Commit 4104267e81 ('Fix memory leak on NFC
DH generation error path') modified the generic (non-OpenSSL)
implementation of dh5_init() to free the previously assigned public key,
if any. However, that commit did not modify the OpenSSL specific version
of this function. Add the same change there to maintain consistent
behavior between these two implementations of the same function.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Commit 4104267e81 ('Fix memory leak on NFC
DH generation error path') modified dh5_init() behavior in the
non-OpenSSL implementation to free the public key (if any was previously
set). However, this did not update one of the callers to make sure the
publ argument in the call is initialized. This could result in trying to
free invalid pointer and segmentation fault when hostapd or
wpa_supplicant was built against some other crypto library than OpenSSL.
Signed-off-by: Rujun Wang <chinawrj@gmail.com>
BoringSSL added OpenSSL 1.1.0's SSL_get_client_random() and friends in
working towards opaquifying the SSL struct. But it, for the moment,
still looks more like 1.0.2 than 1.1.0 and advertises
OPENSSL_VERSION_NUMBER as such. This means that there is no need to
define those in BoringSSL and defining them causes conflicts. (C does
not like having static and non-static functions with the same name.)
As requested, this is conditioned on defined(BORINGSSL_API_VERSION) so
wpa_supplicant may continue to support older BoringSSLs for a time.
(BoringSSL revisions without the accessors predate BoringSSL maintaining
a BORINGSSL_API_VERSION.)
Also add a missing opensslv.h include. tls_openssl.c is sensitive to
OPENSSL_VERSION_NUMBER, so it should include the header directly rather
than rely on another header to do so.
Signed-off-by: David Benjamin <davidben@google.com>
This verifies that pmf=2 is ignored for a non-RSN network while a
network profile specific ieee80211w=2 is enforced.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Since ieee80211w=2 is an explicit configuration to wpa_supplicant, the
connection attempt for such non-PMF (non-RSN) capable networks should be
skipped.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
PMF is available only with RSN and pmf=2 could have prevented open
network connections. Change the global wpa_supplicant pmf parameter to
be interpreted as applying only to RSN cases to allow it to be used with
open networks.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Many of the global configuration parameters are written as strings
without filtering and if there is an embedded newline character in the
value, unexpected configuration file data might be written.
This fixes an issue where wpa_supplicant could have updated the
configuration file global parameter with arbitrary data from the control
interface or D-Bus interface. While those interfaces are supposed to be
accessible only for trusted users/applications, it may be possible that
an untrusted user has access to a management software component that
does not validate the value of a parameter before passing it to
wpa_supplicant.
This could allow such an untrusted user to inject almost arbitrary data
into the configuration file. Such configuration file could result in
wpa_supplicant trying to load a library (e.g., opensc_engine_path,
pkcs11_engine_path, pkcs11_module_path, load_dynamic_eap) from user
controlled location when starting again. This would allow code from that
library to be executed under the wpa_supplicant process privileges.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Most of the cred block parameters are written as strings without
filtering and if there is an embedded newline character in the value,
unexpected configuration file data might be written.
This fixes an issue where wpa_supplicant could have updated the
configuration file cred parameter with arbitrary data from the control
interface or D-Bus interface. While those interfaces are supposed to be
accessible only for trusted users/applications, it may be possible that
an untrusted user has access to a management software component that
does not validate the credential value before passing it to
wpa_supplicant.
This could allow such an untrusted user to inject almost arbitrary data
into the configuration file. Such configuration file could result in
wpa_supplicant trying to load a library (e.g., opensc_engine_path,
pkcs11_engine_path, pkcs11_module_path, load_dynamic_eap) from user
controlled location when starting again. This would allow code from that
library to be executed under the wpa_supplicant process privileges.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Spurious newlines output while writing the config file can corrupt the
wpa_supplicant configuration. Avoid writing these for the network block
parameters. This is a generic filter that cover cases that may not have
been explicitly addressed with a more specific commit to avoid control
characters in the psk parameter.
Signed-off-by: Paul Stewart <pstew@google.com>
WPA/WPA2-Personal passphrase is not allowed to include control
characters. Reject a passphrase configuration attempt if that passphrase
includes an invalid passphrase.
This fixes an issue where wpa_supplicant could have updated the
configuration file psk parameter with arbitrary data from the control
interface or D-Bus interface. While those interfaces are supposed to be
accessible only for trusted users/applications, it may be possible that
an untrusted user has access to a management software component that
does not validate the passphrase value before passing it to
wpa_supplicant.
This could allow such an untrusted user to inject up to 63 characters of
almost arbitrary data into the configuration file. Such configuration
file could result in wpa_supplicant trying to load a library (e.g.,
opensc_engine_path, pkcs11_engine_path, pkcs11_module_path,
load_dynamic_eap) from user controlled location when starting again.
