Jouni Malinen [Fri, 15 Jan 2016 14:27:13 +0000 (16:27 +0200)]
tests: Fix wpas_ctrl_oom
The OpenSSL memory allocation changes broke this test case. Fix this by
removing the cases that do not get triggered anymore and add a separate
wpas_ctrl_error test case to cover the fail_test() versions of errors.
Jouni Malinen [Fri, 15 Jan 2016 12:17:16 +0000 (14:17 +0200)]
OpenSSL: Clean up openssl_digest_vector() to use a single implementation
Use compatibility wrapper functions to allow a single implementation
based on the latest OpenSSL API to be used to implement these functions
instead of having to maintain two conditional implementation based on
the library version.
Jouni Malinen [Fri, 15 Jan 2016 12:14:07 +0000 (14:14 +0200)]
OpenSSL: Clean up crypto_hash_*() to use a single implementation
Use compatibility wrapper functions to allow a single implementation
based on the latest OpenSSL API to be used to implement these functions
instead of having to maintain two conditional implementation based on
the library version.
Jouni Malinen [Fri, 15 Jan 2016 12:06:46 +0000 (14:06 +0200)]
LibreSSL: Fix build with LibreSSL
The changes needed for OpenSSL 1.1.0 had broken this since LibreSSL is
defining OPENSSL_VERSION_NUMBER in a manner that claims it to be newer
than the current OpenSSL version even though it does not support the
current OpenSSL API.
Jouni Malinen [Fri, 15 Jan 2016 11:37:15 +0000 (13:37 +0200)]
EAP-TTLS peer: Fix success after fragmented final Phase 2 message
If the final Phase 2 message needed fragmentation, EAP method decision
was cleared from UNCOND_SUCC or COND_SUCC to FAIL and that resulted in
the authentication failing when the EAP-Success message from the server
got rejected. Fix this by restoring the EAP method decision after
fragmentation.
Jouni Malinen [Thu, 14 Jan 2016 17:46:14 +0000 (19:46 +0200)]
tests: fail_test instead of alloc_fail for aes_{encrypt,decrypt}_init
This is needed to fix ap_wpa2_eap_psk_oom, ap_wpa2_eap_sim_oom,
eap_proto_psk_errors, and ap_ft_oom with the new OpenSSL dynamic memory
allocation design.
Jouni Malinen [Thu, 14 Jan 2016 17:44:47 +0000 (19:44 +0200)]
tests: aes_encrypt_init() and aes_decrypt_init() to use TEST_FAIL
Now the these functions cannot be made to fail by forcing the memory
allocation fail since the OpenSSL-internal version is used, add
TEST_FAIL check to allow OOM test cases to be converted to use the
TEST_FAIL mechanism without reducing coverage.
Jouni Malinen [Thu, 14 Jan 2016 17:30:24 +0000 (19:30 +0200)]
OpenSSL: Use EVP_CIPHER_CTX_new() to work with OpenSSL 1.1.0
The EVP_CIPHER_CTX structure will be made opaque in OpenSSL 1.1.0, so
need to use EVP_CIPHER_CTX_new() with it instead of stack memory. The
design here moves the older OpenSSL versions to use that dynamic
allocation design as well to minimize maintenance effort.
Jouni Malinen [Thu, 14 Jan 2016 17:28:33 +0000 (19:28 +0200)]
OpenSSL: Update session_secret callback to match OpenSSL 1.1.0 API
The SSL_CIPHER **cipher argument was marked const in OpenSSL 1.1.0
pre-release 2 similarly to how this is in BoringSSL. Fix build with that
in preparation for supporting OpenSSL 1.1.0.
Jouni Malinen [Wed, 13 Jan 2016 21:35:53 +0000 (23:35 +0200)]
EAP server: Simplify EAP method registration call
Free the allocated structure in error cases to remove need for each EAP
method to handle the error cases separately. Each registration function
can simply do "return eap_server_method_register(eap);" in the end of
the function.
Jouni Malinen [Wed, 13 Jan 2016 21:25:54 +0000 (23:25 +0200)]
EAP peer: Simplify EAP method registration call
Free the allocated structure in error cases to remove need for each EAP
method to handle the error cases separately. Each registration function
can simply do "return eap_peer_method_register(eap);" in the end of the
function.
