Prevent loading arbitrary executable code based on config at runtime,
while allowing libraries to be specified at compile time when they are
known in advance.
Add the ability to configure libraries to load at compile time.
* CONFIG_PKCS11_ENGINE_PATH - pkcs11_engine library location.
* CONFIG_PKCS11_MODULE_PATH - pkcs11_module library location.
* CONFIG_OPENSC_ENGINE_PATH - opensc_engine library location.
Add flags with the ability to set each of the libraries to NULL and
prevent loading them at runtime.
* CONFIG_NO_PKCS11_ENGINE_PATH - prevents loading pkcs11_engine
library.
* CONFIG_NO_PKCS11_MODULE_PATH - prevents loading pkcs11_module
library.
* CONFIG_NO_OPENSC_ENGINE_PATH - prevents loading opensc_engine
library.
* CONFIG_NO_LOAD_DYNAMIC_EAP - prevents loading EAP libraries at
runtime.
Signed-off-by: David Ruth <druth@chromium.org>
Redesign the identity selection for MK derivation to be done explicitly
based on the last indicated identity (whether it is from
EAP-Response/Identity or method specific AT_IDENTITY) during the current
exchange. This makes the implementation cleaner and avoids cases were
more or less duplicated selection steps ended up being slightly
different. This is not as clean as it could otherwise be due to the
exception needed for the IMSI privacy case where the identity used in MK
derivation is actually not the one exchanged in the EAP messages.
Furthermore, this moves the somewhat confusing EAP method specific
tracking of the lasgt EAP-Response/Identity value from EAP-SIM/AKA into
the main EAP peer implementation.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
A case was missing in the way wpa_supplicant was tracking the identity
that is used when deriving MK if the EAP server does not follow the RFC
guidance on using EAP method specific identity determination (i.e.,
AT_IDENTITY for EAP-SIM/AKA) combined with a fallback from fast
re-authentication to full authentication. wpa_supplicant ended up using
the actual identity instead of the last identity included in an EAP
message even though MK derivation is supposed to use the identity that
was included in the last AT_IDENTITY or in the EAP-Response/Identity if
AT_IDENTITY was not used. This resulted in such an authentication
attempt failing due to incorrect MK being derived and AT_MAC validation
resulting in an mismatch.
Fix this by checking for the case where fast re-authentication is
attempted and the server recognizes the fast re-auth identity, but
decides to fall back to full authentication without a separate EAP
method specific identity exchange using AT_IDENTITY. This allows the
fast re-auth identity from EAP-Response/Identity to be used in MK
derivation in such cases.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
Allow the phase2_auth=2 parameter (in phase1 configuration item) to be
used with EAP-TTLS to require Phase 2 authentication. In practice, this
disables TLS session resumption since EAP-TTLS is defined to skip Phase
2 when resuming a session.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
The previous PEAP client behavior allowed the server to skip Phase 2
authentication with the expectation that the server was authenticated
during Phase 1 through TLS server certificate validation. Various PEAP
specifications are not exactly clear on what the behavior on this front
is supposed to be and as such, this ended up being more flexible than
the TTLS/FAST/TEAP cases. However, this is not really ideal when
unfortunately common misconfiguration of PEAP is used in deployed
devices where the server trust root (ca_cert) is not configured or the
user has an easy option for allowing this validation step to be skipped.
Change the default PEAP client behavior to be to require Phase 2
authentication to be successfully completed for cases where TLS session
resumption is not used and the client certificate has not been
configured. Those two exceptions are the main cases where a deployed
authentication server might skip Phase 2 and as such, where a more
strict default behavior could result in undesired interoperability
issues. Requiring Phase 2 authentication will end up disabling TLS
session resumption automatically to avoid interoperability issues.
Allow Phase 2 authentication behavior to be configured with a new phase1
configuration parameter option:
'phase2_auth' option can be used to control Phase 2 (i.e., within TLS
tunnel) behavior for PEAP:
* 0 = do not require Phase 2 authentication
* 1 = require Phase 2 authentication when client certificate
(private_key/client_cert) is no used and TLS session resumption was
not used (default)
* 2 = require Phase 2 authentication in all cases
Signed-off-by: Jouni Malinen <j@w1.fi>
Return value of crypto_bignum_to_bin() wasn't always checked, resulting
in potential access to uninitialized values. Fix it, as some analyzers
complain about it.
