This allows libeap.a and libeap.so to be built by merging in multiple
libraries from src subdirectories. In addition, this avoids wasting
extra space and time for local builds.
This fixes issues with some GnuTLS versions that seem to be adding
quite a bit of extra data into TLS messages. The EAP server code is
now using the same 300 byte extra room that was already used in the
EAP peer implementation.
The following defines are not really needed in most places, so
remove them to clean up source code and build scripts:
EAP_TLS_FUNCS
EAP_TLS_OPENSSL
EAP_TLS_GNUTLS
CONFIG_TLS_INTERNAL
There seem to be some cases in which wpa_supplicant and
cfg80211/mac80211 seem to have different understanding on
authentication/association state. Since cfg80211/mac80211 is very strict
on when it accepts new authentication/association/scan commands, try our
best at clearing such state mismatches by explicitly deauthenticating
from BSSes with which the driver claims we are associated with if we do
not have local information about such association.
Print what the kernel believes the current BSS status (authenticated
or associated) is in scan results. In addition, check whether this
matches with the state that wpa_supplicant believes the driver to be
in.
This does not change the actual behavior, but will provide information
that will help in debugging potential issues where cfg80211/mac80211
seems to get into a different state from wpa_supplicant. In addition,
this provides an easy location for a workaround that could be added to
clear cfg80211/mac80211 state for unknown BSSes.
This is an initial step in further cleaning up the EAPOL authenticator
use to avoid requiring direct accesses to the internal data structures.
For now, number of external files are still including the internal
definitions from eapol_auth_sm_i.h, but eventually, these direct
references should be removed.
In addition, start ordering header file includes to be in more
consistent order: system header files, src/utils, src/*, same
directory as the *.c file.
This removes the hardcoded definition from Makefile and cleans up
source code by moving the mail HOSTAPD_DUMP_STATE blocks into separate
files to avoid conditional compilation within files.
While this may not include knowledge of all EAP methods since this
depends on build configuration, it is better to not have to include
ieee802_1x.h into eapol_sm.c.
This makes it clearer which files are including header from src/common.
Some of these cases should probably be cleaned up in the future not to
do that.
In addition, src/common/nl80211_copy.h and wireless_copy.h were moved
into src/drivers since they are only used by driver wrappers and do not
need to live in src/common.
This avoids passing the raw scan results into the RSN code and by
doing so, removes the only dependency on src/drivers from the
src/rsn_supp code (or from any src subdirectory for that matter).
Instead of using hardcoded maximum A-MPDU length of 64 kB and no
restrictions on minimum MPDU Start Spacing, use the correct values
reported by the driver.
This is not actually used at all and it looks like the rules for
maintaining the old/new RADIUS configuration are not very clear in the
case the RADIUS client configuration did not change. Consequently, it
is better to just remove this for now and if similar functionality is
ever needed, redesign it to be easier to use without causing hard to
find issues with using freed memory.
Simpler approach to reconfiguring the RADIUS client would involve
just deinitializing the old context unconditionally and initializing
a new one whenever the configuration could have changed.
Use shared functions for converting Primary Device Type between binary
and string formats. In addition, use array of eight octets instead of a
specific structure with multiple fields to reduce code complexity.
Mainly, this is including header files to get definitions for functions
which is good to verify that the parameters match. None of these are
issues that would have shown as incorrect behavior of the program.
This gets rid of previously deprecated driver_ops handlers set_wpa,
set_drop_unencrypted, set_auth_alg, set_mode. The same functionality
can be achieved by using the init/deinit/associate handlers.
These driver wrappers should not be used anymore; WEXT should be used
instead. However, there may still be users stuck on older kernel versions
that may require driver specific wrappers, so the source code still
remains in the repository.
This moves the MTU definition into driver_hostap.c since it was really
meant to be specific to this driver. Since this was the last remaining
definition in hostapd_defs.h, remove that header file as unnecessary.
This code was copied from driver_hostap.c where it is used with the
special wlan#ap interface. It was not supposed to be used to change
the MTU for a normal data interface.
Clean up driver_nl80211.c by gettign rid of the last remaining WEXT use.
This requires that a recent mac80211 version is used to get full protection
in station mode via the authorized flag (IEEE 802.1X PAE).
Clean up driver interface by merging hostapd and wpa_supplicant
specific set_beacon driver_ops into a single one. In addition,
merge set_beacon_int into to the same operation.