This would allow code from that library to be executed under the
wpa_supplicant process privileges.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
This verifies that a WPA2PSK passphrase with control characters gets
rejected in a WPS Credential and that control characters in SSID get
written as a hexdump.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
WPA/WPA2-Personal passphrase is not allowed to include control
characters. Reject a Credential received from a WPS Registrar both as
STA (Credential) and AP (AP Settings) if the credential is for WPAPSK or
WPA2PSK authentication type and includes an invalid passphrase.
This fixes an issue where hostapd or wpa_supplicant could have updated
the configuration file PSK/passphrase parameter with arbitrary data from
an external device (Registrar) that may not be fully trusted. Should
such data include a newline character, the resulting configuration file
could become invalid and fail to be parsed.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
One of supported code paths already allows this scenario. It is used if
driver doesn't report NL80211_ATTR_DEVICE_AP_SME and doesn't support
monitor interface. In such situation:
1) We don't quit if subscribing for WLAN_FC_STYPE_PROBE_REQ fails
2) We don't try subscribing for WLAN_FC_STYPE_ACTION
3) We fallback to AP SME mode after failing to create monitor interface
4) We don't quit if subscribing for WLAN_FC_STYPE_PROBE_REQ fails
Above scenario is used, e.g., with brcmfmac. As you can see - thanks to
events provided by cfg80211 - it's not really required to receive Probe
Request or action frames.
However, the previous implementation did not allow using hostapd with
drivers that:
1) Report NL80211_ATTR_DEVICE_AP_SME
2) Don't support subscribing for PROBE_REQ and/or ACTION frames
In case of using such a driver hostapd will cancel setup after failing
to subscribe for WLAN_FC_STYPE_ACTION. I noticed it after setting flag
WIPHY_FLAG_HAVE_AP_SME in brcmfmac driver for my experiments.
This patch allows working with such drivers with just a small warning
printed as debug message.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
This reduces duration that private keying material might remain in the
process memory by clearing wpabuf data used in WPS operations when there
is possibility of the buffer including keys or related material.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
When a dedicated P2P Device interface is used, its configuration should
be cloned to the group interface. Add a test that covers this both when
a separate group interface is used and not.
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
When the P2P Device interface is used and an existing interface is used
for P2P GO/Client, the P2P Device configuration was not cloned to the
configuration of the existing interface. Thus, configuration parameters
such as idle_group_time, etc., were not propagated to the P2P GO/Client
interface.
Handle this by copying all configuration parameters of the P2P device
interface to the reused interface, with the following exceptions:
1. Copy the NFC key data only if it was not set in the configuration
file.
2. The WPS string fields are set only if they were not previously set
in the configuration of the destination interface (based on the
assumption that these fields should be identical among all
interfaces).
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Use the p2pdev pointer instead of the parent pointer to comply with the
flows when a dedicated P2P Device interface is used and
p2p_no_group_iface == 1 (in which case the parent of the reused
interface isn't necessary the same as p2pdev).
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Use the global control interface to remove P2P network blocks, to
support cases when a dedicated P2P Device interface is used.
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@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>
In p2p_channel_vht80_autogo and p2p_channel_vht80p80_autogo, parse the
P2P-GROUP-STARTED event prior to calling the group_request() method, as
otherwise the group ifname is not set.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Support configurations that use a dedicated P2P Device interface by
using the global control interface and specifying the interface name for
the GET commands fetching the ip_addr_go parameter.
Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Modify the persistent_group_profile_add test to support configurations
that use a dedicated P2P Device interface by sending the ADD_NETWORK and
SET_NETWORK commands on the global control interface and specifying the
P2P Device interface name.
Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Add an attribute to class WpaSupplicant with the name of the
P2P Device interface. If a separate interface is not used for
P2P Device, this attribute will hold the name of the only used
interface (with functions also as the P2P Device management
interface).
This attribute will be used to direct P2P related commands to the
P2P Device interface, which is needed for configurations that use
a separate interface for the P2P Device.
Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Waiting for the P2P-DEVICE-FOUND event should be done on the global
control interface to support configurations that use a dedicated P2P
Device interface.
Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Support configurations that use a dedicated P2P Device interface by
sending the P2P_CONNECT command on the global control interface.
In addition, when a dedicated P2P Device interface is used, there is no
need to manually respond to the Provision Discovery Request since the
request is processed by the P2P Device interface and this interface was
not set for external RX management frames handling.
Signed-off-by: Avraham Stern <avraham.stern@intel.com>
This fixes commit 2f0f69a9ec ('tests: Use
p2ps_provision() and p2ps_connect_pd() in p2ps_connect_p2ps_method()').
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
1. Fail roaming to an AP which exceeded its number of allowed stations.
2. Fail roaming due to passphrase mismatch.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
It is better to use a list of command line arguments for the local
execution case and convert that to a space-separated string for the
remote case.
Signed-off-by: Jouni Malinen <j@w1.fi>
This makes the DFS test cases that use start_dfs_ap() more usable for
testing with remote hosts.
Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com>
