ALPN and NPN for
some time yet.
[1] https://tools.ietf.org/html/draft-ietf-tls-applayerprotoneg-00
Conflicts:
ssl/ssl3.h
ssl/t1_lib.c
Update ecdsatest to use ECDSA_sign_setup and ECDSA_sign_ex, this
avoids the nonce generation which would otherwise break the test.
Reinstate ecdsatest.
For RSA and DSA keys return an appropriate RecipientInfo type. By setting
CMS_RECIPINFO_NONE for DSA keys an appropriate error is returned if
an attempt is made to use DSA with enveloped data.
Add support for ECDH in enveloped data. The CMS ctrls for the EC ASN1
method decode/encode the appropriate parameters from the CMS ASN1 data
and send appropriate data to the EC public key method.
Add support for key wrap algorithms via EVP interface.
Generalise AES wrap algorithm and add to modes, making existing
AES wrap algorithm a special case.
Move test code to evptests.txt
This change updates 8a99cb29 to make the generation of (EC)DSA nonces
using the message digest the default. It also reverts the changes to
(EC)DSA_METHOD structure.
In addition to making it the default, removing the flag from EC_KEY
means that FIPS modules will no longer have an ABI mismatch.
PR #3090
Reported by: Franck Youssef <fry@open.ch>
If no new reason codes are obtained after checking a CRL exit with an
error to avoid repeatedly checking the same CRL.
This will only happen if verify errors such as invalid CRL scope are
overridden in a callback.
Extend RSA ASN1 method to support CMS PSS signatures for both sign
and verify.
For signing the EVP_PKEY_CTX parameters are read and the appropriate
CMS structures set up.
For verification the CMS structures are analysed and the corresponding
parameters in the EVP_PKEY_CTX set.
Also add RSA-OAEP support.
For encrypt the EVP_PKEY_CTX parameters are used.
For decrypt the CMS structure is uses to set the appropriate EVP_PKEY_CTX
parameters.
Add support for customisation of CMS handling of signed and enveloped
data from custom public key parameters.
This will provide support for RSA-PSS and RSA-OAEP but could also be
applied to other algorithms.
Improve RSA sing performance by 20-30% by:
- switching from floating-point to integer conditional moves;
- daisy-chaining sqr-sqr-sqr-sqr-sqr-mul sequences;
- using MONTMUL even during powers table setup;
This change adds the option to calculate (EC)DSA nonces by hashing the
message and private key along with entropy to avoid leaking the private
key if the PRNG fails.
PR#3071
The primary changes made are:
- Updates to the "NAME" section of many pages to correctly reflect the
functions defined on those pages. This section is automatically parsed
by the util/extract-names.pl script, so if it is not correct then
running "man" will not correctly locate the right manual pages.
- Updates to take account of where functions are now deprecated
- Full documentation of the ec sub-library
- A number of other typo corrections and other minor tweaks
Extend OAEP support. Generalise the OAEP padding functions to support
arbitrary digests. Extend EVP_PKEY RSA method to handle the new OAEP
padding functions and add ctrls to set the additional parameters.
requested size. Fixes OpenSSL #2701.
This change does not address the cases of generating safe primes, or
where the |add| parameter is non-NULL.
Conflicts:
crypto/bn/bn.h
crypto/bn/bn_err.c
eliminating them as dead code.
Both volatile and "memory" are used because of some concern that the compiler
may still cache values across the asm block without it, and because this was
such a painful debugging session that I wanted to ensure that it's never
repeated.
Reencode certificates in X509_sign_ctx as well as X509_sign.
This was causing a problem in the x509 application when it modified an
existing certificate.
While ARMv7 in general is capable of unaligned access, not all instructions
actually are. And trouble is that compiler doesn't seem to differentiate
those capable and incapable of unaligned access. Side effect is that kernel
goes into endless loop retrying same instruction triggering unaligned trap.
Problem was observed in xts128.c and ccm128.c modules. It's possible to
resolve it by using (volatile u32*) casts, but letting STRICT_ALIGNMENT
be feels more appropriate.
Check for Suite B support using method flags instead of version numbers:
anything supporting TLS 1.2 cipher suites will also support Suite B.
Return an error if an attempt to use DTLS 1.0 is made in Suite B mode.
If we successfully match a cookie don't set return value to 2 as this
results in other error conditions returning 2 as well.
Instead set return value to -2 which can be checked later if everything
else is OK.
Add new methods DTLS_*_method() which support both DTLS 1.0 and DTLS 1.2 and
pick the highest version the peer supports during negotiation.