Roy Marples [Sat, 9 Jan 2016 01:39:43 +0000 (01:39 +0000)]
bsd: Optimize socket use
Create global init to handle socket calls and route messages.
Register each interface inside the global driver so that
routing messages can find the interface based on rtm_ifindex.
Rafał Miłecki [Mon, 11 Jan 2016 18:31:50 +0000 (19:31 +0100)]
nl80211: Report disassociated STA / lost peer for the correct BSS
We shouldn't use drv->ctx as it always points to the first BSS. When
using FullMAC driver with multi-BSS support it resulted in incorrect
treating nl80211 events. I noticed with with brcmfmac and BCM43602.
Before my change I was getting "disassociated" on a wrong interface:
wlan0-1: STA 78:d6:f0:00:11:22 IEEE 802.11: associated
wlan0-1: STA 78:d6:f0:00:11:22 WPA: pairwise key handshake completed (RSN)
wlan0: STA 78:d6:f0:00:11:22 IEEE 802.11: disassociated
With this patch it works as expected:
wlan0-1: STA 78:d6:f0:00:11:22 IEEE 802.11: associated
wlan0-1: STA 78:d6:f0:00:11:22 WPA: pairwise key handshake completed (RSN)
wlan0-1: STA 78:d6:f0:00:11:22 IEEE 802.11: disassociated
This doesn't apply to hostapd dealing with SoftMAC drivers when handling
AP SME & MLME is done it hostapd not the firmware.
Jouni Malinen [Fri, 8 Jan 2016 22:47:30 +0000 (00:47 +0200)]
Drop OpenSSL 0.9.8 patches to add EAP-FAST support
The OpenSSL project will not support version 0.9.8 anymore. As there
won't be even security fixes for this branch, it is not really safe to
continue using 0.9.8 and we might as well drop the EAP-FAST patches for
it.
Jouni Malinen [Thu, 7 Jan 2016 19:44:49 +0000 (21:44 +0200)]
tests: EAP-SIM with external GSM auth and replacing SIM
These test cases verify that EAP-SIM with external GSM auth supports the
use case of replacing the SIM. The first test case does this incorrectly
by not clearing the pseudonym identity (anonymous_identity in the
network profile) while the second one clears that and shows successful
connection.
Jouni Malinen [Thu, 7 Jan 2016 15:51:40 +0000 (17:51 +0200)]
OpenSSL: Fix client certificate chain building after PKCS#12 use
If wpa_supplicant was first configured with PKCS #12 -based client
certificate chain and then used with another network profile that used a
different certificate chain from a X.509 certificate PEM file instead of
PKCS#12, the extra certificate chain was not reconstructed properly with
older versions of OpenSSL that 1.0.2. This could result in the
authentication failing due to the client certificate chain not being
complete or including incorrect certificates.
Fix this by clearing the extra certificate chain when setting up a new
TLS connection with OpenSSL 1.0.1. This allows OpenSSL to build the
chain using the default mechanism in case the new TLS exchange does not
use PKCS#12.
The following hwsim test case sequence was able to find the issue:
ap_wpa2_eap_tls_pkcs12 ap_wpa2_eap_tls_intermediate_ca_ocsp
Jouni Malinen [Thu, 7 Jan 2016 14:50:53 +0000 (16:50 +0200)]
ACS: Remove unreachable case from a debug print
n_chans can have only values 1, 2, or 4 in this function, so the -1 case
could never be reached. Remove the unreachable case to get rid of static
analyzer warnings.
Jouni Malinen [Thu, 7 Jan 2016 14:47:17 +0000 (16:47 +0200)]
Remove a pointer check that can never be true
chan is set to the result of pointer arithmetic (pointer to an entry in
an array) that can never be NULL. As such, there is no need to check for
it to be non-NULL before deference. Remove this check to avoid
complaints from static analyzers.
Jouni Malinen [Thu, 7 Jan 2016 14:39:10 +0000 (16:39 +0200)]
ACS: Be more consistent with iface->current_mode checks
Offloading of ACS to the driver changed the design a bit in a way that
iface->current_mode could actually be NULL when the offloaded ACS
mechanism supports band selection in addition to channel selection. This
resulted in a combination that is too complex for static analyzers to
notice. While acs_init() can be called with iface->current_mode == NULL
that is only in the case where WPA_DRIVER_FLAGS_ACS_OFFLOAD is in use.
