Kernel/iwd
3
0
Fork 0
mirror of https://git.kernel.org/pub/scm/network/wireless/iwd.git synced 2024-11-09 13:39:23 +01:00
Code Releases Activity
338577bb3a
iwd/src/ap.c

4846 lines
124 KiB
C
Raw Normal View History

ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
/*
*
* Wireless daemon for Linux
*
treewide: Move the Intel copyright forward to 2019
2019-10-25 00:43:08 +02:00
* Copyright (C) 2017-2019 Intel Corporation. All rights reserved.
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*
*/
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif
#include <errno.h>
#include <linux/if_ether.h>
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
#include <netinet/in.h>
#include <arpa/inet.h>
ap: set link/operstate on AP start/stop Prior to this netdev_connect_ok set setting this which really only applies to station mode. In addition this happens for each new station that connects to the AP. Instead set the operstate / link mode when AP starts and stops.
2021-04-28 01:34:52 +02:00
#include <linux/if.h>
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
#include <ell/ell.h>
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
#include "linux/nl80211.h"
treewide: Use ell's useful.h header
2021-03-12 04:46:09 +01:00
#include "ell/useful.h"
ap: fix build with uclibc explicit_bzero is used in src/ap.c since commit d55e00b31d7bccdbb2ea1cdeb0a749df77a51e47 but src/missing.h is not included, as a result build with uclibc fails on: /srv/storage/autobuild/run/instance-1/output-1/host/lib/gcc/xtensa-buildroot-linux-uclibc/9.3.0/../../../../xtensa-buildroot-linux-uclibc/bin/ld: src/ap.o: in function `ap_probe_req_cb': ap.c:(.text+0x23d8): undefined reference to `explicit_bzero' Fixes: - http://autobuild.buildroot.org/results/c7a0096a269bfc52bd8e23d453d36d5bfb61441d
2020-09-10 08:18:00 +02:00
#include "src/missing.h"
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
#include "src/iwd.h"
module: Move declarations into separate header file
2019-11-07 23:33:51 +01:00
#include "src/module.h"
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
#include "src/scan.h"
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
#include "src/netdev.h"
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
#include "src/wiphy.h"
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
#include "src/crypto.h"
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
#include "src/ie.h"
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
#include "src/mpdu.h"
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
#include "src/util.h"
ap: Send PTK handshake frame 1 of 4
2017-09-22 05:06:41 +02:00
#include "src/eapol.h"
#include "src/handshake.h"
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
#include "src/dbus.h"
build: Rename nl80211_util.[ch] into nl80211util.[ch]
2018-10-14 05:41:06 +02:00
#include "src/nl80211util.h"
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
#include "src/frame-xchg.h"
ap: Push Button mode API and beacon changes This adds the API for putting the AP in Push Button mode, which we'll need to P2P GO side but may be useful on its own too. A WSC IE is added to our beacons and probe responses indicating whether the PBC mode is active.
2020-08-28 14:46:46 +02:00
#include "src/wscutil.h"
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
#include "src/eap-wsc.h"
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
#include "src/ip-pool.h"
treewide: Parse EnableNetworkConfiguration in one place Add netconfig_enabled() and use that in all places that want to know whether network configuration is enabled. Drop the enable_network_config deprecated setting, which was only being handled in one of these 5 or so places.
2021-10-21 10:50:30 +02:00
#include "src/netconfig.h"
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
#include "src/ap.h"
ap: add StartProfile DBus method Users can now start an AP from settings based on a profile on disk. The only argument is the SSID which will be used to lookup the profile. If no profile is found a NotFound error will be returned. Any invalid profiles will result in an Invalid return.
2020-11-02 19:28:49 +01:00
#include "src/storage.h"
ap: add AP diagnostic interface This adds a new AccessPointDiagnostic interface. This interface provides similar low level functionality as StationDiagnostic, but for when IWD is in AP mode. This uses netdev_get_all_stations which will dump all stations, parse, and return each station in an individual callback. Once the dump is complete the destroy is called and all data is packaged as an array of dictionaries.
2021-01-22 21:38:45 +01:00
#include "src/diagnostic.h"
treewide: update scan utility usage to use band_*
2021-11-30 00:12:59 +01:00
#include "src/band.h"
ap: add support for scanning while in AP mode Scanning while in AP mode is somewhat of an edge case, but it does have some usefulness specifically with onboarding new devices, i.e. a new device starts an AP, a station connects and provides the new device with network credentials, the new device switches to station mode and connects to the desired network. In addition this could be used later for ACS (though this is a bit overkill for IWD's access point needs). Since AP performance is basically non-existant while scanning this feature is meant to be used in a limited scope. Two DBus API's were added which mirror the station interface: Scan and GetOrderedNetworks. Scan is no different than the station variant, and will perform an active scan on all channels. GetOrderedNetworks diverges from station and simply returns an array of dictionaries containing basic information about networks: { Name: <ssid> SignalStrength: <mBm> Security: <psk, open, or 8021x> } Limitations: - Hidden networks are not supported. This isn't really possible since the SSID's are unknown from the AP perspective. - Sharing scan results with station is not supported. This would be a convenient improvement in the future with respect to onboarding new devices. The scan could be performed in AP mode, then switch to station and connect immediately without needing to rescan. A quick hack could better this situation by not flushing scan results in station (if the kernel retains these on an iftype change).
2022-02-28 19:49:37 +01:00
#include "src/common.h"
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
struct ap_state {
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
struct netdev *netdev;
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
struct l_genl_family *nl80211;
ap: Pass "ops" struct to ap_start() Pass the event callback function pointer in a "struct ap_ops" instead of as individual ap_start() argument to make adding new callbacks easier.
2020-09-16 11:17:49 +02:00
const struct ap_ops *ops;
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap_stopped_func_t stopped_func;
void *user_data;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
char ssid[33];
char passphrase[64];
uint8_t psk[32];
ap: add support for 5GHz frequencies in AP mode This enables IWD to use 5GHz frequencies in AP mode. Currently 6GHz is not supported so we can assume a [General].Channel value 36 or above indicates the 5GHz band. It should be noted that the system will probably need a regulatory domain set in order for 5GHz to be allowed in AP mode. This is due to world roaming (00) restricting any/all 5GHz frequencies. This can be accomplished by setting main.conf [General].Country=CC to the country this AP will operate in.
2022-12-09 20:39:28 +01:00
enum band_freq band;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
uint8_t channel;
ap: generate chandef for starting AP To include HT support a chandef needs to be created for whatever frequency is being used. This allows IWD to provide a secondary channel to the kernel in the case of 40MHz operation. Now the AP will generate a chandef when starting based on the channel set in the user profile (or default). If HT is not supported the chandef width is set to 20MHz no-HT, otherwise band_freq_to_ht_chandef is used.
2022-12-28 23:32:30 +01:00
struct band_chandef chandef;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
uint8_t *authorized_macs;
unsigned int authorized_macs_num;
char wsc_name[33];
struct wsc_primary_device_type wsc_primary_device_type;
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
unsigned int ciphers;
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
enum ie_rsn_cipher_suite group_cipher;
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
uint32_t beacon_interval;
struct l_uintset *rates;
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
uint32_t start_stop_cmd_id;
ap: React to NL80211_CMD_STOP_AP events These events will tell use when our AP gets stopped without our request, for example due to suspend/resume.
2020-02-15 01:17:46 +01:00
uint32_t mlme_watch;
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
uint8_t gtk[CRYPTO_MAX_GTK_LEN];
uint8_t gtk_index;
ap: WSC Probe Request processing logic Implement the caching of WSC probe requests -- when an Enrollee later associates to start registration we need to have its Probe Request on file. Also use this cache for PBC "Session Overlap" detection.
2020-08-28 14:46:47 +02:00
struct l_queue *wsc_pbc_probes;
ap: Push Button mode API and beacon changes This adds the API for putting the AP in Push Button mode, which we'll need to P2P GO side but may be useful on its own too. A WSC IE is added to our beacons and probe responses indicating whether the PBC mode is active.
2020-08-28 14:46:46 +02:00
struct l_timeout *wsc_pbc_timeout;
uint16_t wsc_dpid;
uint8_t wsc_uuid_r[16];
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
uint16_t last_aid;
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
struct l_queue *sta_states;
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
struct l_dhcp_server *netconfig_dhcp;
ap: Save AP address as l_rtnl_address Change the char *addr_str and uint8_t prefix_len pair to an l_rtnl_address object and use ell/rtnl.h utilities that use that directly. Extend broadcast_from_ip to handle prefix_len.
2021-05-28 02:47:35 +02:00
struct l_rtnl_address *netconfig_addr4;
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
uint32_t rtnl_add_cmd;
ap: Add MACs to FILS IP Assignment responses Try to include the gateway and DNS MAC addresses in the corresponding fields in the FILS IP Address Assignment IEs we send to the clients.
2021-08-28 05:23:05 +02:00
uint32_t rtnl_get_gateway4_mac_cmd;
uint32_t rtnl_get_dns4_mac_cmd;
uint8_t netconfig_gateway4_mac[6];
uint8_t netconfig_dns4_mac[6];
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
ap: add support for scanning while in AP mode Scanning while in AP mode is somewhat of an edge case, but it does have some usefulness specifically with onboarding new devices, i.e. a new device starts an AP, a station connects and provides the new device with network credentials, the new device switches to station mode and connects to the desired network. In addition this could be used later for ACS (though this is a bit overkill for IWD's access point needs). Since AP performance is basically non-existant while scanning this feature is meant to be used in a limited scope. Two DBus API's were added which mirror the station interface: Scan and GetOrderedNetworks. Scan is no different than the station variant, and will perform an active scan on all channels. GetOrderedNetworks diverges from station and simply returns an array of dictionaries containing basic information about networks: { Name: <ssid> SignalStrength: <mBm> Security: <psk, open, or 8021x> } Limitations: - Hidden networks are not supported. This isn't really possible since the SSID's are unknown from the AP perspective. - Sharing scan results with station is not supported. This would be a convenient improvement in the future with respect to onboarding new devices. The scan could be performed in AP mode, then switch to station and connect immediately without needing to rescan. A quick hack could better this situation by not flushing scan results in station (if the kernel retains these on an iftype change).
2022-02-28 19:49:37 +01:00
uint32_t scan_id;
struct l_dbus_message *scan_pending;
struct l_queue *networks;
ap: support PTK rekeys This adds support for rekeys to AP mode. A single timer is used and reset to the next station needing a rekey. A default rekey timer of 600 seconds is used unless the profile sets a timeout.
2023-01-18 22:18:00 +01:00
struct l_timeout *rekey_timeout;
unsigned int rekey_time;
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
bool started : 1;
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
bool gtk_set : 1;
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
bool netconfig_set_addr4 : 1;
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
bool in_event : 1;
bool free_pending : 1;
ap: add support for scanning while in AP mode Scanning while in AP mode is somewhat of an edge case, but it does have some usefulness specifically with onboarding new devices, i.e. a new device starts an AP, a station connects and provides the new device with network credentials, the new device switches to station mode and connects to the desired network. In addition this could be used later for ACS (though this is a bit overkill for IWD's access point needs). Since AP performance is basically non-existant while scanning this feature is meant to be used in a limited scope. Two DBus API's were added which mirror the station interface: Scan and GetOrderedNetworks. Scan is no different than the station variant, and will perform an active scan on all channels. GetOrderedNetworks diverges from station and simply returns an array of dictionaries containing basic information about networks: { Name: <ssid> SignalStrength: <mBm> Security: <psk, open, or 8021x> } Limitations: - Hidden networks are not supported. This isn't really possible since the SSID's are unknown from the AP perspective. - Sharing scan results with station is not supported. This would be a convenient improvement in the future with respect to onboarding new devices. The scan could be performed in AP mode, then switch to station and connect immediately without needing to rescan. A quick hack could better this situation by not flushing scan results in station (if the kernel retains these on an iftype change).
2022-02-28 19:49:37 +01:00
bool scanning : 1;
ap: include WMM parameter IE The WMM parameter IE is expected by the linux kernel for any AP supporting HT/VHT etc. IWD won't actually use WMM and its not clear exactly why the kernel uses this restriction, but regardless it must be included to support HT.
2022-12-28 23:32:29 +01:00
bool supports_ht : 1;
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
};
struct sta_state {
uint8_t addr[6];
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
bool associated;
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
bool rsna;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
uint16_t aid;
struct mmpdu_field_capability capability;
uint16_t listen_interval;
struct l_uintset *rates;
uint32_t assoc_resp_cmd_id;
struct ap_state *ap;
ap: Pass frame IEs from clients to the ap_state user Pass the string IEs from the incoming STA association frames to the user in the AP event data. I drop ap_event_station_added_data.rsn_ie because that probably wasn't going to ever be useful and the RSN IE is included in the .assoc_ies array in any case.
2021-03-10 23:06:31 +01:00
uint8_t *assoc_ies;
size_t assoc_ies_len;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
uint8_t *assoc_rsne;
ap: removed code that was moved to eapol All handshake packet handling has been removed from ap and moved to eapol. After association, the AP registers a new authenticator state machine which handles the AP side of the handshake. AP will receive a handshake event once the 4-way handshake is complete.
2018-06-27 19:06:31 +02:00
struct eapol_sm *sm;
struct handshake_state *hs;
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
uint32_t gtk_query_cmd_id;
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
struct l_idle *stop_handshake_work;
struct l_settings *wsc_settings;
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
uint8_t wsc_uuid_e[16];
bool wsc_v2;
ap: Implement P2P GO-side 4-way handshake IP Allocation Use the struct handshake_state::support_ip_allocation field already supported in eapol.c authenticator side to enable the P2P IP Allocation mechanism in ap.c. Add the P2P_GROUP_CAP_IP_ALLOCATION bit in P2P group capabilities to signal the feature is now supported. There's no harm in enabling this feature in every AP (not just P2P Group Owner) but the clients won't know whether we support it other than through that P2P-specific group capability bit.
2021-08-23 16:14:23 +02:00
struct l_dhcp_lease *ip_alloc_lease;
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
bool ip_alloc_sent;
ap: support PTK rekeys This adds support for rekeys to AP mode. A single timer is used and reset to the next station needing a rekey. A default rekey timer of 600 seconds is used unless the profile sets a timeout.
2023-01-18 22:18:00 +01:00
uint64_t rekey_time;
ap: build HT Capabilities/Operation elements If supported this will include the HT capabilities and HT operations elements in beacons/probes. Some shortcuts were taken here since not all the information is currently parsed from the hardware. Namely the HT operation element does not include the basic MCS set. Still, this will at least show stations that the AP is capable of more than just basic rates. The builders themselves are structured similar to the basic rates builder where they build only the contents and return the length. The caller must set the type/length manually. This is to support the two use cases of using with an IE builder vs direct pointer.
2022-12-28 23:32:31 +01:00
bool ht_support : 1;
bool ht_greenfield : 1;
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
};
ap: WSC Probe Request processing logic Implement the caching of WSC probe requests -- when an Enrollee later associates to start registration we need to have its Probe Request on file. Also use this cache for PBC "Session Overlap" detection.
2020-08-28 14:46:47 +02:00
struct ap_wsc_pbc_probe_record {
uint8_t mac[6];
uint8_t uuid_e[16];
uint64_t timestamp;
};
ap: add support for scanning while in AP mode Scanning while in AP mode is somewhat of an edge case, but it does have some usefulness specifically with onboarding new devices, i.e. a new device starts an AP, a station connects and provides the new device with network credentials, the new device switches to station mode and connects to the desired network. In addition this could be used later for ACS (though this is a bit overkill for IWD's access point needs). Since AP performance is basically non-existant while scanning this feature is meant to be used in a limited scope. Two DBus API's were added which mirror the station interface: Scan and GetOrderedNetworks. Scan is no different than the station variant, and will perform an active scan on all channels. GetOrderedNetworks diverges from station and simply returns an array of dictionaries containing basic information about networks: { Name: <ssid> SignalStrength: <mBm> Security: <psk, open, or 8021x> } Limitations: - Hidden networks are not supported. This isn't really possible since the SSID's are unknown from the AP perspective. - Sharing scan results with station is not supported. This would be a convenient improvement in the future with respect to onboarding new devices. The scan could be performed in AP mode, then switch to station and connect immediately without needing to rescan. A quick hack could better this situation by not flushing scan results in station (if the kernel retains these on an iftype change).
2022-02-28 19:49:37 +01:00
struct ap_network {
char ssid[33];
int16_t signal;
enum security security;
};
ap: Refactor global address pool loading Deprecate the global [General].APRanges setting in favour of [IPv4].APAddressPool with an extended (but backwards-compatible) syntax. Drop the existing address pool creation code. The new APAddressPool setting has the same syntax as the profile-local [IPv4].Address setting and the subnet selection code will fall back to the global setting if it's missing, this way we use common code to handle both settings.
2021-05-28 02:47:33 +02:00
static char **global_addr4_strs;
netdev/device: Combine watches There was somewhat overlapping functionality in the device_watch infrastructure as well as the netdev_event_watch. This commit combines the two into a single watch based on the netdev object and cleans up the various interface additions / removals. With this commit the interfaces are created when the netdev/device is switched to Powered=True state AND when the netdev iftype is also in the correct state for that interface. If the device is brought down, then all interfaces except the .Device interface are removed. This will make it easy to implement Device.Mode property properly since most nl80211 devices need to be brought into Powered=False state prior to switching the iftype.