As with SSL/TLS options can change this behaviour specifically
SSL_OP_NO_DTLSv1 and SSL_OP_NO_DTLSv1_2.
Since s->method does not reflect the final client version when a client
hello is sent for SSLv23_client_method it can't be relied on to indicate
if TLS 1.2 ciphers should be used. So use the client version instead.
Add DTLS1.2 support for cached records when computing handshake macs
instead of the MD5+SHA1 case for DTLS < 1.2 (this is a port of the
equivalent TLS 1.2 code to DTLS).
Add correct flags for DTLS 1.2, update s_server and s_client to handle
DTLS 1.2 methods.
Currently no support for version negotiation: i.e. if client/server selects
DTLS 1.2 it is that or nothing.
Since this is always called from DTLS code it is safe to assume the header
length should be the DTLS value. This avoids the need to check the version
number and should work with any version of DTLS (not just 1.0).
Use the enc_flags field to determine whether we should use explicit IV,
signature algorithms or SHA256 default PRF instead of hard coding which
versions support each requirement.
Revise DTLS code. There was a *lot* of code duplication in the
DTLS code that generates records. This makes it harder to maintain and
sometimes a TLS update is omitted by accident from the DTLS code.
Specifically almost all of the record generation functions have code like
this:
some_pointer = buffer + HANDSHAKE_HEADER_LENGTH;
... Record creation stuff ...
set_handshake_header(ssl, SSL_MT_SOMETHING, message_len);
...
write_handshake_message(ssl);
Where the "Record creation stuff" is identical between SSL/TLS and DTLS or
in some cases has very minor differences.
By adding a few fields to SSL3_ENC to include the header length, some flags
and function pointers for handshake header setting and handshake writing the
code can cope with both cases.
Note: although this passes "make test" and some simple DTLS tests there may
be some minor differences in the DTLS code that have to be accounted for.
If an ASN1_INTEGER structure is allocated but not explicitly set encode
it as zero: don't generate an invalid zero length INTEGER.
(cherry picked from commit 1643edc63c)
Add DTLS record header parsing, different client hello format and add
HelloVerifyRequest message type.
Add code to d1_pkt.c to send message headers to the message callback.
Add code to support GCM an CCM modes in evp_test. On encrypt this
will compare the expected ciphertext and tag. On decrypt it will
compare the expected plaintext: tag comparison is done internally.
Add a simple CCM test case and convert all tests from crypto/modes/gcm128.c
Add CMS_RecipientInfo_encrypt: this function encrypts an existing content
encryption key to match the key in the RecipientInfo structure: this is
useful if a new recpient is added to and existing enveloped data structure.
Add documentation.
podlators 2.5.0 has switched to dying on POD syntax errors. This means
that a bunch of long-standing erroneous POD in the openssl documentation
now leads to fatal errors from pod2man, halting installation.
Unfortunately POD constraints mean that you have to sort numeric lists
in ascending order if they start with 1: you cannot do 1, 0, 2 even if
you want 1 to appear first. I've reshuffled such (alas, I wish there
were a better way but I don't know of one).
The version check for DTLS1_VERSION was redundant as
DTLS1_VERSION > TLS1_1_VERSION, however we do need to
check for DTLS1_BAD_VER for compatibility.
PR:2984
MD5 should use little endian order. Fortunately the only ciphersuite
affected is EXP-RC2-CBC-MD5 (TLS_RSA_EXPORT_WITH_RC2_CBC_40_MD5) which
is a rarely used export grade ciphersuite.
(cherry picked from commit f306b87d76)
Kludge alert. This is arranged by passing padding length in unused
bits of SSL3_RECORD->type, so that orig_len can be reconstructed.
(cherry picked from commit 8bfd4c659f)
Break dependency on uint64_t. It's possible to declare bits as
unsigned int, because TLS packets are limited in size and 32-bit
value can't overflow.
(cherry picked from commit cab13fc847)
We have to use EVP in FIPS mode so we can only partially mitigate
timing differences.
Make an extra call to EVP_DigestSignUpdate to hash additonal blocks
to cover any timing differences caused by removal of padding.
(cherry picked from commit b908e88ec1)
The previous CBC patch was bugged in that there was a path through enc()
in s3_pkt.c/d1_pkt.c which didn't set orig_len. orig_len would be left
at the previous value which could suggest that the packet was a
sufficient length when it wasn't.