In other words, the actual ACS functions like acs_cleanup() that would
dereference iface->current_mode are not used in such a case.
Get rid of static analyzer warnings by explicitly checking
iface->current_mode in acs_init() for the case where ACS offloading is
not used.
Jouni Malinen [Thu, 7 Jan 2016 11:41:15 +0000 (13:41 +0200)]
P2P: Print find_start in debug log when ignoring old scan results
This makes it easier to debug issues with old scan results being ignored
during P2P_FIND. A single rx_time would have been fine with
os_gettime(), but with os_get_reltime(), both rx_time and find_start
values are needed.
Jouni Malinen [Thu, 7 Jan 2016 11:30:59 +0000 (13:30 +0200)]
l2_packet: Extend bridge workaround RX processing to cover two frames
There was a race condition in how the l2_packet sockets got read that
could result in the same socket (e.g., non-bridge) to process both the
EAP-Success and the immediately following EAPOL-Key msg 1/4 instead of
each frame going in alternative order between the bridge and non-bridge
sockets. This could be hit, e.g., if the wpa_supplicant process did not
have enough CPU to process all the incoming frames without them getting
buffered and both sockets reporting frames simultaneously.
This resulted in the duplicated EAP-Success frame getting delivered
twice for processing and likely also the EAPOL-Key msg 1/4 getting
processed twice. While the latter does not do much harm, the former did
clear the EAP authentication state and could result in issues.
Fix this by extended the l2_packet Linux packet socket workaround for
bridge to check for duplicates against the last two received frames
instead of just the last one.
Jouni Malinen [Wed, 6 Jan 2016 22:27:50 +0000 (00:27 +0200)]
tests: Make eap_check_auth() error on missing selectedMethod clearer
It was possible to hit an error case in ap_wpa2_eap_in_bridge where the
selectedMethod STATUS field was not available. This resulted in not very
helpful "'selectedMethod'" message in the test log file. Make this
clearer by dumping all received STATUS fields and a clearer exception
message indicating that selectedMethod was missing.
It was possible for the packet socket on the bridge interface to receive
own transmitted frames between the bridge and non-bridge sockets
receiving the same incoming frame from a foreign host. This resulted in
the hash checksum validation step failing to notice a duplicate RX due
to the own frame updating the store hash value.
The own frame did get dropping in RX EAPOL processing, but that was too
late to address the issue with duplicate RX. Fix this by dropping own
frames already in l2_packet layer before checking and updating the last
RX hash value.
Jouni Malinen [Wed, 6 Jan 2016 19:12:08 +0000 (21:12 +0200)]
EAP-PAX: Check hmac_sha1_vector() return value
This function can fail at least in theory, so check its return value
before proceeding. This is mainly helping automated test case coverage
to reach some more error paths.
Jouni Malinen [Wed, 6 Jan 2016 16:49:15 +0000 (18:49 +0200)]
P2P: Clear groups first on FLUSH command
This is needed to get proper P2P group removal processing for some test
cases. discovery_group_client followed by nfc_p2p_client was able to hit
a case where the P2P group idle timeout survived to the next group
instance because of the FLUSH command not clearing the group and this
timeout properly.
Jouni Malinen [Wed, 6 Jan 2016 15:20:20 +0000 (17:20 +0200)]
mesh: Do not force another peering exchange on driver event
If the local driver indicated a peer candidate event when the peer had
already initiated peering exchange in open mesh case, we used to force a
new exchange to be started instead of allowing the previously started
exchange to complete. This is not desirable, so make this initiation of
the new exchange conditional on there not being an already started (or
successfully completed) exchange.
Jouni Malinen [Wed, 6 Jan 2016 15:17:13 +0000 (17:17 +0200)]
mesh: Do not clear link state on driver event if exchange was started
If the local driver event for a new peer candidate arrived only after
the peer had already initiated the peering exchange, we used to clear
the link state. This resulted in the already completed (or in progress)
exchange getting abandoned and a new exchange initiated. This is not
desirable since the already started (or even completed) exchange can be
used. Clear the link state only when adding the new STA entry for the
first time, i.e., use the same !sta->my_lid condition in handling the
driver event similarly to how the peer initiated cases were already
handled.
