To avoid a problem where the beacon socket occasionally
blocks, mark any sockets on the eloop as non-blocking.
The previous patch reordered the code to never send a
command after a socket was put on the eloop, but now also
invalidate the nl handle pointer while it's on there.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
Abstract the handling of sockets on the eloop to avoid
destroying sockets still on the eloop and also to allow
the next patch to mark the socket non-blocking.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
The IBSS code registers the bss nl_mgmt socket for auth
frames when the join event happens, but that is too late
as then the socket is already on the eloop, which could
cause problems when other events are received at the
same time as the registration is done.
Move the auth frame registration to the initial setup
before the socket is put onto the eloop.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
Some driver wrappers may implement this by writing eight octets even
though IPN is only six octets. Use a separate WPA_KEY_RSC_LEN (8) octet
buffer in the call to make sure there is enough buffer room available
for the full returned value and then copy it to IPN field.
The previous implementation used the following igtk field as the extra
buffer and then initialized that field afterwards, so this change does
not fix any real issue in behavior, but it is cleaner to use an explicit
buffer of the maximum length for get_seqnum().
Signed-hostap: Jouni Malinen <j@w1.fi>
wpa_s->conf cannot be NULL because wpa_supplicant_init_iface() would not
allow wpa_supplicant_add_iface() to return wpa_s instance in such state.
Signed-hostap: Jouni Malinen <j@w1.fi>
Even though the length of this buffer is based only on locally
configured information, it is cleaner to include explicit buffer room
validation steps when adding the attributes into the buffer.
Signed-hostap: Jouni Malinen <j@w1.fi>
linux_br_get() was forcing null termination on the buffer, but did not
check whether the string could have been truncated. Make this more
strict by rejecting any truncation case.
Signed-hostap: Jouni Malinen <j@w1.fi>
The allocation of new_bss and its use was separated by a lot of code in
this function. This can be cleaned up by moving the allocation next to
the use, so that this all can be within a single #ifdef HOSTAPD block.
The i802_check_bridge() call was outside type == WPA_IF_AP_BSS case, but
in practice, it is only used for WPA_IF_AP_BSS (and if used for
something else, this would have resulted in NULL pointer dereference
anyway).
Signed-hostap: Jouni Malinen <j@w1.fi>
If SSL_CTX_new() fails in tls_init(), the per-SSL app-data allocation
could have been leaked when multiple TLS instances are allocated.
Signed-hostap: Jouni Malinen <j@w1.fi>
The VHT_CHANWIDTH_160MHZ case fell through to the default case and
printed out a debug message that was not supposed to be shown here.
Signed-hostap: Jouni Malinen <j@w1.fi>
os_strlpcy() should be used instead of os_strncpy() to guarantee null
termination. Since there are no remaining strncpy uses, remove
os_strncpy() definition.
Signed-hostap: Jouni Malinen <j@w1.fi>
Commit 2e5ba4b6d1 moved this to a function
and updated one of the os_snprintf() calls to use the len parameter, but
forgot the other one.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
This was already included in all the other calls to eap_proxy, but
somehow the get_imsi call had been forgotten.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
Trying to access the SIM card details without checking if the eap_proxy
layer has been initialized can results in a crash. Address this by
sending the request for the IMSI through eapol_supp_sm.c which can
verify that eap_proxy has been initialized.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
It is possible for there to be two pending off-channel TX frames, e.g.,
when two devices initiate GO Negotiation at more or less the same time.
This could result in the TX status report for the first frame clearing
wpa_s->pending_action_tx that included the newer frame that has not yet
been transmitted (i.e., is waiting to be sent out). Avoid losing that
frame by confirming that the TX status payload matches the pending frame
before clearing the pending frame and reporting the TX status callback.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
DISCONNECT followed by ENABLE_NETWORK ended up starting a scan for a new
connection due to wpa_supplicant_enable_one_network() setting
wpa_s->reassociate = 1. This was done regardless of wpa_s->disconnected
being 1 which should imply that wpa_supplicant should not try to connect
before asked explicitly with REASSOCIATE or RECONNECT.
Fix this by making ENABLE_NETWORK setting of reassociate = 1 and
starting of scans for connection conditional on wpa_s->disconnected ==
0. This will make ENABLE_NETWORK trigger a connection only if
wpa_supplicant is already in a state where it would try to connect if
there are any enabled networks.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
Previously ACS required valid survey data on all available channels.
This can however not be guaranteed. Instead of just failing, fall back
to the subset of channels that have valid ACS data.
Signed-hostap: Helmut Schaa <helmut.schaa@googlemail.com>
Otherwise hostapd might hang doing nothing anymore. Propagate ACS
errors so we can fail gracefully.
Signed-hostap: Helmut Schaa <helmut.schaa@googlemail.com>
If ACS fails we still need to call hostapd_setup_interface_complete.
Otherwise hostapd will just hang doing nothing anymore. However, pass
an error to hostapd_setup_interface_complete to allow a graceful fail.
