This adds support for a new EAP method: EAP-TEAP (Tunnel Extensible
Authentication Protocol). This should be considered experimental since
RFC 7170 has number of conflicting statements and missing details to
allow unambiguous interpretation. As such, there may be interoperability
issues with other implementations and this version should not be
deployed for production purposes until those unclear areas are resolved.
This does not yet support use of NewSessionTicket message to deliver a
new PAC (either in the server or peer implementation). In other words,
only the in-tunnel distribution of PAC-Opaque is supported for now. Use
of the NewSessionTicket mechanism would require TLS library support to
allow arbitrary data to be specified as the contents of the message.
Signed-off-by: Jouni Malinen <j@w1.fi>
Add a new TLS_CONN_* flag to provide a higher level mechanism for adding
(instead of fully replacing) allowed list of TLS ciphersuites for TEAP
provisioning purposes.
Signed-off-by: Jouni Malinen <j@w1.fi>
This implements "tls-unique" derivation per RFC 5929, Section 3. This
will be needed for channel binding, e.g., with EAP-TEAP.
Signed-off-by: Jouni Malinen <j@w1.fi>
This adds a CTRL-EVENT-EAP-PEER-CERT even for depth=0 even if a depth >
0 certificate results in peer certificate validation error. Previously,
this case resulted in the upper layers not getting any information about
the used peer certificate. Now that information is available, e.g., to
allow server certificate -based overriding of the trust to be done.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
This is needed to allow upper layer software to learn the hash of the
server certificate for allowing user to override trust root
configuration.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Add tod=1 to CTRL-EVENT-EAP-PEER-CERT events if the peer certificate
includes the TOD policy in the X.509v3 Certificate Policies extension.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
This makes it easier to add new information to the callbacks without
having to modify each callback function type in EAPOL and EAP code every
time.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
GnuTLS backend already accepts CA cert blobs in both DER and PEM
formats. Implement similar trial-and-error handling in OpenSSL backend.
Signed-off-by: Santtu Lakkala <santtu.lakkala@jolla.com>
gcc 8.3.0 was apparently clever enough to optimize away the previously
used os_memset() to explicitly clear a stack buffer that contains keys
when that clearing happened just before returning from the function.
Since memset_s() is not exactly portable (or commonly available yet..),
use a less robust mechanism that is still pretty likely to prevent
current compilers from optimizing the explicit clearing of the memory
away.
Signed-off-by: Jouni Malinen <j@w1.fi>
The rcons[] and Td4s[] array values need to be type cast explicitly to
u32 for the left shift 24 operation to be defined due to the implicit
conversion to int not handling the case where MSB would become 1 without
depending on UB.
Credit to OSS-Fuzz: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=14929
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
If OCSP_resp_find_status() fails with the first OCSP_CERTID, the
generation of the second OCSP_CERTID ended up leaking memory. Fix this
by freeing the previously allocated OCSP_CERTID on that code path.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
BN_clear() does not free the BIGNUM; it only clears its value. Fix this
memory leak by using the appropriate BN_clear_free() function instead.
Fixes: b11fa98bcb ("Add explicit checks for peer's DH public key")
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
ec_params needs to be freed before returning from the function.
Extension of this function to support BoringSSL introduced this memory
leak and that was later extended to be the only variant and apply to
OpenSSL and LibreSSL cases as well in commit c23e87d0d1 ("OpenSSL:
Replace EVP_PKEY_paramgen() with EC_KEY_new_by_curve_name()").
Fixes: f29761297b ("BoringSSL: Implement crypto_ecdh_init()")
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Commit 373c796948 ("OpenSSL: Fix compile with OpenSSL 1.1.0 and
deprecated APIs") removed a call to ENGINE_load_dynamic() for newer
versions of OpenSSL, asserting that it should happen automatically.
That appears not to be the case, and loading engines now fails because
the dynamic engine isn't present.
Fix it by calling ENGINE_load_builtin_engines(), which works for all
versions of OpenSSL. Also remove the call to ERR_load_ENGINE_strings()
because that should have happened when SSL_load_error_strings() is
called anyway.
Fixes: 373c796948 ("OpenSSL: Fix compile with OpenSSL 1.1.0 and deprecated APIs")
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
This converts crypto_bignum_to_bin() to use the OpenSSL/BoringSSL
functions BN_bn2binpad()/BN_bn2bin_padded(), when available, to avoid
differences in runtime and memory access patterns depending on the
leading bytes of the BIGNUM value.
OpenSSL 1.0.2 and LibreSSL do not include such functions, so those cases
are still using the previous implementation where the BN_num_bytes()
call may result in different memory access pattern.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
It looks like SSL_CTX_set1_curves_list() command alone is not sufficient
to enable ECDH curve selection with older OpenSSL versions for TLS
server, so enable automatic selection first and specify the exact list
of curves after that.