Jouni Malinen [Wed, 6 Jan 2016 11:13:13 +0000 (13:13 +0200)]
mesh: Connection and group started/removed events into debug log
The messages were sent out with wpa_msg_ctrl() so they were not visible
in the debug log. However, these would be quite helpful strings to
search for in the debug log, so change these messages to use wpa_msg().
Jouni Malinen [Tue, 5 Jan 2016 21:50:50 +0000 (23:50 +0200)]
EAP-PEAP peer: Cryptobinding in fast-reconnect case with inner EAP
This was reported to fail with Windows 2012r2 with "Invalid Compound_MAC
in cryptobinding TLV". It turns out that the server decided to go
through inner EAP method (EAP-MSCHAPv2 in the reported case) even when
using PEAP fast-reconnect. This seems to be against the [MS-PEAP]
specification which claims that inner EAP method is not used in such a
case. This resulted in a different CMK being derived by the server (used
the version that used ISK) and wpa_supplicant (used the version where
IPMK|CMK = TK without ISK when using fast-reconnect).
Fix this interop issue by making wpa_supplicant to use the
fast-reconnect version of CMK derivation only when using TLS session
resumption and the server having not initiated inner EAP method before
going through the cryptobinding exchange.
Jouni Malinen [Mon, 4 Jan 2016 23:03:28 +0000 (01:03 +0200)]
P2P: Try SD Query with each non-ACK peer only once per search iteration
The previous behavior of bursting out all retry attempts of an SD Query
frame during a single search/listen iteration does not look very helpful
in the case where the peer does not ACK the query frame. Since the peer
was found in the search, but is not ACKing frames anymore, it is likely
that it left its listen state and we might as well do something more
useful to burst out a significant number of frames in hopes of seeing
the peer.
Modify the SD Query design during P2P Search to send out only a single
attempt (with likely multiple link-layer retries, if needed) per
search/listen iteration to each peer that has pending SD queries. Once
no more peers with pending queries remain, force another Listen and
Search phase to go through before continuing with the pending SD
queries.
Jouni Malinen [Mon, 4 Jan 2016 21:37:43 +0000 (23:37 +0200)]
Clear wpa_supplicant state to DISCONNECTED on FLUSH command
It was possible for the FLUSH command to trigger auto connect mechanism
to schedule a new scan in 100 ms. This is not desired since all the
network profiles will be removed immediately and the scan or an attempt
to reconnect would not be of any benefit here. Such a scan in 100 ms can
cause issues for cases where multiple test sequences are run back to
back, so prevent this by clearing wpa_supplicant state to DISCONNECTED
(which avoids scheduling of the 100 ms scan trigger on disconnection) if
the state was AUTHENTICATING or higher when the FLUSH command was
issued.
Jouni Malinen [Mon, 4 Jan 2016 21:25:26 +0000 (23:25 +0200)]
tests: Make P2P discovery on non-social channel cases more robust
The test cases discovery_ctrl_char_in_devname and discovery_group_client
tried to allow three P2P_FIND instances to be used before reporting an
error. However, this did not really work properly since the second and
third attempts would likely fail to start the initial special P2P_FIND
scan due to an already ongoing p2p_scan operation. Fix this by stopping
the previous P2P_FIND and waiting for the scan to complete if a retry is
needed.
Jouni Malinen [Mon, 4 Jan 2016 19:34:39 +0000 (21:34 +0200)]
HS 2.0: Add some documentation for OSEN and network block use
This adds notes on how wpa_supplicant can be configured for OSEN for a
link-layer protected online signup connection and how network profiles
can be set for a Hotspot 2.0 data connection when using external
Interworking network selection.
Jouni Malinen [Fri, 1 Jan 2016 15:14:46 +0000 (17:14 +0200)]
tests: wpa_supplicant AP mode - unexpected P2P IE in Association Request
This verifies that there is no NULL pointer dereference when the AP code
processes Probe Request and (Re)Association Request frames with a P2P IE
in case P2P support is explicitly disabled on the AP mode interface.
