If an MD IE is supplied to netdev_connect, pass that MD IE in the
associate request, then validate and handle the MD IE and FT IE in the
associate response from AP.
Don't decide on the AKM suite to use when the bss entries are received
and processed, instead select the suite when the connection is triggered
using a new function device_select_akm_suite, similar to
wiphy_select_cipher(). Describing the AKM suite through flags will be
more difficult when more than 2 suites per security type are supported.
Also handle the wiphy_select_cipher 0 return value when no cipher can be
selected.
The len parameter was only used so it could be validated against ie[1],
but since it was not checked to be > 2, it must have been validated
already, the check was redundant. In any case all users directly
passed ie[1] as len anyway. This makes it consistent with the ie
parsers and builders which didn't require a length.
In many cases the pairwise and group cipher information is not the only
information needed from the BSS RSN/WPA elements in order to make a
decision. For example, th MFPC/MFPR bits might be needed, or
pre-authentication capability bits, group management ciphers, etc.
This patch refactors bss_get_supported_ciphers into the more general
scan_bss_get_rsn_info function
Use the org.freedesktop.DBus.Properties interfaces on objects with
properties and drop the old style GetProperty/SetProperty methods on
individual interfaces. Agent and KnownNetworks have no properties at
this time so don't add org.freedesktop.DBus.Properties interfaces.
This also fixes a potential buffer overflow since the ssid was cast to a
string inside network_create. However, ssid is a buffer of 32 bytes,
and would not be null-terminated in the case of a 32-byte SSID.
Calling device_disassociated inside disconnect_cb was mostly pointless.
Most attributes were already cleared by device_disconnect() when
initiating the disconnection procedure.
This patch also modifies the logic for triggering the autoconnect. If
the user initiated the disconnect call, then autoconnect should not be
triggered. If the disconnect was triggered by other means, then iwd
will still enter autoconnect mode.
All of the abortion logic is invoked when device_disconnect is called.
So there's no point calling device_disassociated in this case. This
also prevents us from entering into autoconnect mode too early.
Also, remove the check for device->state == DEVICE_STATE_CONNECTING.
device_connect_cb should always called when the state is CONNECTING.
If this is not so, it indicates a bug inside the netdev layer.
This was introduced by commit f468fceb02.
However, after commit 2d78f51fac66b9beff03a56f12e5fb8456625f07, the
connect_cb is called from inside netdev_disconnect. This in turn causes
the dbus-reply to be sent out if needed. So by the time we get to the
code in question, connect_pending is always NULL.
Change the path for net.connman.iwd.Device objects to /phyX/Y and
register net.connman.iwd.Adapter at /phyX grouping devices of the same
wiphy.
Turns out no changes to the test/* scripts are needed.
The boolean property indicates if a scan is ongoing. Only the scans
triggered by device.c are reflected (not the ones from WSC) because only
those scans affect the list of networks seen by Dbus.
If Disconnect is called during an ongoing connection attempt send a
CMD_DEAUTHENTICATE command same as when we're already connected, and
send a reply to potential dbus Connect call.
Instead of calling the device added or removed callback when the
interface is detected, call it when interface goes up or down. This
only affects the addition and removal of the WSC interface now.
During the network_info refactoring the adding of the connected BSS to
device->bss_list in case it is not in the scan results has moved to
after the l_hashmap_foreach_remove call meaning that the network could
be removed even though it is still pointed at by
device->connected networks. Reverse the order to what it was before.
Alternatively network_process network could take not of the fact the
network is connected and not call network_remove on it leaving it with
an empty bss_list.
It is probably rare that a disconnect should fail but if it happens the
device->state is not returned to CONNECTED and I'm not sure if it should
be, so the ConnectedNetwork property and other bits should probably be
reset at the start of the disconnection instead of at the end.
Also check if state is CONNECTED before calling network_disconnected
because network_connected may have not been called yet.
The eapol state machine parameters are now built inside device.c when
the network connection is attempted. The reason is that the device
object knows about network settings, wiphy constraints and should
contain the main 'management' logic.
netdev now manages the actual low-level process of building association
messages, detecting authentication events, etc.
Turn netdev watches into device watches. The intent is to refactor out
netdev specific details into its own class and move device specific
logic into device.c away from wiphy.c