2018-08-18 06:40:49 +02:00
static uint32_t netdev_watch;
ap: Make rtnl global static
2021-05-13 20:47:10 +02:00
static struct l_netlink *rtnl;
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
ap: add support for scanning while in AP mode Scanning while in AP mode is somewhat of an edge case, but it does have some usefulness specifically with onboarding new devices, i.e. a new device starts an AP, a station connects and provides the new device with network credentials, the new device switches to station mode and connects to the desired network. In addition this could be used later for ACS (though this is a bit overkill for IWD's access point needs). Since AP performance is basically non-existant while scanning this feature is meant to be used in a limited scope. Two DBus API's were added which mirror the station interface: Scan and GetOrderedNetworks. Scan is no different than the station variant, and will perform an active scan on all channels. GetOrderedNetworks diverges from station and simply returns an array of dictionaries containing basic information about networks: { Name: <ssid> SignalStrength: <mBm> Security: <psk, open, or 8021x> } Limitations: - Hidden networks are not supported. This isn't really possible since the SSID's are unknown from the AP perspective. - Sharing scan results with station is not supported. This would be a convenient improvement in the future with respect to onboarding new devices. The scan could be performed in AP mode, then switch to station and connect immediately without needing to rescan. A quick hack could better this situation by not flushing scan results in station (if the kernel retains these on an iftype change).
2022-02-28 19:49:37 +01:00
static bool network_match_ssid(const void *a, const void *b)
{
const struct ap_network *network = a;
const char *ssid = b;
return !strcmp(network->ssid, ssid);
}
static int network_signal_compare(const void *a, const void *b, void *user)
{
const struct ap_network *new_network = a;
const struct ap_network *network = b;
return (network->signal > new_network->signal) ? 1 : -1;
}
static struct ap_network *ap_network_find(struct ap_state *ap,
struct scan_bss *bss)
{
char ssid[33];
memcpy(ssid, bss->ssid, bss->ssid_len);
ssid[bss->ssid_len] = '\0';
return l_queue_find(ap->networks, network_match_ssid, ssid);
}
static void ap_network_append(struct ap_state *ap, struct scan_bss *bss)
{
struct ap_network *network;
enum security security;
if (util_ssid_is_hidden(bss->ssid_len, bss->ssid))
return;
network = ap_network_find(ap, bss);
if (!network) {
if (scan_bss_get_security(bss, &security) < 0)
return;
network = l_new(struct ap_network, 1);
network->signal = bss->signal_strength;
network->security = security;
memcpy(network->ssid, bss->ssid, bss->ssid_len);
network->ssid[bss->ssid_len] = '\0';
goto insert;
}
if (bss->signal_strength <= network->signal)
return;
l_queue_remove(ap->networks, network);
network->signal = bss->signal_strength;
insert:
l_queue_insert(ap->networks, network, network_signal_compare, NULL);
}
ap: Stop ongoing handshake on reassociation On a new association or re-association, in addition to forgetting a complete RSN Association, also stop the EAPoL SM to stop any ongoing handshake. Do this in a new function ap_stop_handshake that is now used in a few places that had copies of the same few lines. I'll be adding some more lines to this function for WSC support.
2020-08-28 14:46:45 +02:00
static void ap_stop_handshake(struct sta_state *sta)
{
if (sta->sm) {
eapol_sm_free(sta->sm);
sta->sm = NULL;
}
if (sta->hs) {
handshake_state_free(sta->hs);
sta->hs = NULL;
}
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
if (sta->wsc_settings) {
l_settings_free(sta->wsc_settings);
sta->wsc_settings = NULL;
}
if (sta->stop_handshake_work) {
l_idle_remove(sta->stop_handshake_work);
sta->stop_handshake_work = NULL;
}
}
static void ap_stop_handshake_work(struct l_idle *idle, void *user_data)
{
struct sta_state *sta = user_data;
ap_stop_handshake(sta);
ap: Stop ongoing handshake on reassociation On a new association or re-association, in addition to forgetting a complete RSN Association, also stop the EAPoL SM to stop any ongoing handshake. Do this in a new function ap_stop_handshake that is now used in a few places that had copies of the same few lines. I'll be adding some more lines to this function for WSC support.
2020-08-28 14:46:45 +02:00
}
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
static void ap_sta_free(void *data)
{
struct sta_state *sta = data;
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
struct ap_state *ap = sta->ap;
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
ap: remove rates requirement for fmac cards It was seen that some full mac cards/drivers do not include any rate information with the NEW_STATION event. This was causing the NEW_STATION event to be ignored, preventing AP mode from working on these cards. Since the full mac path does not even require sta->rates the parsing can be removed completely.
2021-04-12 20:22:45 +02:00
if (sta->rates)
l_uintset_free(sta->rates);
ap: Pass frame IEs from clients to the ap_state user Pass the string IEs from the incoming STA association frames to the user in the AP event data. I drop ap_event_station_added_data.rsn_ie because that probably wasn't going to ever be useful and the RSN IE is included in the .assoc_ies array in any case.
2021-03-10 23:06:31 +01:00
l_free(sta->assoc_ies);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
if (sta->assoc_resp_cmd_id)
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
l_genl_family_cancel(ap->nl80211, sta->assoc_resp_cmd_id);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
if (sta->gtk_query_cmd_id)
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
l_genl_family_cancel(ap->nl80211, sta->gtk_query_cmd_id);
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
ap: Implement P2P GO-side 4-way handshake IP Allocation Use the struct handshake_state::support_ip_allocation field already supported in eapol.c authenticator side to enable the P2P IP Allocation mechanism in ap.c. Add the P2P_GROUP_CAP_IP_ALLOCATION bit in P2P group capabilities to signal the feature is now supported. There's no harm in enabling this feature in every AP (not just P2P Group Owner) but the clients won't know whether we support it other than through that P2P-specific group capability bit.
2021-08-23 16:14:23 +02:00
if (sta->ip_alloc_lease && ap->netconfig_dhcp)
l_dhcp_server_lease_remove(ap->netconfig_dhcp,
sta->ip_alloc_lease);
ap: Stop ongoing handshake on reassociation On a new association or re-association, in addition to forgetting a complete RSN Association, also stop the EAPoL SM to stop any ongoing handshake. Do this in a new function ap_stop_handshake that is now used in a few places that had copies of the same few lines. I'll be adding some more lines to this function for WSC support.
2020-08-28 14:46:45 +02:00
ap_stop_handshake(sta);
ap: Implement EAPOL frame retries
2017-09-30 04:28:10 +02:00
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
l_free(sta);
}
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
static void ap_reset(struct ap_state *ap)
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
{
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
struct netdev *netdev = ap->netdev;
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
explicit_bzero(ap->passphrase, sizeof(ap->passphrase));
explicit_bzero(ap->psk, sizeof(ap->psk));
if (ap->authorized_macs_num) {
l_free(ap->authorized_macs);
ap->authorized_macs_num = 0;
}
ap: Don't defer ap_reset when ap_free called in event Previously we added logic to defer doing anything in ap_free() to after the AP event handler has returned so that ap_event() has a chance to inform whoever called it that the ap_state has been freed. But there's also a chance that the event handler is destroying both the AP and the netdev it runs on, so after the handler has returned we can't even use netdev_get_wdev_id or netdev_get_ifindex. The easiest solution seems to be to call ap_reset() in ap_free() even if we're within an event handler to ensure we no longer need any external objects. Also make sure ap_reset() can be called multiple times. Another option would be to watch for NETDEV_WATCH_EVENT_DEL and remove our reference to the netdev (because there's no need actually call l_rtnl_ifaddr_delete or frame_watch_wdev_remove if the netdev was destroyed -- frame_watch already tracks netdev removals), or to save just the ifindex and the wdev id...
2022-01-21 11:24:37 +01:00
if (ap->mlme_watch) {
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
l_genl_family_unregister(ap->nl80211, ap->mlme_watch);
ap: Don't defer ap_reset when ap_free called in event Previously we added logic to defer doing anything in ap_free() to after the AP event handler has returned so that ap_event() has a chance to inform whoever called it that the ap_state has been freed. But there's also a chance that the event handler is destroying both the AP and the netdev it runs on, so after the handler has returned we can't even use netdev_get_wdev_id or netdev_get_ifindex. The easiest solution seems to be to call ap_reset() in ap_free() even if we're within an event handler to ensure we no longer need any external objects. Also make sure ap_reset() can be called multiple times. Another option would be to watch for NETDEV_WATCH_EVENT_DEL and remove our reference to the netdev (because there's no need actually call l_rtnl_ifaddr_delete or frame_watch_wdev_remove if the netdev was destroyed -- frame_watch already tracks netdev removals), or to save just the ifindex and the wdev id...
2022-01-21 11:24:37 +01:00
ap->mlme_watch = 0;
}
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
frame_watch_wdev_remove(netdev_get_wdev_id(netdev));
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
ap: Don't defer ap_reset when ap_free called in event Previously we added logic to defer doing anything in ap_free() to after the AP event handler has returned so that ap_event() has a chance to inform whoever called it that the ap_state has been freed. But there's also a chance that the event handler is destroying both the AP and the netdev it runs on, so after the handler has returned we can't even use netdev_get_wdev_id or netdev_get_ifindex. The easiest solution seems to be to call ap_reset() in ap_free() even if we're within an event handler to ensure we no longer need any external objects. Also make sure ap_reset() can be called multiple times. Another option would be to watch for NETDEV_WATCH_EVENT_DEL and remove our reference to the netdev (because there's no need actually call l_rtnl_ifaddr_delete or frame_watch_wdev_remove if the netdev was destroyed -- frame_watch already tracks netdev removals), or to save just the ifindex and the wdev id...
2022-01-21 11:24:37 +01:00
if (ap->start_stop_cmd_id) {
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
l_genl_family_cancel(ap->nl80211, ap->start_stop_cmd_id);
ap: Don't defer ap_reset when ap_free called in event Previously we added logic to defer doing anything in ap_free() to after the AP event handler has returned so that ap_event() has a chance to inform whoever called it that the ap_state has been freed. But there's also a chance that the event handler is destroying both the AP and the netdev it runs on, so after the handler has returned we can't even use netdev_get_wdev_id or netdev_get_ifindex. The easiest solution seems to be to call ap_reset() in ap_free() even if we're within an event handler to ensure we no longer need any external objects. Also make sure ap_reset() can be called multiple times. Another option would be to watch for NETDEV_WATCH_EVENT_DEL and remove our reference to the netdev (because there's no need actually call l_rtnl_ifaddr_delete or frame_watch_wdev_remove if the netdev was destroyed -- frame_watch already tracks netdev removals), or to save just the ifindex and the wdev id...
2022-01-21 11:24:37 +01:00
ap->start_stop_cmd_id = 0;
}
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
ap: Don't defer ap_reset when ap_free called in event Previously we added logic to defer doing anything in ap_free() to after the AP event handler has returned so that ap_event() has a chance to inform whoever called it that the ap_state has been freed. But there's also a chance that the event handler is destroying both the AP and the netdev it runs on, so after the handler has returned we can't even use netdev_get_wdev_id or netdev_get_ifindex. The easiest solution seems to be to call ap_reset() in ap_free() even if we're within an event handler to ensure we no longer need any external objects. Also make sure ap_reset() can be called multiple times. Another option would be to watch for NETDEV_WATCH_EVENT_DEL and remove our reference to the netdev (because there's no need actually call l_rtnl_ifaddr_delete or frame_watch_wdev_remove if the netdev was destroyed -- frame_watch already tracks netdev removals), or to save just the ifindex and the wdev id...
2022-01-21 11:24:37 +01:00
if (ap->rtnl_add_cmd) {
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
l_netlink_cancel(rtnl, ap->rtnl_add_cmd);
ap: Don't defer ap_reset when ap_free called in event Previously we added logic to defer doing anything in ap_free() to after the AP event handler has returned so that ap_event() has a chance to inform whoever called it that the ap_state has been freed. But there's also a chance that the event handler is destroying both the AP and the netdev it runs on, so after the handler has returned we can't even use netdev_get_wdev_id or netdev_get_ifindex. The easiest solution seems to be to call ap_reset() in ap_free() even if we're within an event handler to ensure we no longer need any external objects. Also make sure ap_reset() can be called multiple times. Another option would be to watch for NETDEV_WATCH_EVENT_DEL and remove our reference to the netdev (because there's no need actually call l_rtnl_ifaddr_delete or frame_watch_wdev_remove if the netdev was destroyed -- frame_watch already tracks netdev removals), or to save just the ifindex and the wdev id...
2022-01-21 11:24:37 +01:00
ap->rtnl_add_cmd = 0;
}
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
ap: Add MACs to FILS IP Assignment responses Try to include the gateway and DNS MAC addresses in the corresponding fields in the FILS IP Address Assignment IEs we send to the clients.
2021-08-28 05:23:05 +02:00
if (ap->rtnl_get_gateway4_mac_cmd) {
l_netlink_cancel(rtnl, ap->rtnl_get_gateway4_mac_cmd);
ap->rtnl_get_gateway4_mac_cmd = 0;
}
if (ap->rtnl_get_dns4_mac_cmd) {
l_netlink_cancel(rtnl, ap->rtnl_get_dns4_mac_cmd);
ap->rtnl_get_dns4_mac_cmd = 0;
}
ap: Don't defer ap_reset when ap_free called in event Previously we added logic to defer doing anything in ap_free() to after the AP event handler has returned so that ap_event() has a chance to inform whoever called it that the ap_state has been freed. But there's also a chance that the event handler is destroying both the AP and the netdev it runs on, so after the handler has returned we can't even use netdev_get_wdev_id or netdev_get_ifindex. The easiest solution seems to be to call ap_reset() in ap_free() even if we're within an event handler to ensure we no longer need any external objects. Also make sure ap_reset() can be called multiple times. Another option would be to watch for NETDEV_WATCH_EVENT_DEL and remove our reference to the netdev (because there's no need actually call l_rtnl_ifaddr_delete or frame_watch_wdev_remove if the netdev was destroyed -- frame_watch already tracks netdev removals), or to save just the ifindex and the wdev id...
2022-01-21 11:24:37 +01:00
l_queue_destroy(l_steal_ptr(ap->sta_states), ap_sta_free);
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
if (ap->rates)
ap: Don't defer ap_reset when ap_free called in event Previously we added logic to defer doing anything in ap_free() to after the AP event handler has returned so that ap_event() has a chance to inform whoever called it that the ap_state has been freed. But there's also a chance that the event handler is destroying both the AP and the netdev it runs on, so after the handler has returned we can't even use netdev_get_wdev_id or netdev_get_ifindex. The easiest solution seems to be to call ap_reset() in ap_free() even if we're within an event handler to ensure we no longer need any external objects. Also make sure ap_reset() can be called multiple times. Another option would be to watch for NETDEV_WATCH_EVENT_DEL and remove our reference to the netdev (because there's no need actually call l_rtnl_ifaddr_delete or frame_watch_wdev_remove if the netdev was destroyed -- frame_watch already tracks netdev removals), or to save just the ifindex and the wdev id...
2022-01-21 11:24:37 +01:00
l_uintset_free(l_steal_ptr(ap->rates));
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
ap: Don't defer ap_reset when ap_free called in event Previously we added logic to defer doing anything in ap_free() to after the AP event handler has returned so that ap_event() has a chance to inform whoever called it that the ap_state has been freed. But there's also a chance that the event handler is destroying both the AP and the netdev it runs on, so after the handler has returned we can't even use netdev_get_wdev_id or netdev_get_ifindex. The easiest solution seems to be to call ap_reset() in ap_free() even if we're within an event handler to ensure we no longer need any external objects. Also make sure ap_reset() can be called multiple times. Another option would be to watch for NETDEV_WATCH_EVENT_DEL and remove our reference to the netdev (because there's no need actually call l_rtnl_ifaddr_delete or frame_watch_wdev_remove if the netdev was destroyed -- frame_watch already tracks netdev removals), or to save just the ifindex and the wdev id...
2022-01-21 11:24:37 +01:00
l_queue_destroy(l_steal_ptr(ap->wsc_pbc_probes), l_free);
ap: Fix leaking of ap->wsc_pbc_timeout
2021-06-04 03:50:40 +02:00
l_timeout_remove(ap->wsc_pbc_timeout);
ap: WSC Probe Request processing logic Implement the caching of WSC probe requests -- when an Enrollee later associates to start registration we need to have its Probe Request on file. Also use this cache for PBC "Session Overlap" detection.
2020-08-28 14:46:47 +02:00
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap->started = false;
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
/* Delete IP if one was set by IWD */
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (ap->netconfig_set_addr4) {
ap: Save AP address as l_rtnl_address Change the char *addr_str and uint8_t prefix_len pair to an l_rtnl_address object and use ell/rtnl.h utilities that use that directly. Extend broadcast_from_ip to handle prefix_len.