This is a regression test case for the fixes in the previous commit.
Jouni Malinen [Fri, 1 Jan 2016 15:12:43 +0000 (17:12 +0200)]
Fix wpa_supplicant AP mode P2P IE handling if P2P is disabled
If P2P support is included in wpa_supplicant build (CONFIG_P2P=y), but
P2P functionality is explicitly disabled (e.g., "P2P_SET disabled 1"),
couple of AP management frame processing steps did not check against
hapd->p2p_group being NULL and could end up dereferencing a NULL pointer
if a Probe Request frame or (Re)Association Request frame was received
with a P2P IE in it. Fix this by skipping these steps if hapd->p2p_group
is NULL.
Jouni Malinen [Fri, 1 Jan 2016 14:50:24 +0000 (16:50 +0200)]
Fix wpa_supplicant build with CONFIG_L2_PACKET=pcap
Commit e6dd8196e5daf39e4204ef8ecd26dd50fdca6040 ('Work around Linux
packet socket regression') forgot to add the l2_packet_init_bridge()
wrapper for l2_packet_pcap.c while updating all the other l2_packet
options. This resulted in wpa_supplicant build failing due to missing
l2_packet_init_bridge() function when using CONFIG_L2_PACKET=pcap in
wpa_supplicant/.config. Fix this by adding the wrapper function.
Jouni Malinen [Fri, 1 Jan 2016 11:03:25 +0000 (13:03 +0200)]
WPS: Testing mechanism to force auth/encr type flags
The new wps_force_{auth,encr}_types parameters can be used in test build
(CONFIG_WPS_TESTING) to force wpa_supplicant to use the specified value
in the Authentication/Encryption Type flags attribute. This can be used
to test AP behavior on various error cases for which there are
workarounds to cover deployed device behavior.
Jouni Malinen [Fri, 1 Jan 2016 10:49:27 +0000 (12:49 +0200)]
WPS: Add a workaround for WPA2PSK missing from Enrollee auth flags
Some deployed implementations seem to advertise incorrect information in
this attribute. A value of 0x1b (WPA2 + WPA + WPAPSK + OPEN, but no
WPA2PSK) has been reported to be used. Add WPA2PSK to the list to avoid
issues with building Credentials that do not use the strongest actually
supported authentication option (that device does support WPA2PSK even
when it does not claim it here).
Jouni Malinen [Fri, 1 Jan 2016 09:59:20 +0000 (11:59 +0200)]
WPS: Do not build Credential with unsupported encr combination on AP
It was possible for the Registrar code to generate a Credential with
auth type WPAPSK (i.e., WPA v1) with encr type AES if the Enrollee
claimed support for WPAPSK and not WPA2PSK while the AP was configured
in mixed mode WPAPSK+WPA2PSK regardless of how wpa_pairwise (vs.
rsn_pairwise) was set since encr type was selected from the union of
wpa_pairwise and rsn_pairwise. This could result in the Enrollee
receiving a Credential that it could then not use with the AP.
Fix this by masking the encryption types separately on AP based on the
wpa_pairwise/rsn_pairwise configuration. In the example case described
above, the Credential would get auth=WPAPSK encr=TKIP instead of
auth=WPAPSK encr=AES.
Jouni Malinen [Thu, 31 Dec 2015 20:40:14 +0000 (22:40 +0200)]
tests: Use full prefix of the P2P-GO-NEG-FAILURE
Couple of waits for this event used the "GO-NEG-FAILURE" string instead
of the full event prefix. While this worked in the tests due to a
substring matching, it is better to use the full event prefix here.
Jouni Malinen [Thu, 31 Dec 2015 20:38:55 +0000 (22:38 +0200)]
tests: Do not dump pending events in p2p_go_neg_init timeout=0 case
It was possible for the dump_monitor() call to drop a P2P-GO-NEG-FAILURE
event that was indicated quickly after the P2P_CONNECT command was
issued. This could result in grpform_reject test case failing to see the
expected event and fail the test due to "Rejection not reported".