(cherry picked from commit 6cb19b7681)
This patch makes the decoding of SSLv3 and TLS CBC records constant
time. Without this, a timing side-channel can be used to build a padding
oracle and mount Vaudenay's attack.
This patch also disables the stitched AESNI+SHA mode pending a similar
fix to that code.
In order to be easy to backport, this change is implemented in ssl/,
rather than as a generic AEAD mode. In the future this should be changed
around so that HMAC isn't in ssl/, but crypto/ as FIPS expects.
(cherry picked from commit e130841bcc)
This change adds CRYPTO_memcmp, which compares two vectors of bytes in
an amount of time that's independent of their contents. It also changes
several MAC compares in the code to use this over the standard memcmp,
which may leak information about the size of a matching prefix.
(cherry picked from commit 2ee798880a)
Contributed by: Florian Weimer <fweimer@redhat.com>
Fixes to X509 hostname and email address checking. Wildcard matching support.
New test program and manual page.
to the SSL_CONF APIs.
This is complicated a little because the SSL_CTX structure is not available
when the command line is processed: so just check syntax of commands initially
and store them, ready to apply later.
client hello message. Previously this could only be retrieved on an initial
connection and it was impossible to determine the cipher IDs of any uknown
ciphersuites.
Multiple copies of the ENGINE will cause problems when it is cleaned up as
the methods are stored in static structures which will be overwritten and
freed up more than once.
Set static methods to NULL when the ENGINE is freed so it can be reloaded.
some invalid operations for testing purposes. Currently this can be used
to sign using digests the peer doesn't support, EC curves the peer
doesn't support and use certificates which don't match the type associated
with a ciphersuite.
Reported by: Tomas Mraz <tmraz@redhat.com>
Treat a NULL value passed to drbg_free_entropy callback as non-op. This
can happen if the call to fips_get_entropy fails.
New function X509_chain_up_ref to dup and up the reference count of
a STACK_OF(X509): replace equivalent functionality in several places
by the equivalent call.
by a certificate chain. Add additional tests to handle client
certificates: checks for matching certificate type and issuer name
comparison.
Print out results of checks for each candidate chain tested in
s_server/s_client.
possible to have different stores per SSL structure or one store in
the parent SSL_CTX. Include distint stores for certificate chain
verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
to build and store a certificate chain in CERT structure: returing
an error if the chain cannot be built: this will allow applications
to test if a chain is correctly configured.
Note: if the CERT based stores are not set then the parent SSL_CTX
store is used to retain compatibility with existing behaviour.
details in s_client.
Also add ctrl to set client certificate types. If not used sensible values
will be included based on supported signature algorithms: for example if
we don't include any DSA signing algorithms the DSA certificate type is
omitted.
Fix restriction in old code where certificate types would be truncated
if it exceeded TLS_CT_NUMBER.
the permitted signature algorithms for server and client authentication
are the same but it is now possible to set different algorithms for client
authentication only.
is required by client or server. An application can decide which
certificate chain to present based on arbitrary criteria: for example
supported signature algorithms. Add very simple example to s_server.
This fixes many of the problems and restrictions of the existing client
certificate callback: for example you can now clear existing certificates
and specify the whole chain.
the certificate can be used for (if anything). Set valid_flags field
in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
to have similar checks in it.
Add new "cert_flags" field to CERT structure and include a "strict mode".
This enforces some TLS certificate requirements (such as only permitting
certificate signature algorithms contained in the supported algorithms
extension) which some implementations ignore: this option should be used
with caution as it could cause interoperability issues.
Only store encoded versions of peer and configured signature algorithms.
Determine shared signature algorithms and cache the result along with NID
equivalents of each algorithm.
TLS v1.2. These are sent as an extension for clients and during a certificate
request for servers.
TODO: add support for shared signature algorithms, respect shared algorithms
when deciding which ciphersuites and certificates to permit.
Reported by: Phil Pennock <openssl-dev@spodhuis.org>
Make renegotiation work for TLS 1.2, 1.1 by not using a lower record
version client hello workaround if renegotiating.
BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
in CRYPTO_realloc_clean.
Thanks to Tavis Ormandy, Google Security Team, for discovering this
issue and to Adam Langley <agl@chromium.org> for fixing it. (CVE-2012-2110)
If OPENSSL_MAX_TLS1_2_CIPHER_LENGTH is set then limit the size of client
ciphersuites to this value. A value of 50 should be sufficient.
Document workarounds in CHANGES.