This fixes failures in openssl_ecdh_curves test case when hostapd uses
OpenSSL 1.0.2.
Signed-off-by: Jouni Malinen <j@w1.fi>
Get rid of the branches that depend on the result of the Legendre
operation. This is needed to avoid leaking information about different
temporary results in blinding mechanisms.
This is related to CVE-2019-9494 and CVE-2019-9495.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
This helps in reducing measurable timing differences in operations
involving private information. BoringSSL has removed BN_FLG_CONSTTIME
and expects specific constant time functions to be called instead, so a
bit different approach is needed depending on which library is used.
The main operation that needs protection against side channel attacks is
BN_mod_exp() that depends on private keys (the public key validation
step in crypto_dh_derive_secret() is an exception that can use the
faster version since it does not depend on private keys).
crypto_bignum_div() is currently used only in SAE FFC case with not
safe-prime groups and only with values that do not depend on private
keys, so it is not critical to protect it.
crypto_bignum_inverse() is currently used only in SAE FFC PWE
derivation. The additional protection here is targeting only OpenSSL.
BoringSSL may need conversion to using BN_mod_inverse_blinded().
This is related to CVE-2019-9494 and CVE-2019-9495.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
These wpa_supplicant network profile parameters could be used to specify
a single match string that would be used against the dNSName items in
subjectAltName or CN. There may be use cases where more than one
alternative match string would be useful, so extend these to allow a
semicolon delimited list of values to be used (e.g.,
"example.org;example.com"). If any of the specified values matches any
of the dNSName/CN values in the server certificate, consider the
certificate as meeting this requirement.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Incorrect gen->type value was used to check whether subjectAltName
contained dNSName entries. This resulted in all domain_match and
domain_suffix_match entries failing to find a match and rejecting the
server certificate. Fix this by checking against the correct type
definition for dNSName.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
While commit 1c156e783d ("Fixed tls_prf() to handle keys with
odd length") added support for keys with odd length, the function
never reached this code as the function would return earlier in
case the key length was odd. Fix this by removing the first check
for the key length.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
SSL_use_certificate_chain_file() is not available in the current
BoringSSL even though the defined OPENSSL_VERSION_NUMBER is large enough
to claim that this function would be present in the OpenSSL API.
Fall back to using SSL_use_certificate_file() with BoringSSL to fix the
build.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Allow an additional context value to be passed to TLS exporter as
specified in RFC 5705 section 4.
This does not yet implement it for the internal TLS implementation.
However, as currently nothing uses context yet, this will not break
anything right now. WolfSSL maintainers also stated that they are not
going to add context support yet, but would look into it if/when this is
required by a published draft or a standard.
Signed-off-by: Ervin Oro <ervin.oro@aalto.fi>
X509_get_subject_name() in OpenSSL 1.0.2 does not mark its argument as a
const pointer, so need to type cast this to avoid a build warning.
Fixes: 841205a1ce ("OpenSSL: Add 'check_cert_subject' support for TLS server")
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
This patch added 'check_cert_subject' support to match the value of
every field against the DN of the subject in the client certificate. If
the values do not match, the certificate verification will fail and will
reject the user.
This option allows hostapd to match every individual field in the right
order, also allow '*' character as a wildcard (e.g OU=Development*).
Note: hostapd will match string up to 'wildcard' against the DN of the
subject in the client certificate for every individual field.
Signed-off-by: Paresh Chaudhary <paresh.chaudhary@rockwellcollins.com>
Signed-off-by: Jared Bents <jared.bents@rockwellcollins.com>
Signed-off-by: Jouni Malinen <j@w1.fi>
Pass the group order (if known/specified) to crypto_dh_derive_secret()
(and also to OpenSSL DH_generate_key() in case of Group 5) and verify
that the public key received from the peer meets 1 < pubkey < p and
pubkey^q == 1 mod p conditions.
While all these use cases were using only ephemeral DH keys, it is
better to use more explicit checks while deriving the shared secret to
avoid unexpected behavior.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
The flags variable needs to be initialized to 0 if check_crl is 0 in
the updated configuration.
Fixes: 159a7fbdea ("crl_reload_interval: Add CRL reloading support")
Signed-off-by: Jouni Malinen <j@w1.fi>
Add test-tls program that can be used for fuzzing the internal TLS
client and server implementations. This tool can write client or server
messages into a file as an initialization step and for the fuzzing step,
that file (with potential modifications) can be used to replace the
internally generated message contents.
The TEST_FUZZ=y build parameter is used to make a special build where a
hardcoded random number generator and hardcoded timestamp are used to
force deterministic behavior for the TLS operations.
Signed-off-by: Jouni Malinen <j@w1.fi>
This patch adds a new flag 'crl_reload_interval' to reload CRL
periodically. This can be used to reload ca_cert file and the included
CRL information on every new TLS session if difference between the last
reload and the current time in seconds is greater than
crl_reload_interval.