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Micha Hashkes <micha.hashkes@intel.com>
When using the optimized EAP method sequence within the tunnel, crypto
binding for the previous EAP method can be performed in the same message
with the start of the next EAP method. The Crypto-Binding TLV needs to
be processed before moving to the next EAP method for IMSK to be derived
correctly, so swap the order of these processing steps.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
While RFC 7170 does not describe this, EAP-TEAP has been deployed with
implementations that use the EAP-FAST-MSCHAPv2, instead of the
EAP-MSCHAPv2, way of deriving the MSK for IMSK. Use that design here to
interoperate with other implementations since that seems to be direction
that IETF EMU WG is likely to go with an RFC 7170 update.
This breaks interoperability with earlier hostapd/wpa_supplicant
versions.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
EAP-FAST uses a special variant of EAP-MSHCAPv2 called EAP-FAST-MSCHAPv2
in RFC 5422. The only difference between that and EAP-MSCHAPv2 is in how
the MSK is derived. While this was supposed to be specific to EAP-FAST,
the same design has ended up getting deployed with EAP-TEAP as well.
Move this special handling into EAP-MSCHAPv2 implementation so that it
can be shared for both needs.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
Extend IMSI privacy functionality to allow an attribute (in name=value
format) to be added using the new imsi_privacy_attr parameter.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
Use imsi_privacy_cert as the name of the configuration parameter for the
X.509v3 certificate that contains the RSA public key needed for IMSI
privacy. The only allowed format for this information is a PEM-encoded
X.509 certificate, so the previous name was somewhat confusing.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
This is an allocated resource so it needs to be free on the error path.
Fixes: 42871a5d25 ("EAP-SIM/AKA peer: IMSI privacy")
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
Add a notification message to indicate reason for TLS handshake failure
due to the server not supporting safe renegotiation (RFC 5746).
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
The TLS protocol design for renegotiation was identified to have a
significant security flaw in 2009 and an extension to secure this design
was published in 2010 (RFC 5746). However, some old RADIUS
authentication servers without support for this are still used commonly.
This is obviously not good from the security view point, but since there
are cases where the user of a network service has no realistic means for
getting the authentication server upgraded, TLS handshake may still need
to be allowed to be able to use the network.
OpenSSL 3.0 disabled the client side workaround by default and this
resulted in issues connection to some networks with insecure
authentication servers. With OpenSSL 3.0, the client is now enforcing
security by refusing to authenticate with such servers. The pre-3.0
behavior of ignoring this issue and leaving security to the server can
now be enabled with a new phase1 parameter allow_unsafe_renegotiation=1.
This should be used only when having to connect to a network that has an
insecure authentication server that cannot be upgraded.
The old (pre-2010) TLS renegotiation mechanism might open security
vulnerabilities if the authentication server were to allow TLS
renegotiation to be initiated. While this is unlikely to cause real
issues with EAP-TLS, there might be cases where use of PEAP or TTLS with
an authentication server that does not support RFC 5746 might result in
a security vulnerability.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
The internal flag prot_success_received was not cleared between the
sessions and that resulted in the resumed session not mandating the
protected success indication to be received. Fix this by clearing the
internal flag so that the EAP-TLS handshake using session resumption
with TLS 1.3 takes care of the required check before marking the
authentication successfully completed. This will make the EAP-TLS peer
reject an EAP-Success message should it be received without the
protected success indication.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
Add support for IMSI privacy in the EAP-SIM/AKA peer implementation. If
the new wpa_supplicant network configuration parameter imsi_privacy_key
is used to specify an RSA public key in a form of a PEM encoded X.509v3
certificate, that key will be used to encrypt the permanent identity
(IMSI) in the transmitted EAP messages.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
The DH file parameters are applicable only for the TLS server, so this
parameter did not really have any impact to functionality. Remove it to
get rid of useless code and confusing documentation for the network
block configuration.
Signed-off-by: Jouni Malinen <j@w1.fi>
The default behavior in wpa_supplicant is to disable use of TLSv1.3 in
EAP-TLS unless explicitly enabled in network configuration. The new
CONFIG_EAP_TLSV1_3=y build parameter can be used to change this to
enable TLSv1.3 by default (if supported by the TLS library).
