Jouni Malinen [Sun, 20 Dec 2015 17:58:05 +0000 (19:58 +0200)]
GAS: Do not cancel initial offchannel wait with comeback delay 1
The minimum comeback delay 1 is used to indicate that fragmentation is
needed instead of indicating that the response is going to be available
only after some time. Do not cancel offchannel wait for this case
between the initial and comeback exchanges to avoid delaying the full
operation unnecessarily.
Jouni Malinen [Sun, 20 Dec 2015 18:56:34 +0000 (20:56 +0200)]
tests: Update gas_anqp_oom_hapd to match new implementation
Since wpa_supplicant is now retrying GAS comeback failures once, the
gas_anqp_oom_hapd test case started failing. Fix this by updating the
test case to expect success (on the retry).
Jouni Malinen [Sun, 20 Dec 2015 17:26:01 +0000 (19:26 +0200)]
GAS: Retry full GAS query if comeback response is not received
It is possible for a comeback response to get lost especially when going
through a large GAS exchange fragmented to multiple frames in an
environment with interference or other traffic. Make this less likely to
fail the full exchange by trying full GAS query again and using longer
wait time on the GAS comeback exchanges.
Jouni Malinen [Sun, 20 Dec 2015 19:06:21 +0000 (21:06 +0200)]
tests: Modify gas_malformed_comeback_resp to allow GAS retries
This is in preparation of a wpa_supplicant change to allow GAS retries
which can result in the previous test case design showing failures due
to "unexpected" management frames (GAS Initial Request from the retry).
Jouni Malinen [Sun, 20 Dec 2015 17:23:58 +0000 (19:23 +0200)]
GAS server: Replenish temporary STA entry timeout on comeback request
Previously, the five second timeout was added at the beginning of the
full GAS query and it was not replenished during fragmented exchanges.
This could result in timing out a query if it takes significant time to
go through the possibly multiple fragments and long comeback delay.
Matti Gottlieb [Thu, 10 Dec 2015 10:56:09 +0000 (12:56 +0200)]
GAS: Shorten the duration of the wait for GAS comeback response
When exchanging GAS frames with the AP, the initial response from the AP
may take a while to come, since the AP may need to fetch the info from a
server. The next fragments/comeback response frames should take much
less time since the AP already has all of the info, so the wait time
for these frames can be reduced.
In addition, some drivers, e.g., mac80211, try to combine ROC based flows,
to improve medium utilization. For example, if the requested ROC fits
entirely in a previous requested ROC they can be combined. Thus, reducing
the wait time for the next frames can improve medium utilization.
Shorten the duration of GAS comeback to improve medium utilization and
overall GAS exchange times.
Signed-off-by: Matti Gottlieb <matti.gottlieb@intel.com>
Jouni Malinen [Sun, 20 Dec 2015 17:37:06 +0000 (19:37 +0200)]
GAS: Clear offchannel_tx_started when ending remain-on-channel
Commit 2c0d0ae370f72fbe6248feed7ea0635303e3dc5a ('GAS: End
remain-on-channel due to delayed GAS comeback request') started ending
the remain-on-channel operation between the initial request and the
following comeback request. However, it did not check or update the
offchannel_tx_started variable. While this alone would not necessarily
be problematic, this makes it more difficult to optimize wait time for
offchannel TX operations, so make sure the internal tracking variable
gets updated.
Jouni Malinen [Sun, 20 Dec 2015 11:40:59 +0000 (13:40 +0200)]
Drop any pending EAPOL RX frame when starting a new connection
Such a pending frame cannot be valid anymore, so drop it instead of
risking of using an unexpected EAPOL frame after association if a
previous association received one at the end and the new association can
happen within 100 ms.
Jouni Malinen [Sun, 20 Dec 2015 10:59:19 +0000 (12:59 +0200)]
tests: Fix scan result clearing in ssid_hidden*
These test cases were supposed to clear the cfg80211 and wpa_supplicant
scan caches in the end to avoid causing issues to the following test
cases. This did not work properly after introduction of the support for
aborting a pending scan. Fix this by using the flush_scan_cache()
function and waiting within the test case until the final scan operation
completes.
