If a separate P2P group interface was used, P2P_CONNECT-auto fallback to
GO Negotiation could result in use of freed memory and segmentation
fault. This happened in cases where the peer GO was found in some old
scans, but not in the first scan triggered by the P2P_CONNECT-auto
command ("P2P: Peer was found running GO in older scan -> try to join
the group" shows up in the debug log). In addition, the GO would still
need to reply to PD Request to allow this code path to be triggered.
When five scans for the GO were completed in this sequence, the P2P
group interface was removed as part of falling back to GO Negotiation.
However, that ended up dereferencing the freed wpa_s instance at the end
of scan event processing. Fix this by reordering code a bit and breaking
out from EVENT_SCAN_RESULTS processing if the interface could have been
removed.
Signed-off-by: Jouni Malinen <j@w1.fi>
It is possible for a low powered CPU to take excessively long time to
delete 1000 network blocks when running under valgrind. This would have
resulted in the test case failing and the following reset operation
timing out which would then stop the test sequence completely.
Signed-off-by: Jouni Malinen <j@w1.fi>
When SAE authentication fails, wpa_supplicant retries four times. If all
the retries result in failure, SAE state machine enters BLOCKED state.
Once it enters this state, wpa_supplicant doesn't retry connection. This
commit allow connection retries even if the state machine entered
BLOCKED state.
There could be an opinion "Is this patch needed? User could know the SAE
state machine is in the BLOCKED mode by MESH-SAE-AUTH-BLOCKED event.
Then user can retry connection. By user action, SAE state machine can
change the state from BLOCKED to another.". Yes, this is a true at the
joining mesh STA. However, a STA that is already a member of existing
mesh BSS should not retry connection because if the joining mesh STA
used wrong password, all the existing STA should do something from UI to
retry connection.
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
Send MESH-SAE-AUTH-BLOCKED event if SAE authentication is blocked. The
BLOCK state will finish when a new peer notification event is sent for
the same MAC address.
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
SAE authentication fails likely with wrong password. This commit adds a
notification of the failure to the upper application (UI) so that the
application can notify suspection of a wrong password to the user. The
control interface monitor even for this is "MESH-SAE-AUTH-FAILURE
addr=<peer>".
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
Commit 5a2a6de6a5 ('mesh: Make inactivity
timer configurable') has a problem on 32 bit systems. Setting
NL80211_MESHCONF_PLINK_TIMEOUT to 0xffffffff causes expiration of STA in
a minute by NL80211_CMD_DEL_STATION event. this is the kernel rule for
STA expiration:
(current jiffies) > (frame Rx jiffies + NL80211_MESHCONF_PLINK_TIMEOUT * 250)
On a 32 bit system, the right side could overflow and be unexpected
small value if NL80211_MESHCONF_PLINK_TIMEOUT is sufficiently large. STA
expiration occurs by this reason.
This patch solves the problem by disabling the STA expiration
functionality in mac80211. However, old kernel does not support
disabling it. If so, this patch sets mac80211 inactivity timer 60
seconds into future from the wpa_supplicant inactivity timer.
And I mis-understood that mesh_max_inactivity=0 disables inactivity
timer in wpa_supplicant. This commit fixes it also.
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
Commit 07cb45ccb2 ('mesh: Add no_auto_peer
config option') added a new struct wpa_ssid argument and added an
unnecessary parsing and setting of the value in
wpa_supplicant_ctrl_iface_update_network(). This is not needed since
wpa_config_set() takes care of parsing the parameters.
Signed-off-by: Jouni Malinen <j@w1.fi>
It is possible for a network profile to change its type from P2P
persistent group to a normal network and back. The D-Bus interface uses
different types of objects for those, so the object needs to
re-registered in case of type change. This fixes issues in leaving
behind an incorrect type of object and leaking memory when freeing such
a network block that has had its disabled parameter changed.
Signed-off-by: Jouni Malinen <j@w1.fi>
If HT40 co-ex scan fails due to the driver rejecting scan triggers
multiple times, it was possible for the ap_ht40_scan_retry() timeout
being left behind and it getting run after hapd->drv_priv has been
cleared. This would result in NULL pointer dereference in
driver_nl80211_scan.c. Fix this by canceling the timeout when disabling
the interface.
Signed-off-by: Jouni Malinen <j@w1.fi>
Commit e6dd8196e5 ('Work around Linux
packet socket regression') added a mechanism to close the workaround
bridge socket in l2_packet_receive(). However, it did not take into
account the possibility of the l2->rx_callback() closing the l2_packet
socket altogether. This could result in use of freed memory when usin
RSN pre-authentication. Fix this by reordering the calls to clear the
workaround socket before calling the rx_callback.
Signed-off-by: Jouni Malinen <j@w1.fi>
1. Add get_group_ifname() to wpasupplicant.py
2. Use the function to get the interface name for the bridge.
Signed-off-by: David Spinadel <david.spinadel@intel.com>
run-tests.py is running as root, so sudo does not need to be used
anymore from within each test case.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
There is no need to use sudo and external rm to remove files now that
run-tests.py is required to run as root.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
The new wpa_supplicant configuration writing design (rename instead of
write to original file) did not fail with the symlink-to-self case, so
replace this with the config file being replaced with a directory. In
addition, get rid of unnecessary use of subprocess since run-tests.py is
running as root nowadays.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
There is a chance that wpa_supplicant may get killed during
the time it is writing config file. If this happens, user
information like SSIDs and passwords can be lost forever.