Some servers hang when presented with a client hello record length exceeding
255 bytes but will work with longer client hellos if the TLS record version
in client hello does not exceed TLS v1.0. Unfortunately this doesn't fix all
cases...
enabled instead of requiring an application to hard code a (possibly
inappropriate) parameter set and delve into EC internals we just
automatically use the preferred curve.
Tidy some code up.
Don't allocate a structure to handle ECC extensions when it is used for
default values.
Make supported curves configurable.
Add ctrls to retrieve shared curves: not fully integrated with rest of
ECC code yet.
continue with symmetric decryption process to avoid leaking timing
information to an attacker.
Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
this issue. (CVE-2012-0884)
Reported by: Remi Gacogne <rgacogne-bugs@coredump.fr>
Preserve unused bits value in non-canonicalised ASN1_STRING structures
by using ASN1_STRING_copy which preseves flags.
following reasons:
- it's not the way to engage XPG4v2 mode, defining _XOPEN_SOURCE to
value less than 500 is (see standards(5));
- we need to work out strategy to handle _XOPEN_SOURCE, current state
when we define e.g. _XOPEN_SOURCE to 500 in some files is inappropriate;
- sctp implementation on Solaris is incomplete, in sense that bss_dgram.c
doesn't compile, because not all structures are defined, so that
enabling sctp doesn't work anyway;
Submitted by: Bruce Stephens <bruce.stephens@isode.com>
Use same construct for EXHEADER in srp/Makefile as other makefiles to cope
with possibly empty EXHEADER.
signatures and MDC2 using EVP or RSA_sign. This has become more apparent
when the dgst utility in OpenSSL 1.0.0 and later switched to using the
EVP_DigestSign functions which call RSA_sign.
This means that the signature format OpenSSL 1.0.0 and later used with
dgst -sign and MDC2 is incompatible with previous versions.
Add detection in RSA_verify so either format works.
Note: MDC2 is disabled by default in OpenSSL and very rarely used in practice.
structure.
Before this the only way to add a custom chain was in the parent SSL_CTX
(which is shared by all key types and SSL structures) or rely on auto
chain building (which is performed on each handshake) from the trust store.
certificate chain instead of an X509 structure.
This makes it easier to enhance code in future and the chain
output functions have access to the CERT_PKEY structure being
used.
New function ssl_add_cert_chain which adds a certificate chain to
SSL internal BUF_MEM. Use this function in ssl3_output_cert_chain
and dtls1_output_cert_chain instead of partly duplicating code.
Thanks to Antonio Martin, Enterprise Secure Access Research and
Development, Cisco Systems, Inc. for discovering this bug and
preparing a fix. (CVE-2012-0050)
The cipher definitions of these ciphersuites have been around since SSLeay
but were always disabled. Now OpenSSL supports DH certificates they can be
finally enabled.
Various additional changes were needed to make them work properly: many
unused fixed DH sections of code were untested.
Submitted by: steve
Update maximum message size for certifiate verify messages to support
4096 bit RSA keys again as TLS v1.2 messages is two bytes longer.
Submitted by: Peter Sylvester <peter.sylvester@edelweb.fr>
Reviewed by: steve
- remove some unncessary SSL_err and permit
an srp user callback to allow a worker to obtain
a user verifier.
- cleanup and comments in s_server and demonstration
for asynchronous srp user lookup
New function to retrieve compression method from SSL_SESSION structure.
Delete SSL_SESSION_get_id_len and SSL_SESSION_get0_id functions
as they duplicate functionality of SSL_SESSION_get_id. Note: these functions
have never appeared in any release version of OpenSSL.
Submitted by: Peter Sylvester <peter.sylvester@edelweb.fr>
Reviewed by: steve
Remove unnecessary code for srp and to add some comments to
s_client.
- the callback to provide a user during client connect is
no longer necessary since rfc 5054 a connection attempt
with an srp cipher and no user is terminated when the
cipher is acceptable
- comments to indicate in s_client the (non-)usefulness of
th primalaty tests for non known group parameters.
Submitted by: Peter Sylvester <peter.sylvester@edelweb.fr>
Reviewed by: steve
Make SRP conformant to rfc 5054.
Changes are:
- removal of the addition state after client hello
- removal of all pre-rfc srp alert ids
- sending a fatal alert when there is no srp extension but when the
server wants SRP
- removal of unnecessary code in the client.
Submitted by: Robin Seggelmann <seggelmann@fh-muenster.de>
Reviewed by: steve
Send alert instead of assertion failure for incorrectly formatted DTLS
fragments.
File diff suppressed because it is too large
Load Diff
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