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
RFC 9190 requires protected result indication to be used with TLSv1.3,
so do not allow EAP-TLS to complete successfully if the server does not
send that indication.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
While the drafts for RFC 9190 used a separate Commitment Message term,
that term was removed from the published RFC. Update the debug prints to
match that final language.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
The previously used references were pointing to an obsoleted RFC and
draft versions. Replace these with current versions.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
When ERP initialization was moved from the METHOD state to the SUCCESS
state, the conditions for checking against EAP state being cleared was
missed. The METHOD state verified that sm->m is not NULL while the
SUCCESS state did not have such a check. This opened a window for a race
condition where processing of deauthentication event and EAPOL RX events
could end up delivering an EAP-Success to the EAP peer state machine
after the state had been cleared. This issue has now been worked around
in another manner, but the root cause for this regression should be
fixed as well.
Check that the EAP state machine is properly configured before trying to
initialize ERP in the SUCCESS state.
Fixes: 2a71673e27 ("ERP: Derive ERP key only after successful EAP authentication")
Signed-off-by: Jouni Malinen <j@w1.fi>
Build eap_*.so into the wpa_supplicant similarly with the wpa_supplicant
binary and include the shared helper functions from additional files
into the builds. This got broken at some point with the build system
changes.
Signed-off-by: Jouni Malinen <j@w1.fi>
Recognize the explicitly defined Commitment Message per
draft-ietf-emu-eap-tls13-13 at the conclusion of the EAP-TTLS with TLS
1.3.
Signed-off-by: Alexander Clouter <alex@digriz.org.uk>
Recognize the explicitly defined Commitment Message per
draft-ietf-emu-eap-tls13-13 at the conclusion of the EAP-TLS with TLS
1.3.
Signed-off-by: Alexander Clouter <alex@digriz.org.uk>
This newer Session-Id/Method-Id derivation is used with PEAP and
EAP-TTLS when using TLS 1.3 per draft-ietf-emu-tls-eap-types-00, so do
not limit this to only EAP-TLS.
Signed-off-by: Alexander Clouter <alex@digriz.org.uk>
Use the TLS-Exporter with the label and context as defined in
draft-ietf-emu-tls-eap-types-00 when deriving keys for EAP-TTLS with TLS
1.3.
Signed-off-by: Alexander Clouter <alex@digriz.org.uk>
Use the TLS-Exporter with the label and context as defined in
draft-ietf-emu-tls-eap-types-00 when deriving keys for PEAP with TLS
1.3.
Signed-off-by: Alexander Clouter <alex@digriz.org.uk>
EAP peer does not expect data present when beginning the Phase 2 in
EAP-{TTLS,PEAP} but in TLS 1.3 session tickets are sent after the
handshake completes.
There are several strategies that can be used to handle this, but this
patch picks up from the discussion[1] and implements the proposed use of
SSL_MODE_AUTO_RETRY. SSL_MODE_AUTO_RETRY has already been enabled by
default in OpenSSL 1.1.1, but it needs to be enabled for older versions.
The main OpenSSL wrapper change in tls_connection_decrypt() takes care
of the new possible case with SSL_MODE_AUTO_RETRY for
SSL_ERROR_WANT_READ to indicate that a non-application_data was
processed. That is not really an error case with TLS 1.3, so allow it to
complete and return an empty decrypted application data buffer.
EAP-PEAP/TTLS processing can then use this to move ahead with starting
Phase 2.
[1] https://www.spinics.net/lists/hostap/msg05376.html
Signed-off-by: Alexander Clouter <alex@digriz.org.uk>
This could fail in theory if running out of memory, so better check for
this explicitly instead of allowing the exchange to continue and fail
later due to checkcode mismatch.
Signed-off-by: Jouni Malinen <j@w1.fi>
Use a local variable to try to make ikev2_parse_proposal() easier for
static analyzers to understand. Bounds checking in the loop is really
done by the ikev2_parse_transform() function, so the p->num_transforms
value itself is of no importance for that part and even that was already
implicitly limited in range.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
For EAP-SIM connections, reorder the order of the attributes in
EAP-Response/SIM/Start message: Send AT_IDENTITY first, then
AT_NONCE and AT_VERSION instead of AT_IDENTITY last. Even though there
is no order requirements in the RFC, some implementations expect the
order of the attributes to be exactly as described in the RFC figures.