2021-05-28 02:47:35 +02:00
l_rtnl_ifaddr_delete(rtnl, netdev_get_ifindex(netdev),
ap->netconfig_addr4, NULL, NULL, NULL);
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
ap->netconfig_set_addr4 = false;
}
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
ap: Save AP address as l_rtnl_address Change the char *addr_str and uint8_t prefix_len pair to an l_rtnl_address object and use ell/rtnl.h utilities that use that directly. Extend broadcast_from_ip to handle prefix_len.
2021-05-28 02:47:35 +02:00
l_rtnl_address_free(l_steal_ptr(ap->netconfig_addr4));
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (ap->netconfig_dhcp) {
l_dhcp_server_destroy(ap->netconfig_dhcp);
ap->netconfig_dhcp = NULL;
ap: Do not leak ap->server ap_reset() seems to be called whenever the AP is stopped or removed due to interface shutdown. For some reason ap_reset did not remove the DHCP server object, resulting in leaks: ==211== at 0x483879F: malloc (vg_replace_malloc.c:307) ==211== by 0x46B5AD: l_malloc (util.c:62) ==211== by 0x49B0E2: l_dhcp_server_new (dhcp-server.c:715) ==211== by 0x433AA3: ap_setup_dhcp (ap.c:2615) ==211== by 0x433AA3: ap_load_dhcp (ap.c:2645) ==211== by 0x433AA3: ap_load_config (ap.c:2753) ==211== by 0x433AA3: ap_start (ap.c:2885) ==211== by 0x434A96: ap_dbus_start_profile (ap.c:3329) ==211== by 0x482DA9: _dbus_object_tree_dispatch (dbus-service.c:1815) ==211== by 0x47A4D9: message_read_handler (dbus.c:285) ==211== by 0x4720EB: io_callback (io.c:120) ==211== by 0x47130C: l_main_iterate (main.c:478) ==211== by 0x4713DB: l_main_run (main.c:525) ==211== by 0x4713DB: l_main_run (main.c:507) ==211== by 0x4715EB: l_main_run_with_signal (main.c:647) ==211== by 0x403EE1: main (main.c:550)
2021-05-19 22:48:51 +02:00
}
ap: add support for scanning while in AP mode Scanning while in AP mode is somewhat of an edge case, but it does have some usefulness specifically with onboarding new devices, i.e. a new device starts an AP, a station connects and provides the new device with network credentials, the new device switches to station mode and connects to the desired network. In addition this could be used later for ACS (though this is a bit overkill for IWD's access point needs). Since AP performance is basically non-existant while scanning this feature is meant to be used in a limited scope. Two DBus API's were added which mirror the station interface: Scan and GetOrderedNetworks. Scan is no different than the station variant, and will perform an active scan on all channels. GetOrderedNetworks diverges from station and simply returns an array of dictionaries containing basic information about networks: { Name: <ssid> SignalStrength: <mBm> Security: <psk, open, or 8021x> } Limitations: - Hidden networks are not supported. This isn't really possible since the SSID's are unknown from the AP perspective. - Sharing scan results with station is not supported. This would be a convenient improvement in the future with respect to onboarding new devices. The scan could be performed in AP mode, then switch to station and connect immediately without needing to rescan. A quick hack could better this situation by not flushing scan results in station (if the kernel retains these on an iftype change).
2022-02-28 19:49:37 +01:00
if (ap->scan_id) {
scan_cancel(netdev_get_wdev_id(ap->netdev), ap->scan_id);
ap->scan_id = 0;
}
if (ap->networks) {
l_queue_destroy(ap->networks, l_free);
ap->networks = NULL;
}
ap: support PTK rekeys This adds support for rekeys to AP mode. A single timer is used and reset to the next station needing a rekey. A default rekey timer of 600 seconds is used unless the profile sets a timeout.
2023-01-18 22:18:00 +01:00
if (ap->rekey_timeout) {
l_timeout_remove(ap->rekey_timeout);
ap->rekey_timeout = NULL;
}
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
}
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
static bool ap_event_done(struct ap_state *ap, bool prev_in_event)
{
ap->in_event = prev_in_event;
if (!prev_in_event && ap->free_pending) {
ap: Don't defer ap_reset when ap_free called in event Previously we added logic to defer doing anything in ap_free() to after the AP event handler has returned so that ap_event() has a chance to inform whoever called it that the ap_state has been freed. But there's also a chance that the event handler is destroying both the AP and the netdev it runs on, so after the handler has returned we can't even use netdev_get_wdev_id or netdev_get_ifindex. The easiest solution seems to be to call ap_reset() in ap_free() even if we're within an event handler to ensure we no longer need any external objects. Also make sure ap_reset() can be called multiple times. Another option would be to watch for NETDEV_WATCH_EVENT_DEL and remove our reference to the netdev (because there's no need actually call l_rtnl_ifaddr_delete or frame_watch_wdev_remove if the netdev was destroyed -- frame_watch already tracks netdev removals), or to save just the ifindex and the wdev id...
2022-01-21 11:24:37 +01:00
l_genl_family_free(ap->nl80211);
l_free(ap);
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
return true;
}
return ap->free_pending;
}
/*
* Returns true if the AP is considered freed and the caller must avoid
* accessing @ap.
*/
static bool ap_event(struct ap_state *ap, enum ap_event_type event,
const void *event_data)
{
bool prev = ap->in_event;
if (ap->free_pending)
return true;
ap->in_event = true;
ap->ops->handle_event(event, event_data, ap->user_data);
return ap_event_done(ap, prev);
}
ap: support PTK rekeys This adds support for rekeys to AP mode. A single timer is used and reset to the next station needing a rekey. A default rekey timer of 600 seconds is used unless the profile sets a timeout.
2023-01-18 22:18:00 +01:00
static void ap_check_rekeys(struct ap_state *ap);
ap: use netdev_del_station API Removes del station code from AP in favor of netdev_del_station.
2018-07-03 23:36:35 +02:00
static void ap_del_station(struct sta_state *sta, uint16_t reason,
bool disassociate)
{
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
struct ap_state *ap = sta->ap;
netdev: Move CMD_DEL_STATION builder to nl80211util While here, also get rid of netdev_del_station. The only user of this function was in ap.c and it could easily be replaced by invoking the new nl80211_build_del_station function. The callback used by netdev_build_del_station only printed an error and didn't do anything useful. Get rid of it for now.
2023-11-14 16:49:39 +01:00
uint32_t ifindex = netdev_get_ifindex(ap->netdev);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
struct ap_event_station_removed_data event_data;
bool send_event = false;
netdev: Move CMD_DEL_STATION builder to nl80211util While here, also get rid of netdev_del_station. The only user of this function was in ap.c and it could easily be replaced by invoking the new nl80211_build_del_station function. The callback used by netdev_build_del_station only printed an error and didn't do anything useful. Get rid of it for now.
2023-11-14 16:49:39 +01:00
struct l_genl_msg *msg;
uint8_t subtype = disassociate ?
MPDU_MANAGEMENT_SUBTYPE_DISASSOCIATION :
MPDU_MANAGEMENT_SUBTYPE_DEAUTHENTICATION;
msg = nl80211_build_del_station(ifindex, sta->addr, reason, subtype);
l_genl_family_send(ap->nl80211, msg, NULL, NULL, NULL);
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
ap: use netdev_del_station API Removes del station code from AP in favor of netdev_del_station.
2018-07-03 23:36:35 +02:00
sta->associated = false;
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
if (sta->rsna) {
ap: Pass "ops" struct to ap_start() Pass the event callback function pointer in a "struct ap_ops" instead of as individual ap_start() argument to make adding new callbacks easier.
2020-09-16 11:17:49 +02:00
if (ap->ops->handle_event) {
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
memset(&event_data, 0, sizeof(event_data));
event_data.mac = sta->addr;
event_data.reason = reason;
send_event = true;
}
sta->rsna = false;
}
ap: Stop EAPOL negotiation on Disassociation Make sure we interrupt eapol traffic (4-way handshake) if we receive a Disassociation from station. Actually do this in ap_del_station because it's called from both ap_disassoc_cb and ap_success_assoc_resp_cb and seems to make sense in both cases.
2018-08-24 03:37:54 +02:00
ap: Make station removal safer Replace instances of the ap_del_station() + ap_sta_free()/ap_remove_sta() with calls to ap_station_disconnect to make sure we consistently remove the station from the ap->sta_states queue before using ap_del_station(). ap_del_station() may generate an event to the ap.h API user (e.g. P2P) and this may end up tearing down the AP completely. For that scenario we also don't want ap_sta_free() to access sta->ap so we make sure ap_del_station() performs these cleanup steps so that ap_sta_free() has nothing to do that accesses sta->ap.
2021-08-07 04:10:42 +02:00
if (sta->assoc_resp_cmd_id) {
l_genl_family_cancel(ap->nl80211, sta->assoc_resp_cmd_id);
sta->assoc_resp_cmd_id = 0;
}
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
if (sta->gtk_query_cmd_id) {
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
l_genl_family_cancel(ap->nl80211, sta->gtk_query_cmd_id);
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
sta->gtk_query_cmd_id = 0;
}
ap: Stop ongoing handshake on reassociation On a new association or re-association, in addition to forgetting a complete RSN Association, also stop the EAPoL SM to stop any ongoing handshake. Do this in a new function ap_stop_handshake that is now used in a few places that had copies of the same few lines. I'll be adding some more lines to this function for WSC support.
2020-08-28 14:46:45 +02:00
ap_stop_handshake(sta);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
/*
* If the event handler tears the AP down, we've made sure above that
* a subsequent ap_sta_free(sta) has no need to access sta->ap.
*/
if (send_event)
if (ap_event(ap, AP_EVENT_STATION_REMOVED, &event_data))
return;
ap: Expire client's leases on disconnect If netconfig is enabled tell the DHCP server to expire any leases owned by the client that is disconnecting by using l_dhcp_server_expire_by_mac to return the IPs to the IP pool. They're added to the expired list so they'd only be used if there are no other addresses left in the pool and can be reactivated if the client comes back before the address is used by somebody else. This should ensure that we're always able to offer an address to a new client as long as there are fewer concurrent clients than addresses in the configured subnet or IP range.
2021-08-23 16:14:25 +02:00
/*
* Expire any DHCP leases owned by this client when it disconnects to
* make it harder for somebody to DoS the IP pool. If the client
* comes back and the lease is still in the expired leases list they
* will get their IP back.
*/
if (ap->netconfig_dhcp) {
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
bool prev = ap->in_event;
/*
* If the LEASE_EXPIRED event in ap_dhcp_event_cb triggers an
* ap_free(), delay cleanup to avoid destroying the DHCP
* server midway through l_dhcp_server_expire_by_mac().
*/
ap->in_event = true;
ap: Expire client's leases on disconnect If netconfig is enabled tell the DHCP server to expire any leases owned by the client that is disconnecting by using l_dhcp_server_expire_by_mac to return the IPs to the IP pool. They're added to the expired list so they'd only be used if there are no other addresses left in the pool and can be reactivated if the client comes back before the address is used by somebody else. This should ensure that we're always able to offer an address to a new client as long as there are fewer concurrent clients than addresses in the configured subnet or IP range.
2021-08-23 16:14:25 +02:00
sta->ip_alloc_lease = NULL;
l_dhcp_server_expire_by_mac(ap->netconfig_dhcp, sta->addr);
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
ap_event_done(ap, prev);
}
ap: support PTK rekeys This adds support for rekeys to AP mode. A single timer is used and reset to the next station needing a rekey. A default rekey timer of 600 seconds is used unless the profile sets a timeout.
2023-01-18 22:18:00 +01:00
/*
* Set the rekey time to zero which will skip this station when
* determining the next rekey.
*/
sta->rekey_time = 0;
ap_check_rekeys(ap);
}
static void ap_start_rekey(struct ap_state *ap, struct sta_state *sta)
{
l_debug("Rekey STA "MAC, MAC_STR(sta->addr));
eapol_start(sta->sm);
}
static void ap_rekey_timeout(struct l_timeout *timeout, void *user_data)
{
struct ap_state *ap = user_data;
ap_check_rekeys(ap);
}
/*
* Used to check/start any rekeys which are due and reset the rekey timer to the
* next soonest station needing a rekey.
*
* TODO: Could adapt this to also take into account the next GTK rekey and
* service that as well. But GTK rekeys are not yet supported in AP mode.
*/
static void ap_check_rekeys(struct ap_state *ap)
{
const struct l_queue_entry *e;
uint64_t now = l_time_now();
uint64_t next = 0;
if (!ap->rekey_time)
return;
/* Find the station(s) that need a rekey and start it */
for (e = l_queue_get_entries(ap->sta_states); e; e = e->next) {
struct sta_state *sta = e->data;
if (!sta->associated || !sta->rsna || sta->rekey_time == 0)
continue;
if (l_time_before(now, sta->rekey_time)) {
uint64_t diff = l_time_diff(now, sta->rekey_time);
/* Finding the next rekey time */
if (next < diff)
next = diff;
continue;
}
ap_start_rekey(ap, sta);
}
/*
* Set the next rekey to the station needing it the soonest, or remove
* if a single station and wait until the rekey is complete to reset
* the timer.
*/
if (next)
l_timeout_modify(ap->rekey_timeout, l_time_to_secs(next));
else {
l_timeout_remove(ap->rekey_timeout);
ap->rekey_timeout = NULL;
}
}
static void ap_set_sta_rekey_timer(struct ap_state *ap, struct sta_state *sta)
{
if (!ap->rekey_time)
return;
sta->rekey_time = l_time_now() + ap->rekey_time - 1;
/*
* First/only station authenticated, set rekey timer. Any more stations
* will just set their rekey time and be serviced by the single callback
*/
if (!ap->rekey_timeout)
ap->rekey_timeout = l_timeout_create(
l_time_to_secs(ap->rekey_time),
ap_rekey_timeout, ap, NULL);
ap: use netdev_del_station API Removes del station code from AP in favor of netdev_del_station.
2018-07-03 23:36:35 +02:00
}
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
static bool ap_sta_match_addr(const void *a, const void *b)
{
const struct sta_state *sta = a;
return !memcmp(sta->addr, b, 6);
}
netdev: fixed key setting failure If netdev fails to set the keys, there was no way for device/ap to know. A new handshake event was added for this. The key setting failure function was also fixed to support both AP/station iftypes. It will now automatically send either a disconnect or del_station depending on the interface type. In similar manner, netdev_handshake_failed was also modified to support both AP/station iftypes. Now, any handshake event listeners should call netdev_handshake_failed upon a handshake failure event, including AP.
2018-07-03 23:36:33 +02:00
static void ap_remove_sta(struct sta_state *sta)
{
if (!l_queue_remove(sta->ap->sta_states, sta)) {
src: Fix spelling errors spotted by lintian The debian package checking tool lintian spotted a bunch of spelling errors in the built binaries.
2018-08-04 16:53:35 +02:00
l_error("tried to remove station that doesn't exist");
netdev: fixed key setting failure If netdev fails to set the keys, there was no way for device/ap to know. A new handshake event was added for this. The key setting failure function was also fixed to support both AP/station iftypes. It will now automatically send either a disconnect or del_station depending on the interface type. In similar manner, netdev_handshake_failed was also modified to support both AP/station iftypes. Now, any handshake event listeners should call netdev_handshake_failed upon a handshake failure event, including AP.
2018-07-03 23:36:33 +02:00
return;
}
ap_sta_free(sta);
}
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
static void ap_set_sta_cb(struct l_genl_msg *msg, void *user_data)
{
if (l_genl_msg_get_error(msg) < 0)
l_error("SET_STATION failed: %i", l_genl_msg_get_error(msg));
}
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
static void ap_del_key_cb(struct l_genl_msg *msg, void *user_data)
{
if (l_genl_msg_get_error(msg) < 0)
l_debug("DEL_KEY failed: %i", l_genl_msg_get_error(msg));
}
ap: removed code that was moved to eapol All handshake packet handling has been removed from ap and moved to eapol. After association, the AP registers a new authenticator state machine which handles the AP side of the handshake. AP will receive a handshake event once the 4-way handshake is complete.
2018-06-27 19:06:31 +02:00
static void ap_new_rsna(struct sta_state *sta)
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
{
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
struct ap_state *ap = sta->ap;
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
struct ap_event_station_added_data event_data = {};
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap: removed code that was moved to eapol All handshake packet handling has been removed from ap and moved to eapol. After association, the AP registers a new authenticator state machine which handles the AP side of the handshake. AP will receive a handshake event once the 4-way handshake is complete.
2018-06-27 19:06:31 +02:00
l_debug("STA "MAC" authenticated", MAC_STR(sta->addr));
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
sta->rsna = true;
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap: support PTK rekeys This adds support for rekeys to AP mode. A single timer is used and reset to the next station needing a rekey. A default rekey timer of 600 seconds is used unless the profile sets a timeout.
2023-01-18 22:18:00 +01:00
ap_set_sta_rekey_timer(ap, sta);
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
event_data.mac = sta->addr;
event_data.assoc_ies = sta->assoc_ies;
event_data.assoc_ies_len = sta->assoc_ies_len;
ap_event(ap, AP_EVENT_STATION_ADDED, &event_data);
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
}
ap: removed code that was moved to eapol All handshake packet handling has been removed from ap and moved to eapol. After association, the AP registers a new authenticator state machine which handles the AP side of the handshake. AP will receive a handshake event once the 4-way handshake is complete.