This issue was triggered by ssid_hidden/ssid_hidden2 followed by
ext_password_interworking (though, not every time).
Jouni Malinen [Sun, 20 Dec 2015 08:52:30 +0000 (10:52 +0200)]
WPA: Explicitly clear the buffer used for decrypting Key Data
When AES-WRAP was used to protect the EAPOL-Key Key Data field, this was
decrypted using a temporary heap buffer with aes_unwrap(). That buffer
was not explicitly cleared, so it was possible for the group keys to
remain in memory unnecessarily until the allocated area was reused.
Clean this up by clearing the temporary allocation explicitly before
freeing it.
Jouni Malinen [Sat, 19 Dec 2015 18:22:43 +0000 (20:22 +0200)]
EAP-PEAP server: Add support for fast-connect crypto binding
IPMK and CMK are derived from TK when using TLS session resumption with
PEAPv0 crypto binding. The EAP-PEAP peer implementation already
supported this, but the server side did not.
Jouni Malinen [Sat, 19 Dec 2015 16:53:27 +0000 (18:53 +0200)]
HS 2.0: Remove duplicate icon entries
Only one of the icon entries with a matching BSSID and file name can be
fetched from wpa_supplicant and as such, there is no need to maintain
the old data if it was not explicitly deleted before running a new fetch
for the same BSSID and icon. Remove older duplicated entries whenever
completing a pending icon fetch to optimize memory use.
Jan Nordqvist [Fri, 11 Dec 2015 01:01:46 +0000 (17:01 -0800)]
HS 2.0: Add a command to retrieve icon with in-memory storage
This adds a new command based Hotspot 2.0 icon retrieval option.
In short, here is the new command sequence:
1. REQ_HS20_ICON <bssid> <file-name>
2. event: RX-HS20-ICON <bssid> <file-name> <size>
3. GET_HS20_ICON <bssid> <file-name> <offset> <size>
(if needed, repeat with larger offset values until full icon is
fetched)
5. DEL_HS20_ICON <bssid> <file-name>
REQ_HS20_ICON is a new command that is analogous to HS20_ICON_REQUEST
with the slight difference that an entry to store the icon in memory is
prepared. The RX-HS20-ICON event has been augmented with BSSID,
file-name and size, and GET_HS20_ICON is used to retrieve a chunk of up
to <size> bytes of icon data at offset <offset>. Each chunk is returned
as a base64 encoded fragment, preceded by "HS20-ICON-DATA", BSSID, and
file-name as well as the starting offset of the data.
If there is no entry prepared for the icon when the ANQP result comes
back, hs20_process_icon_binary_file falls back to legacy behavior.
Finally the DEL_HS20_ICON command deletes (all) icons associated with
BSSID and file-name (there could be several if retries are used and they
have different dialog tokens).
Johannes Berg [Thu, 17 Dec 2015 13:54:11 +0000 (15:54 +0200)]
rfkill: Match only the correct expected wiphy rfkill
On systems that have multiple WLAN rfkill instances, the rfkill code
can become confused into thinking that the device was unblocked when
in fact it wasn't, because it only matches on the WLAN type.
Since it then stores the new (unblocked) state from the wrong rfkill
instance, it will never retry the failing IFF_UP operation and the
user has to toggle rfkill again, or otherwise intervene manually, in
this case to get back to operational state.
Fix this by using the existing (but unused) ifname argument when the
rfkill instance is created to match to a specific rfkill index only.
As a P2P Device interface does not have a netdev interface associated
with it, use the name of a sibling interface to initialize the rfkill
context for the P2P Device interface. For nl80211, as the wiphy index
is known only after getting the driver capabilities from the kernel,
move the initialization of the rfkill object to
wpa_driver_nl80211_finish_drv_init().
Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Ilan Peer [Thu, 17 Dec 2015 13:54:10 +0000 (15:54 +0200)]
nl80211: Handle rfkill for P2P Device interface
The rfkill processing in nl80211 driver assumes that the
INTERFACE_ENABLED/DISABLED will be also issued, so does not do much in
the rfkill callbacks. However, as a P2P Device interface is not
associated with a network interface, these events are not issued for it.
Handle rfkill events for the P2P_DEVICE interface by faking the
INTERFACE_ENABLED/DISABLED.
Signed-off-by: Moshe Benji <Moshe.Benji@intel.com> Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Jouni Malinen [Fri, 18 Dec 2015 20:22:25 +0000 (22:22 +0200)]
tests: Prepare ap_open_out_of_memory for modified rfkill init path
The rfkill initialization will be moved out from
wpa_driver_nl80211_drv_init() which would break one step in this OOM
test case due to the memory allocation not existing anymore. Fix this by
skipping that OOM step to avoid causing false failures with the
following commits.
Dmitry Shmidt [Wed, 16 Dec 2015 22:18:18 +0000 (14:18 -0800)]
wpa_supplicant: Add SIGNAL_MONITOR command
SIGNAL_MONITOR THRESHOLD=DD HYSTERESIS=DD command will request signal
strength monitoring events based on there having been requested amount
of drop in the signal strength. The threshold value is the RSSI
threshold in dBm for the event to be sent. 0 threshold can be used to
disable monitoring. The hysteresis value is RSSI hysteresis in dB to
specify the minimum amount of change before a consecutive event is
reported.
With nl80211 driver interface, these values map to the
NL80211_CMD_SET_CQM command with NL80211_ATTR_CQM_RSSI_THOLD and
NL80211_ATTR_CQM_RSSI_HYST attributes to the driver.
This command cannot be used when bgscan module is in use since that
depends on being able to control the connection monitoring parameters.
Matti Gottlieb [Thu, 10 Dec 2015 10:56:08 +0000 (12:56 +0200)]
GAS: End remain-on-channel due to delayed GAS comeback request
During the sequence of exchanging GAS frames with the AP, the AP can
request to come back in X amount of time and resend the GAS request.
Previously, wpa_supplicant did not terminate the remain-on-channel
session, but rather waited until the requested comeback delay had
expired, and then tried to send the GAS frame (potentially to save the
time that is required to schedule a new remain on channel flow).
This might cause unnecessary idle time (can be close to 1000 ms) in
which the device might be off-channel. Ending the current
remain-on-channel session and then rescheduling makes better usage of
the time in this case.
End remain-on-channel session due to receiving a delayed GAS comeback
request from the AP.
Signed-off-by: Matti Gottlieb <matti.gottlieb@intel.com>
Arik Nemtsov [Thu, 10 Dec 2015 10:56:07 +0000 (12:56 +0200)]
TDLS: Ignore incoming TDLS Setup Response retries
The Setup Response timer is relatively fast (500 ms) and there are
instances where it fires on the responder side after the initiator has
already sent out the TDLS Setup Confirm frame. Prevent the processing of
this stale TDLS Setup Response frame on the initiator side.
Eliad Peller [Thu, 10 Dec 2015 10:56:06 +0000 (12:56 +0200)]
wpa_cli: Send ALL_STA command to the correct interface
wpa_ctrl_command_sta(), called by the "ALL_STA" handler, didn't consider
ifname_prefix, resulting in various commands being sent to the global
control interface, rather than the specified interface when IFNAME=
prefix was used.
This in turn caused the unexpected "UNKNOWN COMMAND" result be
considered as valid station, resulting in infinite loop while trying to
get all stations.
Fix it by considering ifname_prefix, similarly to _wpa_ctrl_command().
Ayala Beker [Thu, 10 Dec 2015 10:56:05 +0000 (12:56 +0200)]
Disconnect before trying to switch to a different network
Previously, when wpa_supplicant received bgscan results with a preferred
network, it connected to that network without disconnecting from the
previous one. This might result in an inconsistent state of upper
layers.