If initialization fails, we could potentially try to sendto() on -1
socket which would fail. No point in doing that, so just return early
from the function.
The current MinGW/w32api versions seem to provide all the needed CryptoAPI
functions, so the code for loading these dynamically from the DLL can be
removed.
Need to define the workspace buffer properly to allow compiler to handle
strict aliasing between the incoming unsigned char[64] buffer as an u32
array. The previous version built with strict aliasing enabled can
result in SHA-1 producing incorrect results and consequently, with
4-way handshake failing.
This is based on a report and patch from Dan Williams <dcbw@redhat.com>
but with a different type (the union) used as a fix to avoid needing
extra type casting.
Discovered as part of the investigation of:
https://bugzilla.redhat.com/show_bug.cgi?id=494262#c32
if sha1 is built with gcc without turning off strict aliasing, it will
fail to correctly generate the hashes and will fail its own testcases as
well.
Signed-off-by: Dan Williams <dcbw@redhat.com>
This avoids some issues in cases where the ER has been started and
stopped multiple times on the same address and an AP may have stored
multiple event notification addresses for the same ER. The random
identifier allows the ER to filter out unexpected messages from further
processing.
Do not give the allocated memory to the subscription code since it was
not using it as-is anyway. This makes it easier to understand who owns
the allocation an is responsible of freeing it. This may potentially
fix some memory leaks on error paths.
The WPS 1.0h specification is quite unclear on what exactly should be
used as the MAC Address value in the Credential and AP Settings. It
looks like this should after all be the MAC Address of the Enrollee,
so change Registrar implementation to use that address instead of the
AP BSSID.
In addition, add validation code to the Enrollee implementation to
check the MAC Address value inside Credential (and also inside AP Settings)
to make sure it matches with the Enrollee's own address. However, since
there are deployed implementations that do not follow this interpretation
of the spec, only show the mismatch in debug information to avoid breaking
interoperability with existing devices.
This will need some additional code in wps_er_pbc() to handle PBC mode
enabling for a single AP only. For now, this can only be expected to work
when the ER is connected to a single AP.
UPnP device architecture specification requires all evented variables to
be included in the initial event message after subscription. Since this
can happen before we have seen any events, generated a dummy event
(WSC_ACK with all-zeros nonces) if needed.
This will notify control points of the services going away and allows
them to notice this without having to wait timeout on the
initial advertisements.
This removes following WFA WLANConfig Service actions and the related
state variables: GetAPSettings, SetAPSettings, DelAPSettings,
GetSTASettings, SetSTASettings, DelSTASettings, RebootAP,
ResetAP, RebootSTA, ResetSTA.
While WFA WLANConfig Service version 1.0 claims that some of these are
mandatory to implement for an AP, there are no known implementations
supporting these actions neither in an AP/proxy or an External Registrar
that would use them. These are unlikely to be supported in the future
either and as such, it is just simpler to get rid of them to clean up
the implementation and reduce code size.
When using UPnP transport, the Op-Code is not included, but the WPS
frame processing will need this. Generate a matching Op-Code based
on the message type.
Previously, WSC_MSG was hardcoded for every message from ER, but
this needs to be changed based on message type to send a valid
message to the Enrollee via EAP transport.
New PINs can now be added to WPS ER. This results in the ER code
using SetSelectedRegistrar to modify AP state so that Enrollees
will be able to notice the actice registrar more easily.
This keeps STA/Enrollee entries up to date and sets up registration
protocol session. M1 is processed and M2D generated, but the there
is no code yet to transmit the response back to the AP with
PutWLANResponse.
Always bring down the wlan interface, even when not changing the
BSSID, the interface also needs to be down for changing its type
from managed to AP mode.
The receive Probe Request and EAP-WSC notifications are now parsed
(including the TLVs in them) and contents is shown in the debug log.
Actual processing of the received information is still missing (TODO
comments indicate the needed functionality).
rand would be the address of rand() function and never NULL. The previous
version could have crashed on invalid AKA-AUTS command. Though, these
commands are only from hostapd which sends valid requests and as such,
the actual issue did not show up.
This is like wpa_msg(), but the output is directed only to
ctrl_interface listeners. In other words, the output will not be
shown on stdout or in syslog.