Jouni Malinen [Thu, 31 Dec 2015 19:46:08 +0000 (21:46 +0200)]
HS 2.0: Postpone WNM-Notification sending by 100 ms
This makes it somewhat easier for the station to be able to receive and
process the encrypted WNM-Notification frames that the AP previously
sentt immediately after receiving EAPOL-Key msg 4/4. While the station
is supposed to have the TK configured for receive before sending out
EAPOL-Key msg 4/4, not many actual implementations do that. As such,
there is a race condition in being able to configure the key at the
station and the AP sending out the first encrypted frame after EAPOL-Key
4/4. The extra 100 ms time here makes it more likely for the station to
have managed to configure the key in time.
Jouni Malinen [Thu, 31 Dec 2015 18:10:30 +0000 (20:10 +0200)]
OpenSSL: Share a single openssl_tls_prf() implementation
Add SSL_SESSION_get_master_key() compatibility wrapper for older OpenSSL
versions to be able to use the new openssl_tls_prf() implementation for
OpenSSL 1.1.0 with all supported versions.
Jouni Malinen [Thu, 31 Dec 2015 16:15:09 +0000 (18:15 +0200)]
OpenSSL: Clean up function to fetch client/server random
SSL_get_client_random() and SSL_get_server_random() will be added in
OpenSSL 1.1.0. Provide compatibility wrappers for older versions to
simplify the tls_connection_get_random() implementation.
Jouni Malinen [Thu, 31 Dec 2015 16:05:28 +0000 (18:05 +0200)]
OpenSSL: Drop support for OpenSSL 1.0.0
The OpenSSL project will not support version 1.0.0 anymore. As there
won't be even security fixes for this branch, it is not really safe to
continue using 1.0.0 and we might as well drop support for it to allow
cleaning up the conditional source code blocks.
Jouni Malinen [Thu, 31 Dec 2015 15:55:03 +0000 (17:55 +0200)]
OpenSSL: Drop support for OpenSSL 0.9.8
The OpenSSL project will not support version 0.9.8 anymore. As there
won't be even security fixes for this branch, it is not really safe to
continue using 0.9.8 and we might as well drop support for it to allow
cleaning up the conditional source code blocks.
Jouni Malinen [Wed, 30 Dec 2015 22:53:20 +0000 (00:53 +0200)]
tests: Clear BSS table at the end of rsn_ie_proto_eap_sta
rsn_ie_proto_eap_sta followed by eap_ttls_mschapv2_session_resumption
showed a failure case where the special RSNE from rsn_ie_proto_eap_sta
ended up remaining in a wpa_supplicant BSS entry and the SELECT_NETWORK
command used the previous scan results without checking for changed AP
configuration. This resulted in test failure due to RSN IE being claimed
to be different in EAPOL-Key msg 3/4. This is not really a real world
issue, but try to avoid false failure reports by explicitly clearing the
BSS table at the end of rsn_ie_proto_eap_sta.
Jouni Malinen [Wed, 30 Dec 2015 22:37:45 +0000 (00:37 +0200)]
tests: P2P_LISTEN immediately followed by P2P_FIND
This verifies that the previous commit works correctly by forcing a
P2P_LISTEN command execution to be interupted by a P2P_FIND command
timed in a manner that forces it to show up before the kernel ROC has
started for the Listen.
Jouni Malinen [Wed, 30 Dec 2015 22:35:21 +0000 (00:35 +0200)]
P2P: Fix P2P_FIND while waiting for listen ROC to start in the driver
It was possible for the p2p->pending_listen_freq to be left indicating
that there is a pending ROC for a listen operation if a P2P_FIND command
was timed to arrive suitably between a previous Listen operation issuing
a ROC request and the kernel code starting that request. This could
result in the P2P state machine getting stuck unable to continue the
find ("P2P: p2p_listen command pending already").
Fix this by clearing p2p->pending_listen_freq when starting P2P_FIND
command execution.
Jouni Malinen [Wed, 30 Dec 2015 22:06:06 +0000 (00:06 +0200)]
tests: GO Negotiation stopped after TX start
This verifies that P2P_STOP_FIND stops a pending offchannel TX wait in
the kernel by checking that a listen operation can be started in less
than a second after stopping a pending Action frame TX. This verifies
that the optimization introduced in the previous commit works properly.
