This test case is not really realistic and the second connection attempt
would fail if additional AES-based ciphers get provisioned. Work this
around by dropping to CCMP only if other ciphers are present.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
This is needed to avoid false errors with GCMP-256 and CCMP-256 to be
added in the implementation.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Introduction of the new base64 helper function changed the backtraces
for these OOM test cases and resulted in test failures. Update the test
scripts to work with the new implementation.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
The test step for concurrent HTTP connections seems to be failing quite
frequently when running in a virtual machine with run-tests.py (but not
that much with kvm and vm-run.sh). The failures are due to only 8 or 9
sockets getting a response from the HTTP server. This is sufficient for
testing purposes, to drop the pass criterium from 10 to 8 concurrent
connections. This avoids unnecessary test failures and also allows the
rest of the test case to be performed.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
It was possible for ap_wps_per_station_psk_failure to leave behind scan
entries with active PBC mode if cfg80211 BSS table. This could result in
a following test case failing due PBC overlap. Fix this by clearing the
cfg80211 BSS table explicitly.
This was found with the following test case sequence:
ap_wps_per_station_psk_failure autogo_pbc
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Commit 2d6a526ac3 ('tests: Make
ap_wps_er_http_proto more robust') tried to work around the timeouts
here, but that was not really the best approach since the one second
timeout that was used here for connect() ended up being very close to
the limit even before the kernel change. The longer connect() time is
caused by a sequence where the listen() backlog ignores the connection
instead of accept() followed by close() within the wpa_supplicant ER
HTTP connection handling. The time to retransmit the SYN changed a bit
in the kernel from 1.0 sec to about 1.03 sec. This was enough to push
that over the one second timeout.
Fix this by using a sufficiently long timeout (10 sec) to allow SYN
retransmission to occur to recover from the listen() backlog case.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
It looks like connect() for a TCP socket can time out at least with a
recent kernel. Handle that case more gracefully by ignoring that socket
while allowing the test to continue.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
After successfully passing the 525 tests on a remote setup mark the
tests as remote compatible.
Signed-off-by: Jonathan Afek <jonathanx.afek@intel.com>
Some environments define default system wide HTTP proxy. Using default
system configuration may result in a failure to open some HTTP URLs. Fix
this by ensuring that no proxies are used.
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Pass apdev param to hostapd.add_bss(). Kill hardcoded phy param and get
phy base on apdev. These are needed to support operation with a remote
test host.
Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com>
Pass the full apdev to the add_ap() function instead of just ifname.
This allows us to handle also remote hosts while we can check
apdev['hostname'], apdev['port'].
This step (4) converts the cases that call hostapd.add_ap() from a
helper function that got apdev[i] as an argument.
Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com>
Pass the full apdev to the add_ap() function instead of just ifname.
This allows us to handle also remote hosts while we can check
apdev['hostname'], apdev['port'].
This step (2) converts the cases that use the add_ssdp_ap() helper
function.
Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com>
Pass the full apdev to the add_ap() function instead of just ifname.
This allows us to handle also remote hosts while we can check
apdev['hostname'], apdev['port'].
This step (1) converts the cases where apdev[#]['ifname'] was used as
the argument to hostapd.add_ap().
Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com>
This is needed to avoid reporting failures after a change to remove the
fallback path in PIN generation.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
It was possible to hit a race condition between WPS_CANCEL and
immediately following WPS_PIN command. Wait for a disconnection event to
avoid that. This was seen with the following test case sequence:
ap_wpa2_psk_supp_proto_wrong_group_key_len ap_wps_probe_req_ie_oom
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Only one of the ERs was stopped at the end of the test case and this
could result in the following test case failing, e.g., when executing
this test case sequence: ap_wps_er_multi_add_enrollee ap_wps_upnp.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
This makes it more likely for the two ERs to go through WPS UPnP
exchange in parallel. This was already happening every now and then and
resulted in failures. However, now that there is support for multiple
concurrent exchanges, it is useful to have this test case hit that
possibility more frequently.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
gcc 4.8 vs 5.2 seem to compile eloop_register_sock() differently. With
5.2, that function name does not show up in the backtrace since
eloop_sock_table_add_sock() is used without a separate function call.
This broke the memory allocation failure checking in this test case. Fix
this by matching against the eloop_sock_table_add_sock() function which
shows up in the backtrace for both gcc versions.
Signed-off-by: Jouni Malinen <j@w1.fi>
It was possible for the WPS PBC state to get cached through to the
following test cases and that would trigger false failures. Fix this by
explicitly clearing the scan cache at the end of ap_wps_per_station_psk.
This issue was triggered with the following test case sequence:
ap_wps_per_station_psk autogo_pbc
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Number of test cases did not read all control interface socket events
from the dynamically added wlan5 interface. This could result in hitting
maximum socket TX queue length and failures in the following test cases.
Signed-off-by: Jouni Malinen <j@w1.fi>
This is a regression test case for the issue fixed by the previous
commit (hapd->num_probereq_cb not getting cleared on deinit).
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