Change scan result reporting to use wpa_msg_ctrl() for
CTRL-EVENT-SCAN-RESULTS message at info level and wpa_printf() at
debug level to avoid showing scan result events in syslog in the
common configuration used with NetworkManager.
This adds code to start a HTTP server and to subscribe to UPnP events
from each discovered WPS AP. The event messages are received, but there
is not yet any code to actually parse the contents of the event.
Clean up code so that UPnP implementation does not need to include all
the HTTP functionality. In addition, make it easier to share HTTP server
functionality with other components in the future.
Instead of implementing HTTP client functionality inside
wps_upnp_event.c, use a generic HTTP client module to do this. The HTTP
client code can now be shared more easily for other purposes, too.
This is the first step in adding support for using wpa_supplicant as a
WPS External Registrar to manage APs over UPnP. Only the device
discovery part is implemented in this commit.
While the actual use here would be unlikely to be broken by any C
optimization, it is better to use explicit union construction to let
gcc know about the aliasing and avoid warnings from gcc 4.4.
gcc 4.4 ends up generating strict-aliasing warnings about some very common
networking socket uses that do not really result in a real problem and
cannot be easily avoided with union-based type-punning due to struct
definitions including another struct in system header files. To avoid having
to fully disable strict-aliasing warnings, provide a mechanism to hide the
typecast from aliasing for now. A cleaner solution will hopefully be found
in the future to handle these cases.
Though we have such a timeout when handling SetSelectedRegistrar UPnP
message from an external registrar, it looks like we don't have one when
the internal registrar is activated for PIN connection. Thus we set the
SelectedRegistrar flag when AP is activated for PIN connection but we
never reset it - not by some timeout, nor when registration succeeds.
This lead to situations where AP everlastingly declare that it is
activated for WPS PIN connection when in reality it is not.
Use the same timeout (and also success with PIN) to clear the selected
registrar flag when using internal registrar, too.
If PBC session overlap is detected during an ongoing PBC protocol run,
reject the run (if M8, i.e., credentials, have not yet been sent). This
provides a bit longer monitoring time at the Registrar for PBC mode to
catch some cases where two Enrollees in PBC mode try to enroll
credentials at about the same time.
This provides information about PBC mode result from the WPS Registrar
module. This could be used, e.g., to provide a user notification on the
AP UI on PBC failures.
If PBC session overlap is detected between button press on the registrar
and M1 is reception, report session overlap with the Config Error
attribute in M2D to the Enrollee.
This functionality fits better with src/tls (i.e., internal TLS
implementation), so move it there to make crypto_internal.c more
of a wrapper like other crypto_*.c files.
Private keys can now be used in either unencrypted or encrypted
PKCS #8 encoding. Only the pbeWithMD5AndDES-CBC algorithm (PKCS #5)
is currently supported.
Recognize the PEM header "BEGIN PRIVATE KEY" as base64-decode the data
to be able to use PEM encoded, unencrypted PKCS#8 private keys with the
internal TLS implementation. Previously, only DER encoding of the
PKCS#8 private key was supported.
OpenSSL 0.9.7 does not include get_rfc3526_prime_1536() function, so
provide that functionality internally if needed. In addition, make
sha256_vector() building depend on whether SHA256 support is included
in the OpenSSL library. This with CONFIG_INTERNAL_SHA256=y in .config
allows OpenSSL without SHA256 support to be used.
Both hostapd/wpa_supplicant compilation fails on DragonFly BSD.
This patch solves this issue.
I have tested only compilation. Not functionality.
Because I don't have any device which can work on DragonFly BSD.
mac80211 does not currently allow (re)authentication when we are already
authenticated. In order to work around this, force deauthentication if
nl80211 authentication command fails with EALREADY. Unfortunately, the
workaround code in driver_nl80211.c alone is not enough since the
following disconnection event would clear wpa_supplicant authentication
state. To handle this, add some code to restore authentication state
when using userspace SME.
This workaround will hopefully become unnecessary in some point should
mac80211 start accepting new authentication requests even when in
authenticated state.
Reorganize the TLS/crypto library segments into a single set of blocks
for each library instead of multiple locations handling library-specific
operations. Group crypto functionality together and get wpa_supplicant
and hostapd Makefile closer to eachother in order to make it easier to
eventually move this into a shared makefile.
Crypto library wrappers can now override the internal DH (group 5)
implementation. As a starting point, this is done with OpenSSL. The
new mechanism is currently available only for WPS (i.e., IKEv2 still
depends on the internal DH implementation).