2018-06-27 19:06:31 +02:00
static void ap_drop_rsna(struct sta_state *sta)
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
{
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
struct ap_state *ap = sta->ap;
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
struct l_genl_msg *msg;
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
uint32_t ifindex = netdev_get_ifindex(sta->ap->netdev);
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
uint8_t key_id = 0;
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
struct ap_event_station_removed_data event_data = {};
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
sta->rsna = false;
nl80211: introduce nl80211 utility API's Netdev/AP share several NL80211 commands and each has their own builder API's. These were moved into a common file nl80211_util.[ch]. A helper was added to AP for building NEW_STATION to make the associate callback look cleaner (rather than manually building NEW_STATION).
2018-10-08 22:44:09 +02:00
msg = nl80211_build_set_station_unauthorized(ifindex, sta->addr);
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
l_genl_msg_append_attr(msg, NL80211_ATTR_STA_AID, 2, &sta->aid);
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
if (!l_genl_family_send(ap->nl80211, msg, ap_set_sta_cb, NULL, NULL)) {
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
l_genl_msg_unref(msg);
l_error("Issuing SET_STATION failed");
}
msg = l_genl_msg_new_sized(NL80211_CMD_DEL_KEY, 64);
l_genl_msg_append_attr(msg, NL80211_ATTR_IFINDEX, 4, &ifindex);
l_genl_msg_append_attr(msg, NL80211_ATTR_KEY_IDX, 1, &key_id);
l_genl_msg_append_attr(msg, NL80211_ATTR_MAC, 6, sta->addr);
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
if (!l_genl_family_send(ap->nl80211, msg, ap_del_key_cb, NULL, NULL)) {
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
l_genl_msg_unref(msg);
l_error("Issuing DEL_KEY failed");
}
ap: Stop ongoing handshake on reassociation On a new association or re-association, in addition to forgetting a complete RSN Association, also stop the EAPoL SM to stop any ongoing handshake. Do this in a new function ap_stop_handshake that is now used in a few places that had copies of the same few lines. I'll be adding some more lines to this function for WSC support.
2020-08-28 14:46:45 +02:00
ap_stop_handshake(sta);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
event_data.mac = sta->addr;
ap_event(ap, AP_EVENT_STATION_REMOVED, &event_data);
ap: removed code that was moved to eapol All handshake packet handling has been removed from ap and moved to eapol. After association, the AP registers a new authenticator state machine which handles the AP side of the handshake. AP will receive a handshake event once the 4-way handshake is complete.
2018-06-27 19:06:31 +02:00
}
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
static void ap_set_rsn_info(struct ap_state *ap, struct ie_rsn_info *rsn)
{
memset(rsn, 0, sizeof(*rsn));
rsn->akm_suites = IE_RSN_AKM_SUITE_PSK;
rsn->pairwise_ciphers = ap->ciphers;
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
rsn->group_cipher = ap->group_cipher;
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
}
ap: Handle most WSC IEs using ap_write_extra_ies Make the WSC IE processing and writing more self-contained (i.e. so that it can be more easily moved to a separate file if desired) by using the new ap_write_extra_ies() mechanism.
2021-03-11 14:07:24 +01:00
static void ap_wsc_exit_pbc(struct ap_state *ap)
{
if (!ap->wsc_pbc_timeout)
return;
l_timeout_remove(ap->wsc_pbc_timeout);
ap->wsc_dpid = 0;
ap_update_beacon(ap);
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
ap_event(ap, AP_EVENT_PBC_MODE_EXIT, NULL);
ap: Handle most WSC IEs using ap_write_extra_ies Make the WSC IE processing and writing more self-contained (i.e. so that it can be more easily moved to a separate file if desired) by using the new ap_write_extra_ies() mechanism.
2021-03-11 14:07:24 +01:00
}
struct ap_pbc_record_expiry_data {
uint64_t min_time;
const uint8_t *mac;
};
static bool ap_wsc_pbc_record_expire(void *data, void *user_data)
{
struct ap_wsc_pbc_probe_record *record = data;
const struct ap_pbc_record_expiry_data *expiry_data = user_data;
if (record->timestamp > expiry_data->min_time &&
memcmp(record->mac, expiry_data->mac, 6))
return false;
l_free(record);
return true;
}
#define AP_WSC_PBC_MONITOR_TIME 120
#define AP_WSC_PBC_WALK_TIME 120
static void ap_process_wsc_probe_req(struct ap_state *ap, const uint8_t *from,
const uint8_t *wsc_data,
size_t wsc_data_len)
{
struct wsc_probe_request req;
struct ap_pbc_record_expiry_data expiry_data;
struct ap_wsc_pbc_probe_record *record;
uint64_t now;
bool empty;
uint8_t first_sta_addr[6] = {};
const struct l_queue_entry *entry;
if (wsc_parse_probe_request(wsc_data, wsc_data_len, &req) < 0)
return;
if (!(req.config_methods & WSC_CONFIGURATION_METHOD_PUSH_BUTTON))
return;
if (req.device_password_id != WSC_DEVICE_PASSWORD_ID_PUSH_BUTTON)
return;
/* Save the address of the first enrollee record */
record = l_queue_peek_head(ap->wsc_pbc_probes);
if (record)
memcpy(first_sta_addr, record->mac, 6);
now = l_time_now();
/*
* Expire entries older than PBC Monitor Time. While there also drop
* older entries from the same Enrollee that sent us this new Probe
* Request. It's unclear whether we should also match by the UUID-E.
*/
expiry_data.min_time = now - AP_WSC_PBC_MONITOR_TIME * 1000000;
expiry_data.mac = from;
l_queue_foreach_remove(ap->wsc_pbc_probes, ap_wsc_pbc_record_expire,
&expiry_data);
empty = l_queue_isempty(ap->wsc_pbc_probes);
if (!ap->wsc_pbc_probes)
ap->wsc_pbc_probes = l_queue_new();
/* Add new record */
record = l_new(struct ap_wsc_pbc_probe_record, 1);
memcpy(record->mac, from, 6);
memcpy(record->uuid_e, req.uuid_e, sizeof(record->uuid_e));
record->timestamp = now;
l_queue_push_tail(ap->wsc_pbc_probes, record);
/*
* If queue was non-empty and we've added one more record then we
* now have seen more than one PBC enrollee during the PBC Monitor
* Time and must exit "active PBC mode" due to "session overlap".
* WSC v2.0.5 Section 11.3:
* "Within the PBC Monitor Time, if the Registrar receives PBC
* probe requests from more than one Enrollee [...] then the
* Registrar SHALL signal a "session overlap" error. As a result,
* the Registrar shall refuse to enter active PBC mode and shall
* also refuse to perform a PBC-based Registration Protocol
* exchange [...]"
*/
if (empty)
return;
if (ap->wsc_pbc_timeout) {
l_debug("Exiting PBC mode due to Session Overlap");
ap_wsc_exit_pbc(ap);
}
/*
* "If the Registrar is engaged in PBC Registration Protocol
* exchange with an Enrollee and receives a Probe Request or M1
* Message from another Enrollee, then the Registrar should
* signal a "session overlap" error".
*
* For simplicity just interrupt the handshake with that enrollee.
*/
for (entry = l_queue_get_entries(ap->sta_states); entry;
entry = entry->next) {
struct sta_state *sta = entry->data;
if (!sta->associated || sta->assoc_rsne)
continue;
/*
* Check whether this enrollee is in PBC Registration
* Protocol by comparing its mac with the first (and only)
* record we had in ap->wsc_pbc_probes. If we had more
* than one record we wouldn't have been in
* "active PBC mode".
*/
if (memcmp(sta->addr, first_sta_addr, 6) ||
!memcmp(sta->addr, from, 6))
continue;
l_debug("Interrupting handshake with %s due to Session Overlap",
util_address_to_string(sta->addr));
if (sta->hs) {
netdev_handshake_failed(sta->hs,
MMPDU_REASON_CODE_DISASSOC_AP_BUSY);
sta->sm = NULL;
}
ap_remove_sta(sta);
}
}
ap: Deduplicate writing authorized MACs into wsc_* structs
2021-04-28 17:48:37 +02:00
static void ap_write_authorized_macs(struct ap_state *ap,
size_t out_len, uint8_t *out)
{
size_t len = ap->authorized_macs_num * 6;
if (!len)
return;
if (len > out_len)
len = out_len;
memcpy(out, ap->authorized_macs, len);
}
ap: Handle most WSC IEs using ap_write_extra_ies Make the WSC IE processing and writing more self-contained (i.e. so that it can be more easily moved to a separate file if desired) by using the new ap_write_extra_ies() mechanism.
2021-03-11 14:07:24 +01:00
static size_t ap_get_wsc_ie_len(struct ap_state *ap,
enum mpdu_management_subtype type,
const struct mmpdu_header *client_frame,
size_t client_frame_len)
{
return 256;
}
static size_t ap_write_wsc_ie(struct ap_state *ap,
enum mpdu_management_subtype type,
const struct mmpdu_header *client_frame,
size_t client_frame_len,
uint8_t *out_buf)
{
uint8_t *wsc_data;
size_t wsc_data_size;
uint8_t *wsc_ie;
size_t wsc_ie_size;
size_t len = 0;
/* WSC IE */
ap: include PROBE_RESP attribute if required by driver Certain drivers support/require probe response offloading which IWD did not check for or properly handle. If probe response offloading is required the probe response frame watch will not be added and instead the ATTR_PROBE_RESP will be included with START_AP. The head/tail builders were reused but slightly modified to check if the probe request frame is NULL, since it will be for use with START_AP.
2022-09-28 21:48:52 +02:00
if (type == MPDU_MANAGEMENT_SUBTYPE_PROBE_RESPONSE && client_frame) {
ap: Fix an invalid access in ap_write_wsc_ie client_frame is not valid for a beacon frame as beacons are not sent in response to another frame. Move the access to client_frame->address_2 to the conditional blocks for Probe Response and Association Response frames.
2021-08-07 04:10:43 +02:00
const uint8_t *from = client_frame->address_2;
ap: Handle most WSC IEs using ap_write_extra_ies Make the WSC IE processing and writing more self-contained (i.e. so that it can be more easily moved to a separate file if desired) by using the new ap_write_extra_ies() mechanism.
2021-03-11 14:07:24 +01:00
struct wsc_probe_response wsc_pr = {};
const struct mmpdu_probe_request *req =
mmpdu_body(client_frame);
size_t req_ies_len = (void *) client_frame + client_frame_len -
(void *) req->ies;
ssize_t req_wsc_data_size;
/*
* Process the client Probe Request WSC IE first as it may
* cause us to exit "active PBC mode" and that will be
* immediately reflected in our Probe Response WSC IE.
*/
wsc_data = ie_tlv_extract_wsc_payload(req->ies, req_ies_len,
&req_wsc_data_size);
if (wsc_data) {
ap_process_wsc_probe_req(ap, from, wsc_data,
req_wsc_data_size);
l_free(wsc_data);
}
wsc_pr.version2 = true;
wsc_pr.state = WSC_STATE_CONFIGURED;
if (ap->wsc_pbc_timeout) {
wsc_pr.selected_registrar = true;
wsc_pr.device_password_id = ap->wsc_dpid;
wsc_pr.selected_reg_config_methods =
WSC_CONFIGURATION_METHOD_PUSH_BUTTON;
}
wsc_pr.response_type = WSC_RESPONSE_TYPE_AP;
memcpy(wsc_pr.uuid_e, ap->wsc_uuid_r, sizeof(wsc_pr.uuid_e));
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
wsc_pr.primary_device_type = ap->wsc_primary_device_type;
ap: Handle most WSC IEs using ap_write_extra_ies Make the WSC IE processing and writing more self-contained (i.e. so that it can be more easily moved to a separate file if desired) by using the new ap_write_extra_ies() mechanism.
2021-03-11 14:07:24 +01:00
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
if (ap->wsc_name[0] != '\0')
l_strlcpy(wsc_pr.device_name, ap->wsc_name,
ap: Handle most WSC IEs using ap_write_extra_ies Make the WSC IE processing and writing more self-contained (i.e. so that it can be more easily moved to a separate file if desired) by using the new ap_write_extra_ies() mechanism.
2021-03-11 14:07:24 +01:00
sizeof(wsc_pr.device_name));
wsc_pr.config_methods =
WSC_CONFIGURATION_METHOD_PUSH_BUTTON;
ap: Deduplicate writing authorized MACs into wsc_* structs
2021-04-28 17:48:37 +02:00
ap_write_authorized_macs(ap, sizeof(wsc_pr.authorized_macs),
wsc_pr.authorized_macs);
ap: Handle most WSC IEs using ap_write_extra_ies Make the WSC IE processing and writing more self-contained (i.e. so that it can be more easily moved to a separate file if desired) by using the new ap_write_extra_ies() mechanism.
2021-03-11 14:07:24 +01:00
wsc_data = wsc_build_probe_response(&wsc_pr, &wsc_data_size);
} else if (type == MPDU_MANAGEMENT_SUBTYPE_BEACON) {
struct wsc_beacon wsc_beacon = {};
wsc_beacon.version2 = true;
wsc_beacon.state = WSC_STATE_CONFIGURED;
if (ap->wsc_pbc_timeout) {
wsc_beacon.selected_registrar = true;
wsc_beacon.device_password_id = ap->wsc_dpid;
wsc_beacon.selected_reg_config_methods =
WSC_CONFIGURATION_METHOD_PUSH_BUTTON;
}
ap: Deduplicate writing authorized MACs into wsc_* structs
2021-04-28 17:48:37 +02:00
ap_write_authorized_macs(ap, sizeof(wsc_beacon.authorized_macs),
wsc_beacon.authorized_macs);
ap: Handle most WSC IEs using ap_write_extra_ies Make the WSC IE processing and writing more self-contained (i.e. so that it can be more easily moved to a separate file if desired) by using the new ap_write_extra_ies() mechanism.
2021-03-11 14:07:24 +01:00
wsc_data = wsc_build_beacon(&wsc_beacon, &wsc_data_size);
} else if (L_IN_SET(type, MPDU_MANAGEMENT_SUBTYPE_ASSOCIATION_RESPONSE,
MPDU_MANAGEMENT_SUBTYPE_REASSOCIATION_RESPONSE)) {
ap: Fix an invalid access in ap_write_wsc_ie client_frame is not valid for a beacon frame as beacons are not sent in response to another frame. Move the access to client_frame->address_2 to the conditional blocks for Probe Response and Association Response frames.
2021-08-07 04:10:43 +02:00
const uint8_t *from = client_frame->address_2;
ap: Handle most WSC IEs using ap_write_extra_ies Make the WSC IE processing and writing more self-contained (i.e. so that it can be more easily moved to a separate file if desired) by using the new ap_write_extra_ies() mechanism.
2021-03-11 14:07:24 +01:00
struct wsc_association_response wsc_resp = {};
struct sta_state *sta =
l_queue_find(ap->sta_states, ap_sta_match_addr, from);
if (!sta || sta->assoc_rsne)
return 0;
wsc_resp.response_type = WSC_RESPONSE_TYPE_AP;
wsc_resp.version2 = sta->wsc_v2;
wsc_data = wsc_build_association_response(&wsc_resp,
&wsc_data_size);
} else
return 0;
if (!wsc_data) {
l_error("wsc_build_<mgmt-subtype> error (stype 0x%x)", type);
return 0;
}
wsc_ie = ie_tlv_encapsulate_wsc_payload(wsc_data, wsc_data_size,
&wsc_ie_size);
l_free(wsc_data);
if (!wsc_ie) {
l_error("ie_tlv_encapsulate_wsc_payload error (stype 0x%x)",
type);
return 0;
}
memcpy(out_buf + len, wsc_ie, wsc_ie_size);
len += wsc_ie_size;
l_free(wsc_ie);
return len;
}
ap: make supported rates a common builder. The supported rates IE was being built in two places. This makes that code common. Unfortunately it needs to support both an ie builder and using a pointer directly which is why it only builds the contents of the IE and the caller must set the type/length.
2022-12-20 22:43:09 +01:00
static size_t ap_build_supported_rates(struct ap_state *ap,
uint8_t *rates)
{
uint32_t minr, maxr, count, r;
minr = l_uintset_find_min(ap->rates);
maxr = l_uintset_find_max(ap->rates);
count = 0;
for (r = minr; r <= maxr && count < 8; r++)
if (l_uintset_contains(ap->rates, r)) {
uint8_t flag = 0;
/* Mark only the lowest rate as Basic Rate */
if (count == 0)
flag = 0x80;
*rates++ = r | flag;
count++;
}
return count;
}
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
static size_t ap_get_extra_ies_len(struct ap_state *ap,
enum mpdu_management_subtype type,
const struct mmpdu_header *client_frame,
size_t client_frame_len)
{
size_t len = 0;
ap: Handle most WSC IEs using ap_write_extra_ies Make the WSC IE processing and writing more self-contained (i.e. so that it can be more easily moved to a separate file if desired) by using the new ap_write_extra_ies() mechanism.