Fix this by disconnecting from the current AP before connecting to the
new one when the network profile changes and there is an existing
connection.
Arik Nemtsov [Thu, 10 Dec 2015 10:56:04 +0000 (12:56 +0200)]
Avoid network selection from scan during connection
If scan results arrive during the connection process, the network
selection function was called, interrupting the current connection.
While a regular scan is mutually exclusive with connection establishment
via the nature of radio work, there's no such protection for scheduled
scan. Prevent network selection while a connection is in progress.
Ayala Beker [Thu, 10 Dec 2015 10:56:02 +0000 (12:56 +0200)]
P2P: Fix possible NULL pointer dereference
Fix wpas_p2p_invite() to call p2p_set_own_pref_freq_list() after the
NULL check, to avoid NULL pointer dereference if P2P initialization were
to have failed or P2P module getting deinitialized.
Pali Rohár [Sun, 6 Dec 2015 11:01:32 +0000 (12:01 +0100)]
EAP-TTLS peer: Fix parsing auth= and autheap= phase2 params
This patch fixes an issue with an invalid phase2 parameter value
auth=MSCHAPv2 getting interpreted as auth=MSCHAP (v1) which could
degrade security (though, only within a protected TLS tunnel). Now when
invalid or unsupported auth= phase2 parameter combinations are
specified, EAP-TTLS initialization throws an error instead of silently
doing something.
More then one auth= phase2 type cannot be specified and also both auth= and
autheap= options cannot be specified.
Parsing phase2 type is case sensitive (as in other EAP parts), so phase2
parameter auth=MSCHAPv2 is invalid. Only auth=MSCHAPV2 is correct.
Signed-off-by: Pali Rohár <pali.rohar@gmail.com>
[Use cstr_token() to get rid of unnecessary allocation; cleanup] Signed-off-by: Jouni Malinen <j@w1.fi>
Matt Woods [Sat, 12 Dec 2015 10:25:42 +0000 (10:25 +0000)]
atheros: Unify memory processing functions
There are two types of memory processing functions in the file
atheros_driver.c, such as memory and os_memory. Unify the processing
functions into one type which has the prefix "os_".
Jouni Malinen [Thu, 17 Dec 2015 16:37:19 +0000 (18:37 +0200)]
mesh: Fix VHT Operation information in peering messages
The full VHT channel information was not set in the hostapd data
structures which resulted in incorrect information (all zeros) being
used when building the VHT Operation element for peering messages while
the actual driver mode was set with the full details. We did not seem to
use the VHT information from peering messages, so this does not change
behavior with another wpa_supplicant-based mesh implementation. Anyway,
these elements should match the ones used in Beacon frames.
Jouni Malinen [Thu, 17 Dec 2015 09:45:48 +0000 (11:45 +0200)]
TLS: Report OCSP rejection cases when no valid response if found
This adds a CTRL-EVENT-EAP-TLS-CERT-ERROR and CTRL-EVENT-EAP-STATUS
messages with 'bad certificate status response' for cases where no valid
OCSP response was received, but the network profile requires OCSP to be
used.
Jouni Malinen [Wed, 16 Dec 2015 23:45:51 +0000 (01:45 +0200)]
TLS: Process OCSP SingleResponse(s)
This completes OCSP stapling support on the TLS client side. Each
SingleResponse value is iterated until a response matching the server
certificate is found. The validity time of the SingleResponse is
verified and certStatus good/revoked is reported if all validation step
succeed.
Jouni Malinen [Wed, 16 Dec 2015 22:38:40 +0000 (00:38 +0200)]
TLS: Parse and validate BasicOCSPResponse
This adds the next step in completing TLS client support for OCSP
stapling. The BasicOCSPResponse is parsed, a signing certificate is
found, and the signature is verified. The actual sequence of OCSP
responses (SignleResponse) is not yet processed in this commit.