Signed-hostap: Helmut Schaa <helmut.schaa@googlemail.com>
This makes it easier to go through the P2P channel list operations in
the debug log without having to parse through the hexdump manually.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
To disallow roaming when a scan request's results are read, callers
of the D-Bus Scan() method may add a new "AllowRoam" boolean key
to the scan options dictionary and set that key's value to FALSE.
Signed-hostap: Dan Williams <dcbw@redhat.com>
If a scan is currently running and the scan interval is changed, a
second scan will be started before the current has finished. This will
in turn, if no networks are configured, cause wpa_s->state to be
forced to WPA_INACTIVE before the first scan has finished.
Signed-hostap: Pontus Fuchs <pontus.fuchs@gmail.com>
Incorrect nla_get variants were used to get event type and frequency.
Kernel passes both as u32. This caused issues on tinynl/big-endian hosts
- CAC finished was treated as radar detection and frequency was 0.
Signed-hostap: Michal Kazior <michal.kazior@tieto.com>
It is possible for a vif netdev to be removed by something else than
hostapd and if that happens for a virtual AP interface, if_remove()
handler should still free the local data structure to avoid memory leaks
if something external removes a netdev.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
Previously, the send_mlme->send_frame->send_frame_cmd path that could be
used when a GO sends an offchannel Action frame ended up not updating
drv->send_action_cookie. This can result in an issue with not being able
to cancel wait for the response, e.g., in invitation-to-running-group
case.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
test_p2p_go_invite_auth is similar to test_p2p_go_invite with the main
difference being in the peer device pre-authorizing the invitation
instead of processing invitation at upper layers after having received
it.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
Need to use the pointer to the current ongoing query instead of matching
from the pending list based on the destination address so that we get
the correct query instance when processing the TX status report.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
These operations can have conflicting offchannel requirements, so wait
with a new scan trigger until a pending GAS query has been completed.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
Offchannel operations needed for a GAS query can conflict with ongoing
scan/connection progress, so delay GAS queries if such an operation is
in progress on the current interface or any virtual interface sharing
the same radio.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
It would be possible to issue another GAS query when a previous one is
still in progress and this could result in conflicting offchannel
operations. Prevent that by delaying GAS query initiation until the
previous operation has been completed.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
This makes it less likely to forget WPS single-channel scan optimization
in effect after having completed the WPS operation or in case WPS
operating gets cancelled.
Signed-hostap: Jouni Malinen <j@w1.fi>
There is no point trying to go through a test case if the NOTE command
to write TEST-START entry does not succeed. This avoids some excessive
waits on buildbot trying to forcefully kill the programs on its timeout
if wpa_supplicant gets stuck waiting for something (like the current
issue with libnl events and commands having a chance of hitting a
blocking wait on netlink messages).
Signed-hostap: Jouni Malinen <j@w1.fi>
When a GO or P2P Client invites a peer device to join an already
operating group, the Operating Channel in Invitation Request needs to be
forced to the current operating channel of the group.
Signed-hostap: Jouni Malinen <j@w1.fi>
This fixes issues where a GO used offchannel-TX operation to send an
Invitation Request frame. Wait for the offchannel TX operation needs to
be stopped as soon as the Invitation Response frame has been received.
This addresses some issues where Probe Response frame from the GO
through the monitor interface may end up going out on a wrong channel
(the channel of this offchannel TX operation for invitation).
Signed-hostap: Jouni Malinen <j@w1.fi>
wpa_printf() does not need '\n' so remove the extra newline. In
addition, drop the priority of this message from MSG_INFO to MSG_DEBUG
since this is in no way exceptional operation.
Signed-hostap: Jouni Malinen <j@w1.fi>
Join-a-group needs to force the current operating channel of the target
group as the frequency to use for the PD exchange. When the channel was
selected based on a BSS entry for the GO, this worked only for the first
PD Request frame while the retries reverted to a potentially different
channel based on a P2P peer entry. Fix this by maintaining the forced
channel through the PD retry sequence.
Signed-hostap: Jouni Malinen <j@w1.fi>
This adds more coverage to testing by forcing the GO to be found with an
older entry in the BSS table and with that entry having a different
operating channel. Such a case has found issues with PD retries and
scanning if incorrect frequency and SSID is selected for the group.
Instead of relying on the old BSS entry with different operating channel
to happen based on a specific test case sequence, force this to happen
for this test case every time.
Signed-hostap: Jouni Malinen <j@w1.fi>
This makes it easier to debug issues related to selecting GO information
from the latest updated BSS table entry.
Signed-hostap: Jouni Malinen <j@w1.fi>
P2P Invitation Response frame is required to include the Channel List
attribute only in Status=Success case. Skip the debug message claiming
that a mandatory attribute was not included in non-Success case.
Signed-hostap: Jouni Malinen <j@w1.fi>
This verifies network selection using 3GPP Network Information and
connection using EAP-SIM with GSM authentication handled by an external
program.
Signed-hostap: Jouni Malinen <j@w1.fi>
Use similar mechanism to CONFIG_PCSC=y case to set the IMSI and MNC
length for eap_proxy. This allows automatic 3GPP realm comparison
against the domain list.
Signed-hostap: Jouni Malinen <j@w1.fi>