Replace "nl80211: Ignored unknown event (cmd=33)" with
"nl80211: Scan trigger" to make debug output clearer. We do not
currently do anything with this event apart from showing it in
the debug log.
This allows NSS to be used to derive EAP-TLS/PEAP/TTLS keying material.
NSS requires a patch from
https://bugzilla.mozilla.org/show_bug.cgi?id=507359
to provide the new API. In addition, that patch needs to be modified to
add the 16-bit context length value in SSL_ExportKeyingMaterial() only if
contextlen != 0 in order to match with the EAP-TLS/PEAP/TTLS use cases.
This issue seems to be coming from the unfortunate incompatibility in
draft-ietf-tls-extractor-07.txt (draft-ietf-tls-extractor-00.txt would
have used compatible PRF construction).
At this point, it is unclear how this will be resolved eventually, but
anyway, this shows a mechanism that can be used to implement EAP key
derivation with NSS with a small patch to NSS.
This brings in the first step in adding support for using NSS
(Mozilla Network Security Services) as the crypto and TLS library
with wpa_supplicant. This version is able to run through EAP-PEAP
and EAP-TTLS authentication, but does not yet implement any
certificate/private key configuration. In addition, this does not
implement proper key fetching functions either, so the end result
is not really of much use in real world yet.
On DragonFly BSD, wired IEEE 802.1X fails with this message:
ioctl[SIOC{ADD/DEL}MULTI]: Invalid argument
This patch solves this issue.
I have tested with these:
OS : DragonFly BSD 2.4.0
EAP : EAP-TLS
Switch : Cisco Catalyst 2950
This reverts commit 5aa9cb5cca.
The nested key attribute is using different attribute values
(NL80211_KEY_* vs. NL80211_ATTR_KEY_*), so cannot share the same routine
for these purposes..
Current wpa_supplicant has a bug with WEP keys, it adds a zero-length
sequence counter field to netlink which the kernel doesn't accept.
Additionally, the kernel API slightly changed to accept keys only when
connected, so we need to send it the keys after that. For that to work
with shared key authentication, we also include the default WEP TX key
in the authentication command.
To upload the keys properly _after_ associating, add a new flag
WPA_DRIVER_FLAGS_SET_KEYS_AFTER_ASSOC_DONE indicating that the driver
needs the keys at that point and not earlier.
Commit 0b55b934ee broke this by not
initializing drv->ap = 1 in hostapd case since the mode updating
code ended up unlinking the socket file. Setting drv->ap = 1
removes the mode change and as such, unlinking of the socket file.
It is simpler to just build in all the test driver code regardless
of whether this is for hostapd or wpa_supplicant (which will eventually
get AP mode support with driver_test, too).
radius_server_encapsulate_eap() resets sess->eap->if->eap{Success,Fail}
to FALSE, such that the completion condition is never true.
The net effect is that completed sessions would linger for
RADIUS_SESSION_TIMEOUT seconds.
Signed-off-by: Alex Badea <vamposdecampos@gmail.com>
Previously, the default settings allowed 100 sessions in 60 seconds.
With this fix, the default limit is now 100 sessions per 10 seconds.
[Bug 329]
Store a copy of device attributes during WPS protocol run and make it
available for external programs via the control interface STA MIB
command for associated stations. This gives access to device name and
type which can be useful when showing user information about associated
stations.
wpa_supplicant can now reconfigure the AP by acting as an External
Registrar with the wps_reg command. Previously, this was only used
to fetch the current AP settings, but now the wps_reg command has
optional arguments which can be used to provide the new AP
configuration. When the new parameters are set, the WPS protocol run
is allowed to continue through M8 to reconfigure the AP instead of
stopping at M7.
This is an initial step in adding support for the new connect command.
For now, we just add the capability query. The actual use of the new
command will be added separately.
This is a patch for OpenBSD wired IEEE 802.1X. This is only for wired,
not wireless, because OpenBSD uses wpa_supplicant only on wired now.
http://www.openbsd.org/cgi-bin/cvsweb/ports/security/wpa_supplicant/
I have tested with these.
OS : OpenBSD 4.5
EAP : EAP-TLS
Switch : CentreCOM 8724SL
Previously, both the command replies and unsolicited events were
received from the same socket. This could cause problems if an event
message is received between a command and the response to that command.