2021-03-11 14:07:24 +01:00
len += ap_get_wsc_ie_len(ap, type, client_frame, client_frame_len);
ap: correct extra IEs length calculation for WMM IE The extra IE length for the WMM IE was being set to 26 which is the HT IE length, not WMM. Fix this and use the proper size for the WMM IE of 50 bytes. This shouldn't have caused any problems prior as the tail length is always allocated with 256 or 512 extra bytes of headroom.
2023-01-27 18:29:30 +01:00
/* WMM IE length */
ap: include WMM parameter IE The WMM parameter IE is expected by the linux kernel for any AP supporting HT/VHT etc. IWD won't actually use WMM and its not clear exactly why the kernel uses this restriction, but regardless it must be included to support HT.
2022-12-28 23:32:29 +01:00
if (ap->supports_ht)
ap: correct extra IEs length calculation for WMM IE The extra IE length for the WMM IE was being set to 26 which is the HT IE length, not WMM. Fix this and use the proper size for the WMM IE of 50 bytes. This shouldn't have caused any problems prior as the tail length is always allocated with 256 or 512 extra bytes of headroom.
2023-01-27 18:29:30 +01:00
len += 50;
ap: include WMM parameter IE The WMM parameter IE is expected by the linux kernel for any AP supporting HT/VHT etc. IWD won't actually use WMM and its not clear exactly why the kernel uses this restriction, but regardless it must be included to support HT.
2022-12-28 23:32:29 +01:00
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
if (ap->ops->get_extra_ies_len)
len += ap->ops->get_extra_ies_len(type, client_frame,
client_frame_len,
ap->user_data);
return len;
}
ap: include WMM parameter IE The WMM parameter IE is expected by the linux kernel for any AP supporting HT/VHT etc. IWD won't actually use WMM and its not clear exactly why the kernel uses this restriction, but regardless it must be included to support HT.
2022-12-28 23:32:29 +01:00
/* WMM Specification 2.2.2 WMM Parameter Element */
struct ap_wmm_ac_record {
#if defined(__LITTLE_ENDIAN_BITFIELD)
uint8_t aifsn : 4;
uint8_t acm : 1;
uint8_t aci : 2;
uint8_t reserved : 1;
uint8_t ecw_min : 4;
uint8_t ecw_max : 4;
#elif defined (__BIG_ENDIAN_BITFIELD)
uint8_t reserved : 1;
uint8_t aci : 2;
uint8_t acm : 1;
uint8_t aifsn : 4;
ap: Fix member naming on big endian
2023-01-23 16:45:22 +01:00
uint8_t ecw_max : 4;
uint8_t ecw_min : 4;
ap: include WMM parameter IE The WMM parameter IE is expected by the linux kernel for any AP supporting HT/VHT etc. IWD won't actually use WMM and its not clear exactly why the kernel uses this restriction, but regardless it must be included to support HT.
2022-12-28 23:32:29 +01:00
#else
#error "Please fix <asm/byteorder.h"
#endif
__le16 txop_limit;
} __attribute__((packed));
static size_t ap_write_wmm_ies(struct ap_state *ap, uint8_t *out_buf)
{
unsigned int i;
struct wiphy *wiphy = netdev_get_wiphy(ap->netdev);
/*
* Linux kernel requires APs include WMM Information element if
* supporting HT/VHT/etc.
*
* The only value we can actually get from the kernel is UAPSD. The
* remaining values (AC parameter records) are made up or defaults
* defined in the WMM spec are used.
*/
*out_buf++ = IE_TYPE_VENDOR_SPECIFIC;
*out_buf++ = 24;
memcpy(out_buf, microsoft_oui, sizeof(microsoft_oui));
out_buf += sizeof(microsoft_oui);
*out_buf++ = 2; /* WMM OUI Type */
*out_buf++ = 1; /* WMM Parameter subtype */
*out_buf++ = 1; /* WMM Version */
*out_buf++ = wiphy_supports_uapsd(wiphy) ? 1 << 7 : 0;
*out_buf++ = 0; /* reserved */
for (i = 0; i < 4; i++) {
struct ap_wmm_ac_record ac = { 0 };
ac.aifsn = 2;
ac.acm = 0;
ac.aci = i;
ac.ecw_min = 1;
ac.ecw_max = 15;
l_put_le16(0, &ac.txop_limit);
memcpy(out_buf + (i * 4), &ac, sizeof(struct ap_wmm_ac_record));
}
return 26;
}
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
static size_t ap_write_extra_ies(struct ap_state *ap,
enum mpdu_management_subtype type,
const struct mmpdu_header *client_frame,
size_t client_frame_len,
uint8_t *out_buf)
{
size_t len = 0;
ap: Handle most WSC IEs using ap_write_extra_ies Make the WSC IE processing and writing more self-contained (i.e. so that it can be more easily moved to a separate file if desired) by using the new ap_write_extra_ies() mechanism.
2021-03-11 14:07:24 +01:00
len += ap_write_wsc_ie(ap, type, client_frame, client_frame_len,
out_buf + len);
ap: include WMM parameter IE The WMM parameter IE is expected by the linux kernel for any AP supporting HT/VHT etc. IWD won't actually use WMM and its not clear exactly why the kernel uses this restriction, but regardless it must be included to support HT.
2022-12-28 23:32:29 +01:00
if (ap->supports_ht)
len += ap_write_wmm_ies(ap, out_buf + len);
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
if (ap->ops->write_extra_ies)
len += ap->ops->write_extra_ies(type,
client_frame, client_frame_len,
out_buf + len, ap->user_data);
return len;
}
ap: build HT Capabilities/Operation elements If supported this will include the HT capabilities and HT operations elements in beacons/probes. Some shortcuts were taken here since not all the information is currently parsed from the hardware. Namely the HT operation element does not include the basic MCS set. Still, this will at least show stations that the AP is capable of more than just basic rates. The builders themselves are structured similar to the basic rates builder where they build only the contents and return the length. The caller must set the type/length manually. This is to support the two use cases of using with an IE builder vs direct pointer.
2022-12-28 23:32:31 +01:00
static size_t ap_build_ht_capability(struct ap_state *ap, uint8_t *buf)
{
struct wiphy *wiphy = netdev_get_wiphy(ap->netdev);
size_t ht_capa_len;
const uint8_t *ht_capa = wiphy_get_ht_capabilities(wiphy, ap->band,
&ht_capa_len);
memcpy(buf, ht_capa, ht_capa_len);
return ht_capa_len;
}
static size_t ap_build_ht_operation(struct ap_state *ap, uint8_t *buf)
{
const struct l_queue_entry *e;
unsigned int non_ht = false;
unsigned int non_greenfield = false;
memset(buf, 0, 22);
*buf++ = ap->channel;
/*
* If 40MHz set 'Secondary Channel Offset' (bits 0-1) to above/below
* and set 'STA Channel Width' (bit 2) to indicate non-20Mhz.
*/
if (ap->chandef.channel_width == BAND_CHANDEF_WIDTH_20)
goto check_stas;
else if (ap->chandef.frequency < ap->chandef.center1_frequency)
*buf |= 1 & 0x3;
else
*buf |= 3 & 0x3;
*buf |= 1 << 2;
check_stas:
for (e = l_queue_get_entries(ap->sta_states); e; e = e->next) {
struct sta_state *sta = e->data;
if (!sta->associated)
continue;
if (!sta->ht_support)
non_ht = true;
else if (!sta->ht_greenfield)
non_greenfield = true;
}
if (non_greenfield)
set_bit(buf, 10);
if (non_ht)
set_bit(buf, 12);
/*
* TODO: Basic MCS set for all associated STAs
*/
return 22;
}
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
/*
* Build a Beacon frame or a Probe Response frame's header and body until
* the TIM IE. Except for the optional TIM IE which is inserted by the
* kernel when needed, our contents for both frames are the same.
* See Beacon format in 8.3.3.2 and Probe Response format in 8.3.3.10.
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
*
* 802.11-2016, Section 9.4.2.1:
* "The frame body components specified for many management subtypes result
* in elements ordered by ascending values of the Element ID field and then
* the Element ID Extension field (when present), with the exception of the
* MIC Management element (9.4.2.55)."
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
*/
static size_t ap_build_beacon_pr_head(struct ap_state *ap,
enum mpdu_management_subtype stype,
ap: update to use new TLV builder APIs AP still relies on the get_data/set_length semantics. Its more convenient to still use these since it avoids the need for extra temporary buffers when building the rates IE.
2019-04-23 19:47:26 +02:00
const uint8_t *dest, uint8_t *out_buf,
size_t out_len)
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
{
struct mmpdu_header *mpdu = (void *) out_buf;
uint16_t capability = IE_BSS_CAP_ESS | IE_BSS_CAP_PRIVACY;
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
const uint8_t *bssid = netdev_get_address(ap->netdev);
ap: make supported rates a common builder. The supported rates IE was being built in two places. This makes that code common. Unfortunately it needs to support both an ie builder and using a pointer directly which is why it only builds the contents of the IE and the caller must set the type/length.
2022-12-20 22:43:09 +01:00
size_t len;
ap: Use ie_tlv_builder in ap_build_beacon_pr_head
2017-09-30 04:28:07 +02:00
struct ie_tlv_builder builder;
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
memset(mpdu, 0, 36); /* Zero out header + non-IE fields */
/* Header */
mpdu->fc.protocol_version = 0;
mpdu->fc.type = MPDU_TYPE_MANAGEMENT;
mpdu->fc.subtype = stype;
memcpy(mpdu->address_1, dest, 6); /* DA */
memcpy(mpdu->address_2, bssid, 6); /* SA */
memcpy(mpdu->address_3, bssid, 6); /* BSSID */
/* Body non-IE fields */
l_put_le16(ap->beacon_interval, out_buf + 32); /* Beacon Interval */
l_put_le16(capability, out_buf + 34); /* Capability Info */
ap: Use ie_tlv_builder in ap_build_beacon_pr_head
2017-09-30 04:28:07 +02:00
ap: update to use new TLV builder APIs AP still relies on the get_data/set_length semantics. Its more convenient to still use these since it avoids the need for extra temporary buffers when building the rates IE.
2019-04-23 19:47:26 +02:00
ie_tlv_builder_init(&builder, out_buf + 36, out_len - 36);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
/* SSID IE */
ap: Use ie_tlv_builder in ap_build_beacon_pr_head
2017-09-30 04:28:07 +02:00
ie_tlv_builder_next(&builder, IE_TYPE_SSID);
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
ie_tlv_builder_set_data(&builder, ap->ssid, strlen(ap->ssid));
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
/* Supported Rates IE */
ap: Use ie_tlv_builder in ap_build_beacon_pr_head
2017-09-30 04:28:07 +02:00
ie_tlv_builder_next(&builder, IE_TYPE_SUPPORTED_RATES);
ap: make supported rates a common builder. The supported rates IE was being built in two places. This makes that code common. Unfortunately it needs to support both an ie builder and using a pointer directly which is why it only builds the contents of the IE and the caller must set the type/length.
2022-12-20 22:43:09 +01:00
len = ap_build_supported_rates(ap, ie_tlv_builder_get_data(&builder));
ie_tlv_builder_set_length(&builder, len);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
/* DSSS Parameter Set IE for DSSS, HR, ERP and HT PHY rates */
ap: Use ie_tlv_builder in ap_build_beacon_pr_head
2017-09-30 04:28:07 +02:00
ie_tlv_builder_next(&builder, IE_TYPE_DSSS_PARAMETER_SET);
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
ie_tlv_builder_set_data(&builder, &ap->channel, 1);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
ap: build HT Capabilities/Operation elements If supported this will include the HT capabilities and HT operations elements in beacons/probes. Some shortcuts were taken here since not all the information is currently parsed from the hardware. Namely the HT operation element does not include the basic MCS set. Still, this will at least show stations that the AP is capable of more than just basic rates. The builders themselves are structured similar to the basic rates builder where they build only the contents and return the length. The caller must set the type/length manually. This is to support the two use cases of using with an IE builder vs direct pointer.
2022-12-28 23:32:31 +01:00
if (ap->supports_ht) {
ie_tlv_builder_next(&builder, IE_TYPE_HT_CAPABILITIES);
len = ap_build_ht_capability(ap,
ie_tlv_builder_get_data(&builder));
ie_tlv_builder_set_length(&builder, len);
ie_tlv_builder_next(&builder, IE_TYPE_HT_OPERATION);
len = ap_build_ht_operation(ap,
ie_tlv_builder_get_data(&builder));
ie_tlv_builder_set_length(&builder, len);
}
ie: Change ie_tlv_builder_finalize signature ie_tlv_builder_init takes a size_t as input, yet for some reason ie_tlv_builder_finalize takes an unsigned int argument as output. Fix the latter to use size_t as well.
2021-07-09 21:31:27 +02:00
ie_tlv_builder_finalize(&builder, &out_len);
return 36 + out_len;
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
}
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
static size_t ap_build_country_ie(struct ap_state *ap, uint8_t *out_buf,
size_t buf_len)
{
size_t len;
size_t i;
int spacing;
uint8_t *pos = out_buf;
uint8_t nchans = 1;
struct wiphy *wiphy = netdev_get_wiphy(ap->netdev);
const struct band_freq_attrs *last = NULL;
const struct band_freq_attrs *list = wiphy_get_frequency_info_list(
wiphy, ap->band, &len);
if (!list || wiphy_country_is_unknown(wiphy))
return 0;
if (L_WARN_ON(buf_len < 5))
goto no_space;
*pos++ = IE_TYPE_COUNTRY;
/* length not yet known */
pos++;
wiphy_get_reg_domain_country(wiphy, (char *)pos);
pos += 2;
*pos++ = ' ';
buf_len -= 5;
if (ap->band == BAND_FREQ_2_4_GHZ)
spacing = 1;
else
spacing = 4;
/*
* Construct a list of subband triplet entries. Each entry contains a
* starting channel and a number of channels which are spaced evenly
* and use the same TX power. Any deviation from this results in a new
* channel group.
*
tree-wide: correct the spelling Ghz -> GHz This mispelling was present in the configuration, so I retained parsing of the legacy BandModifier*Ghz options for compatibility. Without this change anyone spelling GHz correctly in their configs would be very confused.
2023-11-08 02:33:54 +01:00
* TODO: 6GHz requires operating triplets, not subband triplets.
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
*/
for (i = 0; i < len; i++) {
const struct band_freq_attrs *attr = &list[i];
if (!attr->supported || attr->disabled)
continue;
if (!last) {
/* Room for one complete triplet */
if (L_WARN_ON(buf_len < 3))
goto no_space;
*pos++ = i;
last = attr;
continue;
}
if (spacing != attr - last ||
attr->tx_power != last->tx_power) {
/* finish current group */
*pos++ = nchans;
*pos++ = last->tx_power;
buf_len -= 3;
/* start a new group */
if (L_WARN_ON(buf_len < 3))
goto no_space;
*pos++ = i;
nchans = 1;
} else
nchans++;
last = attr;
}
/* finish final group */
ap: check that the last band_freq_attrs was set Static analysis complains that 'last' could be NULL which is true. This really could only happen if every frequency was disabled which likely is impossible but in any case, check before dereferencing the pointer.
2024-02-29 19:12:13 +01:00
if (last) {
*pos++ = nchans;
*pos++ = last->tx_power;
}
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
len = pos - out_buf - 2;
/* Pad to even byte */
if (len & 1) {
if (L_WARN_ON(buf_len < 1))
goto no_space;
*pos++ = 0;
len++;
}
out_buf[1] = len;
return out_buf[1] + 2;
no_space:
return 0;
}
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
/* Beacon / Probe Response frame portion after the TIM IE */
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
static size_t ap_build_beacon_pr_tail(struct ap_state *ap,
enum mpdu_management_subtype stype,
const struct mmpdu_header *req,
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
size_t req_len, uint8_t *out_buf,
size_t buf_len)
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
{
size_t len;
struct ie_rsn_info rsn;
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
len = ap_build_country_ie(ap, out_buf, buf_len);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
/* RSNE */
ap_set_rsn_info(ap, &rsn);
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
if (!ie_build_rsne(&rsn, out_buf + len))
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
return 0;
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
len += 2 + out_buf[len + 1];
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
len += ap_write_extra_ies(ap, stype, req, req_len, out_buf + len);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
return len;
}
ap: Push Button mode API and beacon changes This adds the API for putting the AP in Push Button mode, which we'll need to P2P GO side but may be useful on its own too. A WSC IE is added to our beacons and probe responses indicating whether the PBC mode is active.
2020-08-28 14:46:46 +02:00
static void ap_set_beacon_cb(struct l_genl_msg *msg, void *user_data)
{
int error = l_genl_msg_get_error(msg);
if (error < 0)
l_error("SET_BEACON failed: %s (%i)", strerror(-error), -error);
}
ap: Make ap_update_beacon public Let users call ap_update_beacon when a value has changed which should be reflected in the beacon IEs.
2021-03-10 23:06:30 +01:00
void ap_update_beacon(struct ap_state *ap)
ap: Push Button mode API and beacon changes This adds the API for putting the AP in Push Button mode, which we'll need to P2P GO side but may be useful on its own too. A WSC IE is added to our beacons and probe responses indicating whether the PBC mode is active.