Peer Authenticator
| |
| +------------------------------+
| | Server does not have a |
| | Subscriber identity available|
| | When starting EAP-SIM |
| +------------------------------+
| |
| EAP-Request/SIM/Start |
| (AT_ANY_ID_REQ, AT_VERSION_LIST) |
|<------------------------------------------------|
| |
| |
| EAP-Response/SIM/Start |
| (AT_IDENTITY, AT_NONCE_MT, |
| AT_SELECTED_VERSION) |
|------------------------------------------------>|
| |
Signed-off-by: Hai Shalom <haishalom@google.com>
If wpa_supplicant is built with dynamic EAP methods,
the *.so files land here. Add them to .gitignore.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The install target at the beginning of src/eap_peer/Makefile was
confusing make about the build rules for libeap_peer.a and overriding of
the install target between src/eap_peer/Makefile and src/lib.rules was
breaking installation of dynamic EAP peer *.so files.
Fix this by lib.rules defining a default for the install target so that
src/*/Makefile can override that and by moving the install target for
eap_peer to the end of the Makefile.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
We don't really need to duplicate more of this, so just
move the lib.rules include to the end and do more of the
stuff that's common anyway there.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Derive the library name from the directory name, and let each
library Makefile only declare the objects that are needed.
This reduces duplicate code for the ar call. While at it, also
pretty-print that call.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
This is something I hadn't previously done, but there are
cases where it's needed, e.g., building 'wlantest' and then
one of the tests/fuzzing/*/ projects, they use a different
configuration (fuzzing vs. not fuzzing).
Perhaps more importantly, this gets rid of the last thing
that was dumped into the source directories, apart from
the binaries themselves.
Note that due to the use of thin archives, this required
building with absolute paths.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Instead of building in the source tree, put most object
files into the build/ folder at the root, and put each
thing that's being built into a separate folder.
This then allows us to build hostapd and wpa_supplicant
(or other combinations) without "make clean" inbetween.
For the tests keep the objects in place for now (and to
do that, add the build rule) so that we don't have to
rewrite all of that with $(call BUILDOBJS,...) which is
just noise there.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The EAP-TEAP server may skip Phase 2 if the client authentication could
be completed during Phase 1 based on client certificate. Handle this
similarly to the case of PAC use.
Signed-off-by: Jouni Malinen <j@w1.fi>
EAP-SIM full authentication starts with one or more SIM/Start rounds, so
reject an unexpected SIM/Challenge round without any preceeding
SIM/Start rounds to avoid unexpected behavior. In practice, an attempt
to start with SIM/Challenge would have resulted in different MK being
derived and the Challenge message getting rejected due to mismatching
AT_MAC unless the misbehaving server has access to valid Kc, so the end
result is identical, but it is cleaner to reject the unexpected message
explicitly to avoid any risk of trying to proceed without NONCE_MT.
Signed-off-by: Jouni Malinen <j@w1.fi>
Allow any pointer to be used as source for encoding and use char * as
the return value from encoding and input value for decoding to reduce
number of type casts needed in the callers.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
This code block with dependency on PCSC_FUNCS was missed when conf->pin
was moved to conf->cert.pin. Fix this to get rid of compilation issues
with CONFIG_PCSC=y builds.
Fixes: b99c4cadb7 ("EAP peer: Move certificate configuration params into shared struct")
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
This is needed to allow clean transition from one inner EAP
authentication method to another one if EAP method negotiation is needed
within Phase 2.
Signed-off-by: Jouni Malinen <j@w1.fi>
This allows EAP-TLS to be used within an EAP-TEAP tunnel when there is
an explicit request for machine credentials. The network profile
parameters are otherwise same as the Phase 1 parameters, but each one
uses a "machine_" prefix for the parameter name.
Signed-off-by: Jouni Malinen <j@w1.fi>
OCSP configuration is applicable to each instance of TLS-based
authentication and as such, the configuration might need to be different
for Phase 1 and Phase 2. Move ocsp into struct eap_peer_cert_config and
add a separate ocsp2 network profile parameter to set this for Phase 2.
Signed-off-by: Jouni Malinen <j@w1.fi>