This change works around that by writing config to a
temporary file and then renaming the file to the correct name.
Signed-off-by: Vinit Deshpande <vinitd@google.com>
We should not enable HT if WEP or TKIP is configured.
Without the patch and WEP configuration we will get message:
Association request to the driver failed
Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com>
Previously, a channel with even a single scan/survey result missing
information was skipped in ACS. This may not be desirable in cases when
multiple scan iterations are used (which is the case by default in
hostapd). Instead, use all channels that provided at least one complete
set of results. Calculate the average interference factor as an average
of the iterations that did provide complete values.
This seems to help with some cases, e.g., when ath9k may not be able to
report the noise floor for all channels from the first scan iteration
immediately after the driver has been loaded, but then returns it for
all other scan iterations.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
The new acs_chan_bias configuration parameter is a space-separated list
of <channel>:<bias> pairs. It can be used to increase (or decrease) the
likelihood of a specific channel to be selected by the ACS algorithm.
The total interference factor for each channel gets multiplied by the
specified bias value before finding the channel with the lowest value.
In other words, values between 0.0 and 1.0 can be used to make a channel
more likely to be picked while values larger than 1.0 make the specified
channel less likely to be picked. This can be used, e.g., to prefer the
commonly used 2.4 GHz band channels 1, 6, and 11 (which is the default
behavior on 2.4 GHz band if no acs_chan_bias parameter is specified).
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
The interference factors for adjacent 2.4 GHz channels were summed
together without doing any kind of weighted average on them. This
resulted in the channels at the band edges getting undue preference due
to only including interference factors from three channels vs. five for
the channels in the middle of the band.
While it is somewhat unclear whether the design here was supposed to
count overlapping channels together in this way or whether that is
already covered in channel survey results, it is clear that this summing
of three to five values together and then comparing the sum rather than
average of some kind gives too much preference to the channels at the
edges of the band by assuming that there is no interference whatsoever
outside the band.
Use weighted average of the interference factors rather than a sum from
different number of values. For now, the adjacent 2.4 GHz channels get
weight of 0.85 (1.0 for the main channel itself) and the neighboring
channels to those adjacent ones get 0.55 weight. Band-edge channels are
handled in a way that takes average over the channels that were actually
considered instead of assuming zero interference from neighboring bands.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
1. Modify discovery_stop to use global control interface when calling
P2P_FLUSH.
2. Modify p2p_listen_and_offchannel_tx to use the global control
interface when waiting for P2P PD event.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Some of the tests used p2p_dev_addr() that can be different from
own_addr() if a dedicated P2P Device interface is used.
Signed-off-by: Ben Rosenfeld <ben.rosenfeld@intel.com>
Some of the tests used p2p_dev_addr() that can be different from
own_addr() if a dedicated P2P Device interface is used.
Signed-off-by: Ben Rosenfeld <ben.rosenfeld@intel.com>
The test used p2p_dev_addr() that can be different from own_addr()
if a dedicated P2P Device interface is used.
Signed-off-by: Ben Rosenfeld <ben.rosenfeld@intel.com>
The test used p2p_dev_addr() that can be different from own_addr()
if a dedicated P2P Device interface is used.
Signed-off-by: Ben Rosenfeld <ben.rosenfeld@intel.com>
Fix the destination address that is sent in the WNM-Notification to be
the BSS address opposed to the P2P address.
Signed-off-by: Matti Gottlieb <matti.gottlieb@intel.com>
This is needed for cases that the group interface differs from the main
interface, i.e., when a dedicated P2P Device interface is used.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
When the 'SET' command is used to configure parameters related to P2P
operation use the global control interface and not the per interface one
as otherwise the setting will only have effect on the interface and will
work if a dedicated P2P_DEVICE is used.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Once the connection is established need to call group_form_result() on
the invited device, as otherwise the group interface name is not updated
and the connectivity test is done with the main interface instead of the
group interface.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
These are far from perfect since timing is quite difficult to match for
the case that behaved incorrectly. Anyway, it looks loke
p2p_service_discovery_peer_not_listening was able to hit the error now
and then, so this should be sufficient as a regression test case.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
If p2p_find_timeout triggers after starting SD but before getting TX
status for send action, unwanted search could get triggered again when
TX status arrives though p2p_find_timeout moved the state to P2P_IDLE by
then. p2p_continue_find() would then move the state to P2P_SEARCH again.
Do not trigger the find operation from this context if state is
P2P_IDLE to avoid this.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
This is going to be required for OpenSSL 1.1.0 which makes the SSL
structure opaque. Older versions starting from OpenSSL 1.0.1 include
this function, so start using it now based on OPENSSL_VERSION_NUMBER.
Signed-off-by: Jouni Malinen <j@w1.fi>