2020-08-28 14:46:46 +02:00
{
struct l_genl_msg *cmd;
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
uint8_t head[256];
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
size_t tail_len = 256 + ap_get_extra_ies_len(ap,
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
MPDU_MANAGEMENT_SUBTYPE_BEACON,
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
NULL, 0);
L_AUTO_FREE_VAR(uint8_t *, tail) = malloc(tail_len);
size_t head_len;
ap: Push Button mode API and beacon changes This adds the API for putting the AP in Push Button mode, which we'll need to P2P GO side but may be useful on its own too. A WSC IE is added to our beacons and probe responses indicating whether the PBC mode is active.
2020-08-28 14:46:46 +02:00
uint64_t wdev_id = netdev_get_wdev_id(ap->netdev);
static const uint8_t bcast_addr[6] = {
0xff, 0xff, 0xff, 0xff, 0xff, 0xff
};
ap: Make ap_update_beacon public Let users call ap_update_beacon when a value has changed which should be reflected in the beacon IEs.
2021-03-10 23:06:30 +01:00
if (L_WARN_ON(!ap->started))
return;
ap: Push Button mode API and beacon changes This adds the API for putting the AP in Push Button mode, which we'll need to P2P GO side but may be useful on its own too. A WSC IE is added to our beacons and probe responses indicating whether the PBC mode is active.
2020-08-28 14:46:46 +02:00
head_len = ap_build_beacon_pr_head(ap, MPDU_MANAGEMENT_SUBTYPE_BEACON,
bcast_addr, head, sizeof(head));
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
tail_len = ap_build_beacon_pr_tail(ap, MPDU_MANAGEMENT_SUBTYPE_BEACON,
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
NULL, 0, tail, tail_len);
ap: Push Button mode API and beacon changes This adds the API for putting the AP in Push Button mode, which we'll need to P2P GO side but may be useful on its own too. A WSC IE is added to our beacons and probe responses indicating whether the PBC mode is active.
2020-08-28 14:46:46 +02:00
if (L_WARN_ON(!head_len || !tail_len))
return;
cmd = l_genl_msg_new_sized(NL80211_CMD_SET_BEACON,
32 + head_len + tail_len);
l_genl_msg_append_attr(cmd, NL80211_ATTR_WDEV, 8, &wdev_id);
l_genl_msg_append_attr(cmd, NL80211_ATTR_BEACON_HEAD, head_len, head);
l_genl_msg_append_attr(cmd, NL80211_ATTR_BEACON_TAIL, tail_len, tail);
l_genl_msg_append_attr(cmd, NL80211_ATTR_IE, 0, "");
l_genl_msg_append_attr(cmd, NL80211_ATTR_IE_PROBE_RESP, 0, "");
l_genl_msg_append_attr(cmd, NL80211_ATTR_IE_ASSOC_RESP, 0, "");
if (l_genl_family_send(ap->nl80211, cmd, ap_set_beacon_cb, NULL, NULL))
return;
l_genl_msg_unref(cmd);
l_error("Issuing SET_BEACON failed");
}
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
static uint32_t ap_send_mgmt_frame(struct ap_state *ap,
const struct mmpdu_header *frame,
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
size_t frame_len,
frame_xchg_cb_t callback,
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
void *user_data)
{
ap: add support for 5GHz frequencies in AP mode This enables IWD to use 5GHz frequencies in AP mode. Currently 6GHz is not supported so we can assume a [General].Channel value 36 or above indicates the 5GHz band. It should be noted that the system will probably need a regulatory domain set in order for 5GHz to be allowed in AP mode. This is due to world roaming (00) restricting any/all 5GHz frequencies. This can be accomplished by setting main.conf [General].Country=CC to the country this AP will operate in.
2022-12-09 20:39:28 +01:00
uint32_t ch_freq = band_channel_to_freq(ap->channel, ap->band);
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
uint64_t wdev_id = netdev_get_wdev_id(ap->netdev);
struct iovec iov[2];
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
iov[0].iov_base = (void *) frame;
iov[0].iov_len = frame_len;
iov[1].iov_base = NULL;
return frame_xchg_start(wdev_id, iov, ch_freq, 0, 0, 0, 0,
callback, user_data, NULL, NULL);
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
}
ap: Implement P2P GO-side 4-way handshake IP Allocation Use the struct handshake_state::support_ip_allocation field already supported in eapol.c authenticator side to enable the P2P IP Allocation mechanism in ap.c. Add the P2P_GROUP_CAP_IP_ALLOCATION bit in P2P group capabilities to signal the feature is now supported. There's no harm in enabling this feature in every AP (not just P2P Group Owner) but the clients won't know whether we support it other than through that P2P-specific group capability bit.
2021-08-23 16:14:23 +02:00
#define IP4_FROM_STR(str) \
(__extension__ ({ \
struct in_addr ia; \
inet_pton(AF_INET, str, &ia) == 1 ? ia.s_addr : 0; \
}))
ap: Fix handshake state gtk not being set handshake_state_set_authenticator_ie must be called to set group_cipher in struct handshake_shake before handshake_set_gtk_state, otherwise handshake_set_gtk_state is unable to determine the key length to set handshake state gtk. Fixes: 4bc20a097965 ("ap: Start EAP-WSC authentication with WSC enrollees")
2020-11-14 15:21:45 +01:00
static void ap_start_handshake(struct sta_state *sta, bool use_eapol_start,
const uint8_t *gtk_rsc)
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
{
struct ap_state *ap = sta->ap;
const uint8_t *own_addr = netdev_get_address(ap->netdev);
struct ie_rsn_info rsn;
eapol,ap: Remove assumption of single cipher in authenticator IE Allow the user of the eapol_sm & handshake_state APIs to have multiple pairwise ciphers listed in the authenticator IE.
2021-02-01 14:40:42 +01:00
uint8_t bss_rsne[64];
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
handshake_state_set_ssid(sta->hs, (void *) ap->ssid, strlen(ap->ssid));
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
handshake_state_set_authenticator_address(sta->hs, own_addr);
handshake_state_set_supplicant_address(sta->hs, sta->addr);
ap_set_rsn_info(ap, &rsn);
/*
* Note: This assumes the length that ap_set_rsn_info() requires. If
* ap_set_rsn_info() changes then this will need to be updated.
*/
ie_build_rsne(&rsn, bss_rsne);
handshake_state_set_authenticator_ie(sta->hs, bss_rsne);
ap: Fix handshake state gtk not being set handshake_state_set_authenticator_ie must be called to set group_cipher in struct handshake_shake before handshake_set_gtk_state, otherwise handshake_set_gtk_state is unable to determine the key length to set handshake state gtk. Fixes: 4bc20a097965 ("ap: Start EAP-WSC authentication with WSC enrollees")
2020-11-14 15:21:45 +01:00
if (gtk_rsc)
handshake_state_set_gtk(sta->hs, sta->ap->gtk,
sta->ap->gtk_index, gtk_rsc);
ap: Implement P2P GO-side 4-way handshake IP Allocation Use the struct handshake_state::support_ip_allocation field already supported in eapol.c authenticator side to enable the P2P IP Allocation mechanism in ap.c. Add the P2P_GROUP_CAP_IP_ALLOCATION bit in P2P group capabilities to signal the feature is now supported. There's no harm in enabling this feature in every AP (not just P2P Group Owner) but the clients won't know whether we support it other than through that P2P-specific group capability bit.
2021-08-23 16:14:23 +02:00
if (ap->netconfig_dhcp)
sta->hs->support_ip_allocation = true;
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
sta->sm = eapol_sm_new(sta->hs);
if (!sta->sm) {
ap_stop_handshake(sta);
l_error("could not create sm object");
goto error;
}
eapol_sm_set_listen_interval(sta->sm, sta->listen_interval);
eapol_sm_set_use_eapol_start(sta->sm, use_eapol_start);
eapol_register(sta->sm);
eapol_start(sta->sm);
return;
error:
ap_del_station(sta, MMPDU_REASON_CODE_UNSPECIFIED, true);
}
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
static bool ap_sta_get_dhcp4_lease(struct sta_state *sta)
{
if (sta->ip_alloc_lease)
return true;
if (!sta->ap->netconfig_dhcp)
return false;
sta->ip_alloc_lease = l_dhcp_server_discover(sta->ap->netconfig_dhcp,
0, NULL, sta->addr);
if (!sta->ip_alloc_lease) {
l_error("l_dhcp_server_discover() failed, see IWD_DHCP_DEBUG "
"output");
return false;
}
return true;
}
ap: remove ap_error_deauth_sta Instead of manually sending a deauth/disassociate to a station during an error or removal, the kernel can do it automatically with DEL_STATION by including the MGMT_SUBTYPE attribute. This removes the need for ap_error_deauth_sta and introduces ap_deauthenticate_sta. Now AP can be explicit when it chooses to deauth or disassociate.
2018-06-28 18:12:22 +02:00
static void ap_handshake_event(struct handshake_state *hs,
handshake: Convert handshake event callbacks variadic functions Convert the handshake event callback type to use variable argument list to allow for more flexibility in event-specific arguments passed to the callbacks. Note the uint16_t reason code is promoted to an int when using variable arguments so va_arg(args, int) has to be used.
2019-10-28 15:04:57 +01:00
enum handshake_event event, void *user_data, ...)
ap: remove ap_error_deauth_sta Instead of manually sending a deauth/disassociate to a station during an error or removal, the kernel can do it automatically with DEL_STATION by including the MGMT_SUBTYPE attribute. This removes the need for ap_error_deauth_sta and introduces ap_deauthenticate_sta. Now AP can be explicit when it chooses to deauth or disassociate.
2018-06-28 18:12:22 +02:00
{
struct sta_state *sta = user_data;
ap: Implement P2P GO-side 4-way handshake IP Allocation Use the struct handshake_state::support_ip_allocation field already supported in eapol.c authenticator side to enable the P2P IP Allocation mechanism in ap.c. Add the P2P_GROUP_CAP_IP_ALLOCATION bit in P2P group capabilities to signal the feature is now supported. There's no harm in enabling this feature in every AP (not just P2P Group Owner) but the clients won't know whether we support it other than through that P2P-specific group capability bit.
2021-08-23 16:14:23 +02:00
struct ap_state *ap = sta->ap;
handshake: Convert handshake event callbacks variadic functions Convert the handshake event callback type to use variable argument list to allow for more flexibility in event-specific arguments passed to the callbacks. Note the uint16_t reason code is promoted to an int when using variable arguments so va_arg(args, int) has to be used.
2019-10-28 15:04:57 +01:00
va_list args;
va_start(args, user_data);
ap: removed code that was moved to eapol All handshake packet handling has been removed from ap and moved to eapol. After association, the AP registers a new authenticator state machine which handles the AP side of the handshake. AP will receive a handshake event once the 4-way handshake is complete.
2018-06-27 19:06:31 +02:00
ap: remove ap_error_deauth_sta Instead of manually sending a deauth/disassociate to a station during an error or removal, the kernel can do it automatically with DEL_STATION by including the MGMT_SUBTYPE attribute. This removes the need for ap_error_deauth_sta and introduces ap_deauthenticate_sta. Now AP can be explicit when it chooses to deauth or disassociate.
2018-06-28 18:12:22 +02:00
switch (event) {
case HANDSHAKE_EVENT_COMPLETE:
ap: Implement P2P GO-side 4-way handshake IP Allocation Use the struct handshake_state::support_ip_allocation field already supported in eapol.c authenticator side to enable the P2P IP Allocation mechanism in ap.c. Add the P2P_GROUP_CAP_IP_ALLOCATION bit in P2P group capabilities to signal the feature is now supported. There's no harm in enabling this feature in every AP (not just P2P Group Owner) but the clients won't know whether we support it other than through that P2P-specific group capability bit.
2021-08-23 16:14:23 +02:00
if (sta->ip_alloc_lease) {
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
if (hs->support_ip_allocation)
sta->ip_alloc_sent = true;
ap: Implement P2P GO-side 4-way handshake IP Allocation Use the struct handshake_state::support_ip_allocation field already supported in eapol.c authenticator side to enable the P2P IP Allocation mechanism in ap.c. Add the P2P_GROUP_CAP_IP_ALLOCATION bit in P2P group capabilities to signal the feature is now supported. There's no harm in enabling this feature in every AP (not just P2P Group Owner) but the clients won't know whether we support it other than through that P2P-specific group capability bit.
2021-08-23 16:14:23 +02:00
/*
* Move the lease from offered to active state if the
* client has actually used it. In any case drop our
* reference to the lease, the server owns the lease
* and if we want to keep our reference we'd need to
* react to relevant server events.
*/
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
if (sta->ip_alloc_sent)
ap: Implement P2P GO-side 4-way handshake IP Allocation Use the struct handshake_state::support_ip_allocation field already supported in eapol.c authenticator side to enable the P2P IP Allocation mechanism in ap.c. Add the P2P_GROUP_CAP_IP_ALLOCATION bit in P2P group capabilities to signal the feature is now supported. There's no harm in enabling this feature in every AP (not just P2P Group Owner) but the clients won't know whether we support it other than through that P2P-specific group capability bit.
2021-08-23 16:14:23 +02:00
l_dhcp_server_request(ap->netconfig_dhcp,
sta->ip_alloc_lease);
sta->ip_alloc_lease = NULL;
}
ap: remove ap_error_deauth_sta Instead of manually sending a deauth/disassociate to a station during an error or removal, the kernel can do it automatically with DEL_STATION by including the MGMT_SUBTYPE attribute. This removes the need for ap_error_deauth_sta and introduces ap_deauthenticate_sta. Now AP can be explicit when it chooses to deauth or disassociate.
2018-06-28 18:12:22 +02:00
ap_new_rsna(sta);
break;
case HANDSHAKE_EVENT_FAILED:
handshake: Convert handshake event callbacks variadic functions Convert the handshake event callback type to use variable argument list to allow for more flexibility in event-specific arguments passed to the callbacks. Note the uint16_t reason code is promoted to an int when using variable arguments so va_arg(args, int) has to be used.
2019-10-28 15:04:57 +01:00
netdev_handshake_failed(hs, va_arg(args, int));
netdev: fixed key setting failure If netdev fails to set the keys, there was no way for device/ap to know. A new handshake event was added for this. The key setting failure function was also fixed to support both AP/station iftypes. It will now automatically send either a disconnect or del_station depending on the interface type. In similar manner, netdev_handshake_failed was also modified to support both AP/station iftypes. Now, any handshake event listeners should call netdev_handshake_failed upon a handshake failure event, including AP.
2018-07-03 23:36:33 +02:00
/* fall through */
case HANDSHAKE_EVENT_SETTING_KEYS_FAILED:
ap: eapol_sm lifecycle fixes On one hand when we're called with HANDSHAKE_EVENT_FAILED or HANDSHAKE_EVENT_SETTING_KEYS_FAILED the eapol_sm will be freed in eapol.c, fix a double-free by setting it to NULL before ap_free_sta is called. On the other hand make sure we call eapol_sm_free before setting sta->sm to NULL in ap_drop_rsna to avoid potential leak and avoid the eapol_sm continuing to use the handshake_state we freed.
2018-08-24 03:37:53 +02:00
sta->sm = NULL;
netdev: fixed key setting failure If netdev fails to set the keys, there was no way for device/ap to know. A new handshake event was added for this. The key setting failure function was also fixed to support both AP/station iftypes. It will now automatically send either a disconnect or del_station depending on the interface type. In similar manner, netdev_handshake_failed was also modified to support both AP/station iftypes. Now, any handshake event listeners should call netdev_handshake_failed upon a handshake failure event, including AP.
2018-07-03 23:36:33 +02:00
ap_remove_sta(sta);
ap: Implement P2P GO-side 4-way handshake IP Allocation Use the struct handshake_state::support_ip_allocation field already supported in eapol.c authenticator side to enable the P2P IP Allocation mechanism in ap.c. Add the P2P_GROUP_CAP_IP_ALLOCATION bit in P2P group capabilities to signal the feature is now supported. There's no harm in enabling this feature in every AP (not just P2P Group Owner) but the clients won't know whether we support it other than through that P2P-specific group capability bit.
2021-08-23 16:14:23 +02:00
break;
case HANDSHAKE_EVENT_P2P_IP_REQUEST:
{
char own_addr_str[INET_ADDRSTRLEN];
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
if (!ap_sta_get_dhcp4_lease(sta))
ap: Implement P2P GO-side 4-way handshake IP Allocation Use the struct handshake_state::support_ip_allocation field already supported in eapol.c authenticator side to enable the P2P IP Allocation mechanism in ap.c. Add the P2P_GROUP_CAP_IP_ALLOCATION bit in P2P group capabilities to signal the feature is now supported. There's no harm in enabling this feature in every AP (not just P2P Group Owner) but the clients won't know whether we support it other than through that P2P-specific group capability bit.
2021-08-23 16:14:23 +02:00
break;
ap: Use _u32 dhcp_lease getters
2021-10-14 00:38:44 +02:00
sta->hs->client_ip_addr =
l_dhcp_lease_get_address_u32(sta->ip_alloc_lease);
sta->hs->subnet_mask =
l_dhcp_lease_get_netmask_u32(sta->ip_alloc_lease);
ap: Implement P2P GO-side 4-way handshake IP Allocation Use the struct handshake_state::support_ip_allocation field already supported in eapol.c authenticator side to enable the P2P IP Allocation mechanism in ap.c. Add the P2P_GROUP_CAP_IP_ALLOCATION bit in P2P group capabilities to signal the feature is now supported. There's no harm in enabling this feature in every AP (not just P2P Group Owner) but the clients won't know whether we support it other than through that P2P-specific group capability bit.