Jouni Malinen [Wed, 30 Dec 2015 22:03:21 +0000 (00:03 +0200)]
P2P: Stop offchannel TX wait on P2P_STOP_FIND/P2P_LISTEN
Previously it was possible for the pending Action frame TX to be
cleared, but the offchannel TX operation being left in wait state in the
kernel. This would delay start of the next operation (e.g., that listen
operation requested by P2P_LISTEN) until the wait time for the
previously pending Action frame had expired.
Optimize this by explicitly stopping any pending offchannel Action frame
TX when clearing the internal offchannel TX state in
wpas_p2p_clear_pending_action_tx().
Jouni Malinen [Wed, 30 Dec 2015 19:51:52 +0000 (21:51 +0200)]
Revert "tests: vm: Output everything on console"
This reverts commit be9fe3d8aff394ea6868f1a2347e8c12609b086e. While I
did manage to complete multiple test runs without failures, it looks
like this change increases full test run duration by about 30 seconds
when using seven VMs. The most visible reason for that seems to be in
"breaking" active scanning quite frequently with the Probe Response
frame coming out about 40 ms (or more) after the Probe Request frame
which is long enough for the station to already have left the channel.
Since this logging change is not critical, it is simplest to revert it
for now rather than make changes to huge number of test cases to allow
more scan attempts to be performed before timing out.
Jouni Malinen [Wed, 30 Dec 2015 18:00:01 +0000 (20:00 +0200)]
tests: Increase connection timeout for number of EAP test cases
The previously used 10 second timeout allowed only two scan attempts
(five seconds between scans) and it was possible to hit a failure every
now and then when running under heavy load and the Probe Response frame
got delayed by 40 ms or so twice in a row. Add more time for one more
scan attempt to reduce the likelihood of this happening.
Jouni Malinen [Wed, 30 Dec 2015 17:52:42 +0000 (19:52 +0200)]
tests: Make scan test cases more robust by allowing retries
These test caases depended on a single active scan round finding the AP.
It is possible for the Probe Response frame to get delayed sufficiently
to miss the response especially when testing under heavy load with
multiple parallel VMs. Allow couple of scan retries to avoid reporting
failures from these test cases.
Jouni Malinen [Wed, 30 Dec 2015 17:14:40 +0000 (19:14 +0200)]
tests: Make P2PS join-a-group cases more robust
Use the group SSID (if known) when requesting a join operation. This
makes some of the P2PS test cases more robust in cases where previously
executed tests have added older groups into the cached scan results with
the same MAC addresses and an incorrect BSS could have been picked
previously.
Jouni Malinen [Wed, 30 Dec 2015 17:12:38 +0000 (19:12 +0200)]
P2P: Add an option to specify group SSID in P2P_CONNECT join case
The new optional ssid=<hexdump> argument to P2P_CONNECT can be used to
make P2P Client operations during join-an-existing-group more robust by
filtering out scan results based on the SSID in addition to the P2P
Device/Interface Address. This can help if the same MAC address has been
used in multiple groups recently and the cached scan results may still
include an older BSS.
Jouni Malinen [Wed, 30 Dec 2015 17:08:51 +0000 (19:08 +0200)]
P2P: Provide group SSID, if specified, to P2P Client join step
At least one of the wpas_p2p_connect() callers (NFC join case) already
had access to the Group SSID. Pass that information through
wpas_p2p_connect() to wpas_p2p_join() so that the join operation can
filter out incorrect groups more easily.
Jouni Malinen [Wed, 30 Dec 2015 17:06:19 +0000 (19:06 +0200)]
P2P: Do not accept any GO BSS entry if SSID is specified for join
Accept only a BSS entry matching the SSID when trying to find the
operating channel of a GO during join operation for which the SSID was
already specified. Previously, it could have been possible to pick an
incorrect BSS entry if the new GO was not found in the latest scan and
there was an older cached scan entry for the same BSSID.
Jouni Malinen [Wed, 30 Dec 2015 17:04:43 +0000 (19:04 +0200)]
P2P: Use join SSID in the skip-PD cases
It was already possible to limit join operation to accept only a
specific SSID. However, this constraint was not used when starting a P2P
Client interface as a WPS Enrollee without going through a Provision
Discovery exchange.