Using two sockets avoids this issue.
When Linux has Path MTU discovery enabled, it sets by default the DF bit
on all outgoing datagrams, also UDP ones. If a RADIUS message is bigger
than the smallest MTU size to the target, it will be discarded.
This effectively limits RADIUS messages to ~ 1500 Bytes, while they can
be up to 4k according to RFC2865. In practice, this can mean trouble
when doing EAP-TLS with many RADIUS attributes besides the EAP-Message.
[Bug 326]
When Linux has Path MTU discovery enabled, it sets by default the DF bit
on all outgoing datagrams, also UDP ones. If a RADIUS message is bigger
than the smallest MTU size to the target, it will be discarded.
This effectively limits RADIUS messages to ~ 1500 Bytes, while they can
be up to 4k according to RFC2865. In practice, this can mean trouble
when doing EAP-TLS with many RADIUS attributes besides the EAP-Message.
[Bug 326]
These could, at least in theory, be used to generate unexpected common
name or subject alternative name matches should a CA sign strings with
NUL (C string termination) in them. For now, just reject the certificate
if an embedded NUL is detected. In theory, all the comparison routines
could be made to compare these strings as binary blobs (with additional
X.509 rules to handle some exceptions) and display NUL characters
somehow. Anyway, just rejecting the certificate will get rid of
potential problems with the C string getting terminated and it should
not really be used in certificates, so this should not break valid use
cases.
The BLOCK_SIZE define can be made more specific by using AES_ prefix and
by moving it to aes.h. After this, most aes-*.c do not really need to
include anything from the internal aes_i.h header file. In other words,
aes_i.h can now be used only for the code that uses the internal AES
block operation implementation and none of the code that can use AES
implementation from an external library do not need to include this
header file.
Better not specify EVP_CIPHER again for the second init call since that
will override key length with the default value. The previous version
was likely to work since most use cases would be likely to use the
default key length. Anyway, better make this handle variable length
ciphers (mainly, RC4), too, just in case it is needed in the future.
This is not really of that much use since rc4_skip() can be used as
easily. In addition, rc4 has caused some symbol conflicts in the past,
so it is easier to live without that as an exported symbol.
wpa_supplicant can now be built with FIPS capable OpenSSL for FIPS mode
operation. Currently, this is only enabling the FIPS mode in OpenSSL
without providing any higher level enforcement in wpa_supplicant.
Consequently, invalid configuration will fail during the authentication
run. Proper configuration (e.g., WPA2-Enterprise with EAP-TLS) allows
the connection to be completed.
Instead of using low level, digest-specific functions, use the generic
EVP interface for digest functions. In addition, report OpenSSL errors
in more detail.
Some crypto libraries can return in these functions (e.g., if a specific
hash function is disabled), so we better provide the caller a chance to
check whether the call failed. The return values are not yet used
anywhere, but they will be needed for future changes.
On NetBSD 5.0, when I use wired 802.1X, "Invalid argument" occurs
on SIOCADDMULTI ioctl and 802.1X fails.
I tried FreeBSD code, but "Address family not supported by protocol family"
occurs on SIOCADDMULTI ioctl and 802.1X fails, too.
This patch solves this issue.
I have tested with these:
OS : NetBSD 5.0
EAP : EAP-MD5
Switch : CentreCOM 8724SL
This is a (hopefully) temporary workaround to allow the same source code
tree to be used for building hostapd and wpa_supplicant without having
to manually force recompilation of some files. Currently, some of the
driver wrapper files need to be built separately for hostapd and
wpa_supplicant (#ifdef's in the files based on AP functionality).
This is somewhat racy as far as parallel make execution is concerned,
i.e., it may be necessary to run "make -j#" twice (plain "make" works
fine. Since this is supposed to be a temporary workaround, there is not
much point in trying to fix this with any more complex make processing.
Instead of having all driver stuff collected across wpa_supplicant
and hostapd, create a common snippet that they both include and
that handles the build configuration.
If base64_encode() were to be used with a huge data array, the
previous version could have resulted in overwriting the allocated
buffer due to an integer overflow as pointed out in
http://www.freebsd.org/cgi/query-pr.cgi?pr=137484. However, there
are no know use cases in hostapd or wpa_supplicant that would do that.
Anyway, the recommended change looks reasonable and provides additional
protection should the base64_encode() function be used for something
else in the future.