2021-08-23 16:14:23 +02:00
l_rtnl_address_get_address(ap->netconfig_addr4, own_addr_str);
sta->hs->go_ip_addr = IP4_FROM_STR(own_addr_str);
break;
}
ap: support PTK rekeys This adds support for rekeys to AP mode. A single timer is used and reset to the next station needing a rekey. A default rekey timer of 600 seconds is used unless the profile sets a timeout.
2023-01-18 22:18:00 +01:00
case HANDSHAKE_EVENT_REKEY_COMPLETE:
ap_set_sta_rekey_timer(ap, sta);
return;
ap: remove ap_error_deauth_sta Instead of manually sending a deauth/disassociate to a station during an error or removal, the kernel can do it automatically with DEL_STATION by including the MGMT_SUBTYPE attribute. This removes the need for ap_error_deauth_sta and introduces ap_deauthenticate_sta. Now AP can be explicit when it chooses to deauth or disassociate.
2018-06-28 18:12:22 +02:00
default:
break;
}
handshake: Convert handshake event callbacks variadic functions Convert the handshake event callback type to use variable argument list to allow for more flexibility in event-specific arguments passed to the callbacks. Note the uint16_t reason code is promoted to an int when using variable arguments so va_arg(args, int) has to be used.
2019-10-28 15:04:57 +01:00
va_end(args);
ap: removed code that was moved to eapol All handshake packet handling has been removed from ap and moved to eapol. After association, the AP registers a new authenticator state machine which handles the AP side of the handshake. AP will receive a handshake event once the 4-way handshake is complete.
2018-06-27 19:06:31 +02:00
}
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
static void ap_start_rsna(struct sta_state *sta, const uint8_t *gtk_rsc)
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
{
ap: removed code that was moved to eapol All handshake packet handling has been removed from ap and moved to eapol. After association, the AP registers a new authenticator state machine which handles the AP side of the handshake. AP will receive a handshake event once the 4-way handshake is complete.
2018-06-27 19:06:31 +02:00
/* this handshake setup assumes PSK network */
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
sta->hs = netdev_handshake_state_new(sta->ap->netdev);
handshake: Rename own_ie/ap_ie and related setters To avoid confusion in case of an authenticator side handshake_state structure and eapol_sm structure, rename own_ie to supplicant_ie and ap_ie to authenticator_ie. Also rename handshake_state_set_{own,ap}_{rsn,wpa} and fix when we call handshake_state_setup_own_ciphers. As a result handshake_state_set_authenticator, if needed, should be called before handshake_state_set_{own,ap}_{rsn,wpa}.
2018-08-25 03:54:24 +02:00
handshake_state_set_authenticator(sta->hs, true);
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
handshake_state_set_event_func(sta->hs, ap_handshake_event, sta);
handshake: simplify IE setters The handshake object had 4 setters for authenticator/supplicant IE. Since the IE ultimately gets put into the same buffer, there really only needs to be a single setter for authenticator/supplicant. The handshake object can deal with parsing to decide what kind of IE it is (WPA or RSN).
2019-06-07 21:17:00 +02:00
handshake_state_set_supplicant_ie(sta->hs, sta->assoc_rsne);
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
handshake_state_set_pmk(sta->hs, sta->ap->psk, 32);
ap: Fix handshake state gtk not being set handshake_state_set_authenticator_ie must be called to set group_cipher in struct handshake_shake before handshake_set_gtk_state, otherwise handshake_set_gtk_state is unable to determine the key length to set handshake state gtk. Fixes: 4bc20a097965 ("ap: Start EAP-WSC authentication with WSC enrollees")
2020-11-14 15:21:45 +01:00
ap_start_handshake(sta, false, gtk_rsc);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
}
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
static void ap_gtk_query_cb(struct l_genl_msg *msg, void *user_data)
{
struct sta_state *sta = user_data;
nl80211: added verifier/parser for GET_KEY AdHoc will also need the same functionality to verify and parse the key sequence from GET_KEY. This block of code was moved from AP's GET_KEY callback into nl80211_parse_get_key_seq.
2018-10-08 22:44:11 +02:00
const void *gtk_rsc;
ap: handle GET_KEY error setting RSC to zero Rather than the previous hack which disabled group traffic it was found that the GTK RSC could be manually set to zero which allows group traffic. This appears to fix AP mode on brcmfmac along with the previous fixes. This is not documented in nl80211, but appears to work with this driver.
2021-01-29 19:47:25 +01:00
uint8_t zero_gtk_rsc[6];
ap: handle -ENOTSUP when getting the GTK Some hardware does not support this, so use a zero RSC in this case rather than failing the association.
2023-02-24 19:51:47 +01:00
int err;
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
sta->gtk_query_cmd_id = 0;
ap: handle -ENOTSUP when getting the GTK Some hardware does not support this, so use a zero RSC in this case rather than failing the association.
2023-02-24 19:51:47 +01:00
err = l_genl_msg_get_error(msg);
if (err == -ENOTSUP)
goto zero_rsc;
else if (err < 0)
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
goto error;
ap: allow no group traffic on GET_KEY error Some fullmac cards were found to be buggy with getting the GTK where it returns a BIP key for the GTK index, even after creating a GTK with NEW_KEY explicitly. In an effort to get these cards semi-working we can treat this just as a warning and continue with the handshake without a GTK set which disables group traffic. A warning is printed in this case so the user is not completely in the dark.
2021-01-28 19:33:52 +01:00
gtk_rsc = nl80211_parse_get_key_seq(msg);
if (!gtk_rsc) {
ap: handle -ENOTSUP when getting the GTK Some hardware does not support this, so use a zero RSC in this case rather than failing the association.
2023-02-24 19:51:47 +01:00
zero_rsc:
ap: handle GET_KEY error setting RSC to zero Rather than the previous hack which disabled group traffic it was found that the GTK RSC could be manually set to zero which allows group traffic. This appears to fix AP mode on brcmfmac along with the previous fixes. This is not documented in nl80211, but appears to work with this driver.
2021-01-29 19:47:25 +01:00
memset(zero_gtk_rsc, 0, 6);
gtk_rsc = zero_gtk_rsc;
ap: allow no group traffic on GET_KEY error Some fullmac cards were found to be buggy with getting the GTK where it returns a BIP key for the GTK index, even after creating a GTK with NEW_KEY explicitly. In an effort to get these cards semi-working we can treat this just as a warning and continue with the handshake without a GTK set which disables group traffic. A warning is printed in this case so the user is not completely in the dark.
2021-01-28 19:33:52 +01:00
}
nl80211: added verifier/parser for GET_KEY AdHoc will also need the same functionality to verify and parse the key sequence from GET_KEY. This block of code was moved from AP's GET_KEY callback into nl80211_parse_get_key_seq.
2018-10-08 22:44:11 +02:00
ap_start_rsna(sta, gtk_rsc);
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
return;
error:
ap_del_station(sta, MMPDU_REASON_CODE_UNSPECIFIED, true);
}
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
static void ap_stop_handshake_schedule(struct sta_state *sta)
{
if (sta->stop_handshake_work)
return;
sta->stop_handshake_work = l_idle_create(ap_stop_handshake_work,
sta, NULL);
}
static void ap_wsc_handshake_event(struct handshake_state *hs,
enum handshake_event event, void *user_data, ...)
{
struct sta_state *sta = user_data;
va_list args;
struct ap_event_registration_success_data event_data;
struct ap_pbc_record_expiry_data expiry_data;
va_start(args, user_data);
switch (event) {
case HANDSHAKE_EVENT_FAILED:
sta->sm = NULL;
ap_stop_handshake_schedule(sta);
/*
* Some diagrams in WSC v2.0.5 indicate we should
* automatically deauthenticate the Enrollee. The text
* generally indicates the Enrollee may disassociate
* meaning that we should neither deauthenticate nor
* disassociate it automatically. Some places indicate
* that the enrollee can send a new EAPoL-Start right away
* on an unsuccessful registration, we don't implement
* this for now. STA remains associated but not authorized
* and basically has no other option than to re-associate
* or disassociate/deauthenticate.
*/
break;
case HANDSHAKE_EVENT_EAP_NOTIFY:
if (va_arg(args, unsigned int) != EAP_WSC_EVENT_CREDENTIAL_SENT)
break;
/*
* WSC v2.0.5 Section 11.3:
* "If the Registrar successfully runs the PBC method to
* completion with an Enrollee, that Enrollee's probe requests
* are removed from the Monitor Time check the next time the
* Registrar's PBC button is pressed."
*/
expiry_data.min_time = 0;
expiry_data.mac = sta->addr;
l_queue_foreach_remove(sta->ap->wsc_pbc_probes,
ap_wsc_pbc_record_expire,
&expiry_data);
event_data.mac = sta->addr;
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
ap_event(sta->ap, AP_EVENT_REGISTRATION_SUCCESS, &event_data);
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
break;
default:
break;
}
va_end(args);
}
static void ap_start_eap_wsc(struct sta_state *sta)
{
struct ap_state *ap = sta->ap;
/*
* WSC v2.0.5 Section 8.2: "The AP is allowed to send
* EAP-Request/Identity to the station before EAPOL-Start is received
* if a WSC IE is included in the (re)association request and the
* WSC IE is version 2.0 or higher.
*/
bool wait_for_eapol_start = !sta->wsc_v2;
sta->wsc_settings = l_settings_new();
l_settings_set_string(sta->wsc_settings, "Security", "EAP-Method",
"WSC-R");
l_settings_set_string(sta->wsc_settings, "WSC", "EnrolleeMAC",
util_address_to_string(sta->addr));
treewide: Use l_settings_{set,get}_bytes
2020-09-16 11:17:54 +02:00
l_settings_set_bytes(sta->wsc_settings, "WSC", "UUID-R",
ap->wsc_uuid_r, 16);
l_settings_set_bytes(sta->wsc_settings, "WSC", "UUID-E",
sta->wsc_uuid_e, 16);
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
l_settings_set_uint(sta->wsc_settings, "WSC", "RFBand",
WSC_RF_BAND_2_4_GHZ);
l_settings_set_uint(sta->wsc_settings, "WSC", "ConfigurationMethods",
WSC_CONFIGURATION_METHOD_PUSH_BUTTON);
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
l_settings_set_string(sta->wsc_settings, "WSC", "WPA2-SSID", ap->ssid);
ap: Support working without passphrase Add a "psk" setting to allow the user to pass the binary PSK directly instead of generating it from the passphrase and the SSID. In that case we'll only send the PSK to WSC enrollees.
2020-09-16 11:17:50 +02:00
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
if (ap->passphrase[0])
ap: Support working without passphrase Add a "psk" setting to allow the user to pass the binary PSK directly instead of generating it from the passphrase and the SSID. In that case we'll only send the PSK to WSC enrollees.
2020-09-16 11:17:50 +02:00
l_settings_set_string(sta->wsc_settings,
"WSC", "WPA2-Passphrase",
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
ap->passphrase);
ap: Support working without passphrase Add a "psk" setting to allow the user to pass the binary PSK directly instead of generating it from the passphrase and the SSID. In that case we'll only send the PSK to WSC enrollees.
2020-09-16 11:17:50 +02:00
else
l_settings_set_bytes(sta->wsc_settings,
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
"WSC", "WPA2-PSK", ap->psk, 32);
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
sta->hs = netdev_handshake_state_new(ap->netdev);
handshake_state_set_authenticator(sta->hs, true);
handshake_state_set_event_func(sta->hs, ap_wsc_handshake_event, sta);
handshake_state_set_8021x_config(sta->hs, sta->wsc_settings);
ap: Fix handshake state gtk not being set handshake_state_set_authenticator_ie must be called to set group_cipher in struct handshake_shake before handshake_set_gtk_state, otherwise handshake_set_gtk_state is unable to determine the key length to set handshake state gtk. Fixes: 4bc20a097965 ("ap: Start EAP-WSC authentication with WSC enrollees")
2020-11-14 15:21:45 +01:00
ap_start_handshake(sta, wait_for_eapol_start, NULL);
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
}
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
static struct l_genl_msg *ap_build_cmd_del_key(struct ap_state *ap)
{
uint32_t ifindex = netdev_get_ifindex(ap->netdev);
struct l_genl_msg *msg;
msg = l_genl_msg_new_sized(NL80211_CMD_DEL_KEY, 128);
l_genl_msg_append_attr(msg, NL80211_ATTR_IFINDEX, 4, &ifindex);
l_genl_msg_enter_nested(msg, NL80211_ATTR_KEY);
l_genl_msg_append_attr(msg, NL80211_KEY_IDX, 1, &ap->gtk_index);
l_genl_msg_leave_nested(msg);
return msg;
}
nl80211: introduce nl80211 utility API's Netdev/AP share several NL80211 commands and each has their own builder API's. These were moved into a common file nl80211_util.[ch]. A helper was added to AP for building NEW_STATION to make the associate callback look cleaner (rather than manually building NEW_STATION).
2018-10-08 22:44:09 +02:00
static struct l_genl_msg *ap_build_cmd_new_station(struct sta_state *sta)
{
struct l_genl_msg *msg;
uint32_t ifindex = netdev_get_ifindex(sta->ap->netdev);
struct nl80211_sta_flag_update flags = {
ap: check FULL_AP_CLIENT_STATE for NEW_STATION nl80211 has a check which returns -EINVAL for NEW_STATION if this feature is not enabled and the mask contains Authenticated/Associated flags.
2023-02-24 19:51:46 +01:00
.mask = (1 << NL80211_STA_FLAG_AUTHORIZED) |
nl80211: introduce nl80211 utility API's Netdev/AP share several NL80211 commands and each has their own builder API's. These were moved into a common file nl80211_util.[ch]. A helper was added to AP for building NEW_STATION to make the associate callback look cleaner (rather than manually building NEW_STATION).
2018-10-08 22:44:09 +02:00
(1 << NL80211_STA_FLAG_MFP),
.set = (1 << NL80211_STA_FLAG_AUTHENTICATED) |
(1 << NL80211_STA_FLAG_ASSOCIATED),
};
ap: check FULL_AP_CLIENT_STATE for NEW_STATION nl80211 has a check which returns -EINVAL for NEW_STATION if this feature is not enabled and the mask contains Authenticated/Associated flags.
2023-02-24 19:51:46 +01:00
/*
* Without this feature nl80211 rejects NEW_STATION if the mask contains
* auth/assoc flags
*/
if (wiphy_has_feature(netdev_get_wiphy(sta->ap->netdev),
NL80211_FEATURE_FULL_AP_CLIENT_STATE))
flags.mask |= (1 << NL80211_STA_FLAG_ASSOCIATED) |
(1 << NL80211_STA_FLAG_AUTHENTICATED);
nl80211: introduce nl80211 utility API's Netdev/AP share several NL80211 commands and each has their own builder API's. These were moved into a common file nl80211_util.[ch]. A helper was added to AP for building NEW_STATION to make the associate callback look cleaner (rather than manually building NEW_STATION).
2018-10-08 22:44:09 +02:00
msg = l_genl_msg_new_sized(NL80211_CMD_NEW_STATION, 300);
l_genl_msg_append_attr(msg, NL80211_ATTR_IFINDEX, 4, &ifindex);
l_genl_msg_append_attr(msg, NL80211_ATTR_MAC, 6, sta->addr);
l_genl_msg_append_attr(msg, NL80211_ATTR_STA_FLAGS2, 8, &flags);
return msg;
}
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
static void ap_gtk_op_cb(struct l_genl_msg *msg, void *user_data)
{
if (l_genl_msg_get_error(msg) < 0) {
uint8_t cmd = l_genl_msg_get_command(msg);
const char *cmd_name =
cmd == NL80211_CMD_NEW_KEY ? "NEW_KEY" :
cmd == NL80211_CMD_SET_KEY ? "SET_KEY" :
"DEL_KEY";
l_error("%s failed for the GTK: %i",
cmd_name, l_genl_msg_get_error(msg));
}
}
static void ap_associate_sta_cb(struct l_genl_msg *msg, void *user_data)
{
struct sta_state *sta = user_data;
struct ap_state *ap = sta->ap;
if (l_genl_msg_get_error(msg) < 0) {
l_error("NEW_STATION/SET_STATION failed: %i",
l_genl_msg_get_error(msg));
return;
}
ap: Start EAP-WSC authentication with WSC enrollees After association and sending the SET_STATION commands, set up the handshake_state and eapol_sm for EAP-WSC and start the handshake.
2020-08-28 14:46:49 +02:00
/*
* WSC v2.0.5 Section 8.2:
* "Therefore if a WSC IE is present in the (re)association request,
* the AP shall engage in EAP-WSC with the station and shall not
* attempt any other security handshake."
*
* So no need for group traffic, skip the GTK setup below.
*/
if (!sta->assoc_rsne) {
ap_start_eap_wsc(sta);
return;
}
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
/*
* Set up the group key. If this is our first STA then we have
* to add the new GTK to the kernel. In theory we should be
* able to supply our own RSC (e.g. generated randomly) and use it
* immediately for our 4-Way Handshake without querying the kernel.