Jouni Malinen [Mon, 14 Dec 2015 15:23:47 +0000 (17:23 +0200)]
tests: Make key-lifetime-in-memory more robust for GTK check
The decrypted copy of a GTK from EAPOL-Key is cleared from memory only
after having sent out CTRL-EVENT-CONNECTED. As such, there was a race
condition on the test case reading the wpa_supplicant process memory
after the connection. This was unlikely to occur due to the one second
sleep, but even with that, it would be at least theorically possible to
hit this race under heavy load (e.g., when using large number of VMs to
run parallel testing). Avoid this by running a PING command to make sure
wpa_supplicant has returned to eloop before reading the process memory.
This should make it less likely to report false positives on GTK being
found in memory.
Jouni Malinen [Mon, 14 Dec 2015 10:31:35 +0000 (12:31 +0200)]
TLS: Parse OCSPResponse to extract BasicOCSPResponse
This adds the next step for OCSP stapling. The received OCSPResponse is
parsed to get the BasicOCSPResponse. This commit does not yet process
the BasicOCSPResponse.
Jouni Malinen [Sun, 13 Dec 2015 21:11:32 +0000 (23:11 +0200)]
TLS: Parse CertificateStatus message
This allows the internal TLS client implementation to accept
CertificateStatus message from the server when trying to use OCSP
stapling. The actual OCSPResponse is not yet processed in this commit,
but the CertificateStatus message is accepted to allow the TLS handshake
to continue.
Jouni Malinen [Sun, 13 Dec 2015 20:46:25 +0000 (22:46 +0200)]
TLS: Add status_request ClientHello extension if OCSP is requested
This allows the internal TLS implementation to request server
certificate status using OCSP stapling. This commit is only adding code
to add the request. The response is not yet used.
Jouni Malinen [Sun, 13 Dec 2015 18:52:43 +0000 (20:52 +0200)]
TLS: Add minimal support for PKCS #12
This allows the internal TLS implementation to parse a private key and a
certificate from a PKCS #12 file protected with
pbeWithSHAAnd3-KeyTripleDES-CBC.
Jouni Malinen [Sat, 12 Dec 2015 16:16:54 +0000 (18:16 +0200)]
EAP peer: External server certificate chain validation
This adds support for optional functionality to validate server
certificate chain in TLS-based EAP methods in an external program.
wpa_supplicant control interface is used to indicate when such
validation is needed and what the result of the external validation is.
This external validation can extend or replace the internal validation.
When ca_cert or ca_path parameter is set, the internal validation is
used. If these parameters are omitted, only the external validation is
used. It needs to be understood that leaving those parameters out will
disable most of the validation steps done with the TLS library and that
configuration is not really recommend.
By default, the external validation is not used. It can be enabled by
addingtls_ext_cert_check=1 into the network profile phase1 parameter.
When enabled, external validation is required through the CTRL-REQ/RSP
mechanism similarly to other EAP authentication parameters through the
control interface.
The request to perform external validation is indicated by the following
event:
CTRL-REQ-EXT_CERT_CHECK-<id>:External server certificate validation needed for SSID <ssid>
Before that event, the server certificate chain is provided with the
CTRL-EVENT-EAP-PEER-CERT events that include the cert=<hexdump>
parameter. depth=# indicates which certificate is in question (0 for the
server certificate, 1 for its issues, and so on).
The result of the external validation is provided with the following
command:
CTRL-RSP-EXT_CERT_CHECK-<id>:<good|bad>
It should be noted that this is currently enabled only for OpenSSL (and
BoringSSL/LibreSSL). Due to the constraints in the library API, the
validation result from external processing cannot be reported cleanly
with TLS alert. In other words, if the external validation reject the
server certificate chain, the pending TLS handshake is terminated
without sending more messages to the server.
Jouni Malinen [Sat, 12 Dec 2015 10:00:28 +0000 (12:00 +0200)]
EAP-FAST peer: Fix PAC parser error messages
Do not override the parsing error with the "PAC block not terminated
with END" message if the reason for the END line not yet being seen is
in failure to parse an earlier line.