Jouni Malinen [Wed, 30 Dec 2015 17:02:33 +0000 (19:02 +0200)]
P2P: Do not accept any BSS entry for join if SSID is already known
Use wpa_bss_get() with the specific Group SSID instead of
wpa_bss_get_bssid_latest() if the SSID is already known. This makes the
P2P join operations more robust in case the frequency of the group was
not yet known and the same P2P Interface Address may have been used in
multiple group instances with an older group entry still present in the
cached scan results.
Jouni Malinen [Wed, 30 Dec 2015 17:00:56 +0000 (19:00 +0200)]
P2P: Use group SSID, if known, for join operation even if no BSS entry
This allows the cases where a specific group SSID is known to filter out
groups on the P2P Client even if the specific BSS entry for the target
group is not yet available.
Jouni Malinen [Wed, 30 Dec 2015 16:59:03 +0000 (18:59 +0200)]
P2PS: Add group SSID, if known, to the P2PS-PROV-DONE event
The new optional group_ssid=<hexdump> argument in the P2PS-PROV-DONE
event can be used to help in identifying the exact group if there have
been multiple groups with the same P2P Interface Address in short period
of time.
Jouni Malinen [Wed, 30 Dec 2015 11:54:18 +0000 (13:54 +0200)]
tests: Dump control interface sockets during FST operations
This makes it less likely to hit issues with running out of control
interface TX queue when running multiple FST test in a row. Number of
the FST operation sequences seemed to leave quite a few event messages
pending in one of the attached control interface sockets for wlan5 which
could result in test failure if the buffer space ran out and some of the
wpa_supplicant events were not delivered.
Jouni Malinen [Wed, 30 Dec 2015 11:29:32 +0000 (13:29 +0200)]
tests: Clean up at the end of connect_cmd_disconnect_event
This test case used to leave the dynamically added wlan5 interface in a
state where it was still trying to reconnect to a network. This could
result in the following test cases being unable to clear the cfg80211
scan cache. Avoid this type of issues by explicitly stopping the
connection attempt and making sure that there are no scan results in the
cache at the end of connect_cmd_disconnect_event.
The following test case sequence triggered a failure due to the
remaining BSS table entry:
connect_cmd_disconnect_event connect_cmd_wep ap_hs20_random_mac_addr
Dedy Lansky [Tue, 29 Dec 2015 07:24:51 +0000 (09:24 +0200)]
FST: Fix handling of Rx FST Setup Request when session already exists
When we receive FST Setup Request when session already exists, the
following validations take place:
1. we drop the frame if needed according to MAC comparison
2. we drop the frame if the session is "not pending", i.e., if FST
Setup Response was already exchanged (sent or received).
There are two issues with the above:
1. MAC comparison is relevant only before the Setup Response exchange.
In other words, Setup Request should not be dropped due to MAC
comparison after Setup Response has been exchanged.
2. Receiving Setup Request after Setup Response exchange most likely
means that FST state machine is out of sync with the peer. Dropping
the Setup Request will not help solve this situation.
The fix is:
1. do MAC comparison only if session is "pending", i.e., Setup Response
was not yet exchanged.
2. In case Setup Response was already exchanged, reset our session and
handle the Setup Request as if it arrived when session doesn't exist.
Jouni Malinen [Mon, 28 Dec 2015 16:15:18 +0000 (18:15 +0200)]
TLS: Remove storing of never-read value
While this could in theory be claimed to be ready for something to be
added to read a field following the server_write_IV, it does not look
likely that such a use case would show up. As such, just remove the
unused incrementing of pos at the end of the function to get rid of a
useless static analyzer complaint.
Jouni Malinen [Mon, 28 Dec 2015 16:12:03 +0000 (18:12 +0200)]
Remove unnecessary cleanup assignment in SHA1Final()
This makes some static analyzers complain about stored value never being
read. While it is good to clear some other temporary variables, this
local variable i has no security private information (it has a fixed
value of 20 here) and trying to clear it to 0 does not add any value.
Remove that part of the "wipe variables" to avoid one useless static
analyzer complaint.
projects / thirdparty / hostap.git / log