This reloading is used for cases where check_crl is 1 or 2 and the CRL
is included in the ca_file.
Signed-off-by: Paresh Chaudhary <paresh.chaudhary@rockwellcollins.com>
Signed-off-by: Jared Bents <jared.bents@rockwellcollins.com>
When using LibreSSL build fails with:
../src/crypto/tls_openssl.o: in function `tls_connection_client_cert':
../src/crypto/tls_openssl.c:2817: undefined reference to `SSL_use_certificate_chain_file'
collect2: error: ld returned 1 exit status
make: *** [Makefile:1901: wpa_supplicant] Error 1
There is no such function in LibreSSL.
Signed-off-by: Stefan Strogin <stefan.strogin@gmail.com>
Some distributions (e.g., Debian) have started introducting systemwide
OpenSSL policies to disable older protocol versions and ciphers
throughout all programs using OpenSSL. This can result in significant
number of interoperability issues with deployed EAP implementations.
Allow explicit wpa_supplicant (EAP peer) and hostapd (EAP server)
parameters to be used to request systemwide policies to be overridden if
older versions are needed to be able to interoperate with devices that
cannot be updated to support the newer protocol versions or keys. The
default behavior is not changed here, i.e., the systemwide policies will
be followed if no explicit override configuration is used. The overrides
should be used only if really needed since they can result in reduced
security.
In wpa_supplicant, tls_disable_tlsv1_?=0 value in the phase1 network
profile parameter can be used to explicitly enable TLS versions that are
disabled in the systemwide configuration. For example,
phase1="tls_disable_tlsv1_0=0 tls_disable_tlsv1_1=0" would request TLS
v1.0 and TLS v1.1 to be enabled even if the systemwide policy enforces
TLS v1.2 as the minimum version. Similarly, openssl_ciphers parameter
can be used to override systemwide policy, e.g., with
openssl_ciphers="DEFAULT@SECLEVEL=1" to drop from security level 2 to 1
in Debian to allow shorter keys to be used.
In hostapd, tls_flags parameter can be used to configure similar
options. E.g., tls_flags=[ENABLE-TLSv1.0][ENABLE-TLSv1.1]
Signed-off-by: Jouni Malinen <j@w1.fi>
SSL_use_certificate_chain_file() was added in OpenSSL 1.1.0, so need to
maintain the old version using SSL_use_certificate_file() for backwards
compatibility.
Fixes: 658c39809b ("OpenSSL: Load chain certificates from client_cert file")
Signed-off-by: Jouni Malinen <j@w1.fi>
The PADDING array used when adding padding bits in MD4 never change
so can be made const. Making it const puts the array in .rodata
section and can save a few bytes of RAM for systems running without
virtual memory.
Signed-off-by: Mikael Kanstrup <mikael.kanstrup@sony.com>
The function sha512_compress() has a local variable that consumes 640
bytes. This is very heavy for embedded devices that have limited stack
resources. Handle this by replacing the static allocation with a dynamic
one.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
According to random(4) manual, /dev/random is essentially deprecated on
Linux for quite some time:
"The /dev/random interface is considered a legacy interface, and
/dev/urandom is preferred and sufficient in all use cases, with the
exception of applications which require randomness during early boot
time; for these applications, getrandom(2) must be used instead, because
it will block until the entropy pool is initialized."
An attempt to use it would cause unnecessary blocking on machines
without a good hwrng even when it shouldn't be needed. Since Linux 3.17,
a getrandom(2) call is available that will block only until the
randomness pool has been seeded.
It is probably not a good default yet as it requires a fairly recent
kernel and glibc (3.17 and 2.25 respectively).
Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
This helps the server to build the chain to trusted CA when PEM encoding
of client_cert is used with multiple listed certificates. This was
already done for the server certificate configuration, but the client
certificate was limited to using only the first certificate in the file.
Signed-off-by: Isaac Boukris <iboukris@gmail.com>
Add the ability to ignore time-based CRL errors from OpenSSL by
specifying a new configuration parameter, check_crl_strict=0.
This causes the following:
- This setting does nothing when CRL checking is not enabled.
- When CRL is enabled, "strict mode" will cause CRL time errors to not
be ignored and will continue behaving as it currently does.
- When CRL is enabled, disabling strict mode will cause CRL time
errors to be ignored and will allow connections.
By default, check_crl_strict is set to 1, or strict mode, to keep
current functionality.
Signed-off-by: Sam Voss <sam.voss@rockwellcollins.com>
Build configurations with CONFIG_TLS=internal and NEED_SHA512 failed due
to missing sha512.c file. Add that file even though this is not really
used in the currently available configuration combinations since DPP and
OWE are the only users of it and the internal crypto implementation
supports neither.
Signed-off-by: Jouni Malinen <j@w1.fi>