Jouni Malinen [Thu, 10 Dec 2015 16:11:39 +0000 (18:11 +0200)]
OpenSSL: Support new API for HMAC/EVP_MD_CTX in OpenSSL 1.1.x-pre1
The EVP_MD_CTX and HMAC_CTX definitions are now hidden from applications
using OpenSSL. Fix compilation issues with OpenSSL 1.1.x-pre1 by using
the new API for allocating these structures.
Jouni Malinen [Wed, 9 Dec 2015 22:12:59 +0000 (00:12 +0200)]
FT: Fix FTIE generation for EAPOL-Key msg 3/4
This FTIE needs to be an exact copy of the one in (Re)Association
Response frame. Copy the stored element rather than building a new copy
that would not have the correct MIC value. This is needed to fix PTK
rekeying after FT protocol run.
Jouni Malinen [Wed, 9 Dec 2015 22:11:00 +0000 (00:11 +0200)]
FT: Fix sm->assoc_resp_ftie storing on the AP side
The FTIE from (Re)Association Response frame was copied before
calculating the MIC. This resulted in incorrect value being used when
comparing the EAPOL-Key msg 2/4 value in case PTK rekeying was used
after FT protocol run. Fix this by storing the element after the MIC
field has been filled in.
Jouni Malinen [Wed, 9 Dec 2015 22:06:06 +0000 (00:06 +0200)]
FT: Fix FTIE generation for 4-way handshake after FT protocol run
wpa_insert_pmkid() did not support cases where the original RSN IE
included any PMKIDs. That case can happen when PTK rekeying through
4-way handshake is used after FT protocol run. Such a 4-way handshake
used to fail with wpa_supplicant being unable to build the EAPOL-Key msg
2/4.
Fix this by extending wpa_insert_pmkid() to support removal of the old
PMKIDs, if needed.
Jouni Malinen [Sun, 6 Dec 2015 18:31:18 +0000 (20:31 +0200)]
tests: Make pmksa_cache_opportunistic_connect more robust
This test case could fail if the cfg80211 scan cache brought in a BSS
entry from an earlier test case and a new scan did not get executed
prior to the ROAM command. Fix this by forcing the scan to go through
prior to roaming to AP2
This issue showed up with the following test case sequence:
connect_cmd_roam pmksa_cache_opportunistic_connect
Jouni Malinen [Sun, 6 Dec 2015 17:30:59 +0000 (19:30 +0200)]
Add TEST_ASSOC_IE for WPA/RSN IE testing on AP side
The new wpa_supplicant control interface command "TEST_ASSOC_IE
<hexdump>" can now be used to override the WPA/RSN IE for Association
Request frame and following 4-way handshake to allow protocol testing of
AP side processing of WPA/RSN IE.
Jouni Malinen [Sun, 6 Dec 2015 15:48:43 +0000 (17:48 +0200)]
tests: Write GTK locations into debug log in key_lifetime_in_memory
It looks like it is possible for the GTK to be found from memory every
now and then. This makes these test cases fail. Write the memory
addresses in which the GTK was found to the log to make it somewhat
easier to try to figure out where the key can be left in memory.
Jouni Malinen [Sun, 6 Dec 2015 15:10:06 +0000 (17:10 +0200)]
tests: Make dbus_wps_oom more robust
It was possible for the BSSs object property change signal to be
generated during the OOM test case for Get(). If that happened, the
signal was not sent out, but the following Get(BSSs) operation succeeded
unexpectedly which resulted in a test failure. Make this less likely to
happen by waiting 50 ms between the scan and Get(BSSs) operation. This
should be sufficient to cover most cases since wpa_supplicant uses 5 ms
timeout for D-Bus property changed updates.
Jouni Malinen [Sun, 6 Dec 2015 11:35:40 +0000 (13:35 +0200)]
tests: Make ap_max_num_sta_no_probe_resp more robust
It was possible for the AP's Beacon frame to be seen by dev[0] when
running a scan. This is not an error case. Make this test case more
robust by verifying with a sniffer whether a Probe Response frame was
sent to unexpected STA.