* However NL80211_CMD_NEW_KEY only lets us set the receive RSC --
* the Rx PN for CCMP and the Rx IV for TKIP -- and the
* transmit RSC always starts as all zeros. There's effectively
* no way to set the Tx RSC or query the Rx RSC through nl80211.
* So we query the Tx RSC in both scenarios just in case some
* driver/hardware uses a different initial Tx RSC.
*
* Optimally we would get called back by the EAPoL state machine
* only when building the step 3 of 4 message to query the RSC as
* late as possible but that would complicate EAPoL.
*/
if (ap->group_cipher != IE_RSN_CIPHER_SUITE_NO_GROUP_TRAFFIC &&
!ap->gtk_set) {
enum crypto_cipher group_cipher =
ie_rsn_cipher_suite_to_cipher(ap->group_cipher);
int gtk_len = crypto_cipher_key_len(group_cipher);
/*
* Generate our GTK. Not following the example derivation
* method in 802.11-2016 section 12.7.1.4 because a simple
* l_getrandom is just as good.
*/
l_getrandom(ap->gtk, gtk_len);
ap->gtk_index = 1;
nl80211: introduce nl80211 utility API's Netdev/AP share several NL80211 commands and each has their own builder API's. These were moved into a common file nl80211_util.[ch]. A helper was added to AP for building NEW_STATION to make the associate callback look cleaner (rather than manually building NEW_STATION).
2018-10-08 22:44:09 +02:00
msg = nl80211_build_new_key_group(
netdev_get_ifindex(ap->netdev),
group_cipher, ap->gtk_index,
nl80211: support per-mac GTK on _new_key_group AdHoc will require a per-mac GTK to be set. For this reason nl80211_build_new_key_group has been updated to optionally take a MAC address.
2018-10-08 22:44:12 +02:00
ap->gtk, gtk_len, NULL,
0, NULL);
nl80211: introduce nl80211 utility API's Netdev/AP share several NL80211 commands and each has their own builder API's. These were moved into a common file nl80211_util.[ch]. A helper was added to AP for building NEW_STATION to make the associate callback look cleaner (rather than manually building NEW_STATION).
2018-10-08 22:44:09 +02:00
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
if (!l_genl_family_send(ap->nl80211, msg, ap_gtk_op_cb, NULL,
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
NULL)) {
l_genl_msg_unref(msg);
l_error("Issuing NEW_KEY failed");
goto error;
}
nl80211: make SET_KEY/GET_KEY common Moved AP helpers for building SET_KEY/GET_KEY commands into nl80211_util.c
2018-10-08 22:44:10 +02:00
msg = nl80211_build_set_key(netdev_get_ifindex(ap->netdev),
ap->gtk_index);
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
if (!l_genl_family_send(ap->nl80211, msg, ap_gtk_op_cb, NULL,
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
NULL)) {
l_genl_msg_unref(msg);
l_error("Issuing SET_KEY failed");
goto error;
}
/*
* Set the flag now because any new associating STA will
* just use NL80211_CMD_GET_KEY from now.
*/
ap->gtk_set = true;
}
if (ap->group_cipher == IE_RSN_CIPHER_SUITE_NO_GROUP_TRAFFIC)
ap_start_rsna(sta, NULL);
else {
nl80211: make SET_KEY/GET_KEY common Moved AP helpers for building SET_KEY/GET_KEY commands into nl80211_util.c
2018-10-08 22:44:10 +02:00
msg = nl80211_build_get_key(netdev_get_ifindex(ap->netdev),
ap->gtk_index);
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
sta->gtk_query_cmd_id = l_genl_family_send(ap->nl80211, msg,
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
ap_gtk_query_cb,
sta, NULL);
if (!sta->gtk_query_cmd_id) {
l_genl_msg_unref(msg);
l_error("Issuing GET_KEY failed");
goto error;
}
}
return;
error:
ap_del_station(sta, MMPDU_REASON_CODE_UNSPECIFIED, true);
}
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
static void ap_associate_sta(struct ap_state *ap, struct sta_state *sta)
{
struct l_genl_msg *msg;
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
uint32_t ifindex = netdev_get_ifindex(ap->netdev);
nl80211: introduce nl80211 utility API's Netdev/AP share several NL80211 commands and each has their own builder API's. These were moved into a common file nl80211_util.[ch]. A helper was added to AP for building NEW_STATION to make the associate callback look cleaner (rather than manually building NEW_STATION).
2018-10-08 22:44:09 +02:00
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
uint8_t rates[256];
uint32_t r, minr, maxr, count = 0;
uint16_t capability = l_get_le16(&sta->capability);
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
if (sta->associated)
nl80211: introduce nl80211 utility API's Netdev/AP share several NL80211 commands and each has their own builder API's. These were moved into a common file nl80211_util.[ch]. A helper was added to AP for building NEW_STATION to make the associate callback look cleaner (rather than manually building NEW_STATION).
2018-10-08 22:44:09 +02:00
msg = nl80211_build_set_station_associated(ifindex, sta->addr);
else
msg = ap_build_cmd_new_station(sta);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
sta->associated = true;
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
sta->rsna = false;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
minr = l_uintset_find_min(sta->rates);
maxr = l_uintset_find_max(sta->rates);
for (r = minr; r <= maxr; r++)
if (l_uintset_contains(sta->rates, r))
rates[count++] = r;
l_genl_msg_append_attr(msg, NL80211_ATTR_STA_AID, 2, &sta->aid);
l_genl_msg_append_attr(msg, NL80211_ATTR_STA_SUPPORTED_RATES,
count, &rates);
l_genl_msg_append_attr(msg, NL80211_ATTR_STA_LISTEN_INTERVAL, 2,
&sta->listen_interval);
l_genl_msg_append_attr(msg, NL80211_ATTR_STA_CAPABILITY, 2,
&capability);
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
if (!l_genl_family_send(ap->nl80211, msg, ap_associate_sta_cb,
sta, NULL)) {
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
l_genl_msg_unref(msg);
nl80211: introduce nl80211 utility API's Netdev/AP share several NL80211 commands and each has their own builder API's. These were moved into a common file nl80211_util.[ch]. A helper was added to AP for building NEW_STATION to make the associate callback look cleaner (rather than manually building NEW_STATION).
2018-10-08 22:44:09 +02:00
if (l_genl_msg_get_command(msg) == NL80211_CMD_NEW_STATION)
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
l_error("Issuing NEW_STATION failed");
else
l_error("Issuing SET_STATION failed");
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
}
}
static bool ap_common_rates(struct l_uintset *ap_rates,
struct l_uintset *sta_rates)
{
ap: Simplify ap_common_rates If we've checked that the STA has included our Basic Rate we've already found one common rate, no need for further checks.
2017-09-30 04:28:11 +02:00
uint32_t minr = l_uintset_find_min(ap_rates);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
/* Our lowest rate is a Basic Rate so must be supported */
ap: Simplify ap_common_rates If we've checked that the STA has included our Basic Rate we've already found one common rate, no need for further checks.
2017-09-30 04:28:11 +02:00
if (l_uintset_contains(sta_rates, minr))
return true;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
return false;
}
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
static void ap_success_assoc_resp_cb(int err, void *user_data)
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
{
struct sta_state *sta = user_data;
struct ap_state *ap = sta->ap;
sta->assoc_resp_cmd_id = 0;
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
if (err) {
if (err == -ECOMM)
l_error("AP (Re)Association Response received no ACK");
else
l_error("AP (Re)Association Response not sent %s (%i)",
strerror(-err), -err);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
/* If we were in State 3 or 4 go to back to State 2 */
if (sta->associated)
ap: use netdev_del_station API Removes del station code from AP in favor of netdev_del_station.
2018-07-03 23:36:35 +02:00
ap_del_station(sta, MMPDU_REASON_CODE_UNSPECIFIED,
true);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
return;
}
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
/* If we were in State 2, 3 or 4 also go to State 3 */
ap_associate_sta(ap, sta);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
l_info("AP (Re)Association Response ACK received");
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
}
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
static void ap_fail_assoc_resp_cb(int err, void *user_data)
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
{
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
if (err == -ECOMM)
l_error("AP (Re)Association Response with an error status "
"received no ACK");
else if (err)
l_error("AP (Re)Association Response with an error status "
"not sent: %s (%i)", strerror(-err), -err);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
else
src: Fix spelling errors spotted by lintian The debian package checking tool lintian spotted a bunch of spelling errors in the built binaries.
2018-08-04 16:53:35 +02:00
l_info("AP (Re)Association Response with an error status "
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
"delivered OK");
}
static uint32_t ap_assoc_resp(struct ap_state *ap, struct sta_state *sta,
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
const uint8_t *dest,
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
enum mmpdu_reason_code status_code,
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
bool reassoc, const struct mmpdu_header *req,
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
size_t req_len,
const struct ie_fils_ip_addr_request_info *
ip_req_info, frame_xchg_cb_t callback)
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
{
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
const uint8_t *addr = netdev_get_address(ap->netdev);
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
enum mpdu_management_subtype stype = reassoc ?
MPDU_MANAGEMENT_SUBTYPE_REASSOCIATION_RESPONSE :
MPDU_MANAGEMENT_SUBTYPE_ASSOCIATION_RESPONSE;
L_AUTO_FREE_VAR(uint8_t *, mpdu_buf) =
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
l_malloc(256 + ap_get_extra_ies_len(ap, stype, req, req_len));
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
struct mmpdu_header *mpdu = (void *) mpdu_buf;
struct mmpdu_association_response *resp;
size_t ies_len = 0;
ap: make supported rates a common builder. The supported rates IE was being built in two places. This makes that code common. Unfortunately it needs to support both an ie builder and using a pointer directly which is why it only builds the contents of the IE and the caller must set the type/length.
2022-12-20 22:43:09 +01:00
size_t len;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
uint16_t capability = IE_BSS_CAP_ESS | IE_BSS_CAP_PRIVACY;
memset(mpdu, 0, sizeof(*mpdu));
/* Header */
mpdu->fc.protocol_version = 0;
mpdu->fc.type = MPDU_TYPE_MANAGEMENT;
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
mpdu->fc.subtype = stype;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
memcpy(mpdu->address_1, dest, 6); /* DA */
memcpy(mpdu->address_2, addr, 6); /* SA */
memcpy(mpdu->address_3, addr, 6); /* BSSID */
/* Association Response body */
resp = (void *) mmpdu_body(mpdu);
l_put_le16(capability, &resp->capability);
resp->status_code = L_CPU_TO_LE16(status_code);
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
resp->aid = sta ? L_CPU_TO_LE16(sta->aid | 0xc000) : 0;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
/* Supported Rates IE */
resp->ies[ies_len++] = IE_TYPE_SUPPORTED_RATES;
ap: make supported rates a common builder. The supported rates IE was being built in two places. This makes that code common. Unfortunately it needs to support both an ie builder and using a pointer directly which is why it only builds the contents of the IE and the caller must set the type/length.
2022-12-20 22:43:09 +01:00
len = ap_build_supported_rates(ap, resp->ies + ies_len + 1);
resp->ies[ies_len++] = len;
ies_len += len;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
ap: build HT Capabilities/Operation elements If supported this will include the HT capabilities and HT operations elements in beacons/probes. Some shortcuts were taken here since not all the information is currently parsed from the hardware. Namely the HT operation element does not include the basic MCS set. Still, this will at least show stations that the AP is capable of more than just basic rates. The builders themselves are structured similar to the basic rates builder where they build only the contents and return the length. The caller must set the type/length manually. This is to support the two use cases of using with an IE builder vs direct pointer.
2022-12-28 23:32:31 +01:00
if (ap->supports_ht) {
resp->ies[ies_len++] = IE_TYPE_HT_CAPABILITIES;
len = ap_build_ht_capability(ap, resp->ies + ies_len + 1);
resp->ies[ies_len++] = len;
ies_len += len;
resp->ies[ies_len++] = IE_TYPE_HT_OPERATION;
len = ap_build_ht_operation(ap, resp->ies + ies_len + 1);
resp->ies[ies_len++] = len;
ies_len += len;
}
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
ies_len += ap_write_extra_ies(ap, stype, req, req_len,
resp->ies + ies_len);
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
if (ip_req_info) {
struct ie_fils_ip_addr_response_info ip_resp_info = {};
if (ip_req_info->ipv4 && sta && ap_sta_get_dhcp4_lease(sta)) {
uint32_t lease_lifetime =
l_dhcp_lease_get_lifetime(sta->ip_alloc_lease);
ap: Use _u32 dhcp_lease getters
2021-10-14 00:38:44 +02:00
uint32_t gw =
l_dhcp_lease_get_gateway_u32(
sta->ip_alloc_lease);
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
char **lease_dns_str_list =
l_dhcp_lease_get_dns(sta->ip_alloc_lease);
ap: Use _u32 dhcp_lease getters
2021-10-14 00:38:44 +02:00
ip_resp_info.ipv4_addr = l_dhcp_lease_get_address_u32(
sta->ip_alloc_lease);
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
ip_resp_info.ipv4_prefix_len =
ap: Use l_dhcp_lease_get_prefix_length
2021-10-14 00:01:33 +02:00
l_dhcp_lease_get_prefix_length(
sta->ip_alloc_lease);
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
if (lease_lifetime != 0xffffffff)
ip_resp_info.ipv4_lifetime = lease_lifetime;
ap: Use _u32 dhcp_lease getters
2021-10-14 00:38:44 +02:00
if (gw) {
ip_resp_info.ipv4_gateway = gw;
ap: Add MACs to FILS IP Assignment responses Try to include the gateway and DNS MAC addresses in the corresponding fields in the FILS IP Address Assignment IEs we send to the clients.
2021-08-28 05:23:05 +02:00
memcpy(ip_resp_info.ipv4_gateway_mac,
ap->netconfig_gateway4_mac, 6);
}
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
ap: Add MACs to FILS IP Assignment responses Try to include the gateway and DNS MAC addresses in the corresponding fields in the FILS IP Address Assignment IEs we send to the clients.
2021-08-28 05:23:05 +02:00
if (lease_dns_str_list && lease_dns_str_list[0]) {
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
ip_resp_info.ipv4_dns =
IP4_FROM_STR(lease_dns_str_list[0]);
ap: Add MACs to FILS IP Assignment responses Try to include the gateway and DNS MAC addresses in the corresponding fields in the FILS IP Address Assignment IEs we send to the clients.
2021-08-28 05:23:05 +02:00
memcpy(ip_resp_info.ipv4_dns_mac,
ap->netconfig_dns4_mac, 6);
}
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
l_strv_free(lease_dns_str_list);
sta->ip_alloc_sent = true;
} else if (ip_req_info->ipv4 || ip_req_info->ipv6) {
/*
* 802.11ai-2016 Section 11.47.3.3: "If the AP is unable
* to assign an IP address in the (Re)Association
* Response frame, then the AP sets the IP address
* assignment pending flag in the IP Address Response
* Control field of the FILS IP Address Assignment
* element to 1 and sets the IP address request timeout
* to 0 in (Re)Association Response frame."
*/
ip_resp_info.response_pending = 1;
ip_resp_info.response_timeout = 0;
}
ie_build_fils_ip_addr_response(&ip_resp_info,
resp->ies + ies_len);
ies_len += 2 + resp->ies[ies_len + 1];
}
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
return ap_send_mgmt_frame(ap, mpdu, resp->ies + ies_len - mpdu_buf,
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
callback, sta);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
}
ap: move ie_parse_supported_rates into ap.c Supported rates will soon be parsed along with HT/VHT capabilities to determine the best data rate. This will remove the need for the supported_rates uintset element in scan_bss, as well as the single API to only parse the supported rates IE. AP still does rely on this though (since it only supports basic rates), so the parsing function was moved into ap.c.
2019-02-23 00:39:38 +01:00
static int ap_parse_supported_rates(struct ie_tlv_iter *iter,
struct l_uintset **set)
{
const uint8_t *rates;
unsigned int len;
unsigned int i;
len = ie_tlv_iter_get_length(iter);
if (ie_tlv_iter_get_tag(iter) == IE_TYPE_SUPPORTED_RATES && len == 0)
return -EINVAL;
rates = ie_tlv_iter_get_data(iter);
if (!*set)
*set = l_uintset_new(108);
for (i = 0; i < len; i++) {
if (rates[i] == 0xff)
continue;
l_uintset_put(*set, rates[i] & 0x7f);
}
return 0;
}
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
/*
* This handles both the Association and Reassociation Request frames.
* Association Request is documented in 802.11-2016 9.3.3.6 (frame format),
* 802.11-2016 11.3.5.3 (MLME/SME) and Reassociation in 802.11-2016
* 9.3.3.8 (frame format), 802.11-2016 11.3.5.3 (MLME/SME).
*
* The difference between Association and Reassociation procedures is
* documented in 11.3.5.1 "General" but seems inconsistent with specific
* instructions in 11.3.5.3 vs. 11.3.5.5 and 11.3.5.2 vs. 11.3.5.4.
* According to 11.3.5.1:
* 1. Reassociation requires the STA to be already associated in the ESS,
* Association doesn't.
* 2. Unsuccessful Reassociation should not cause a state transition of
* the authentication state between the two STAs.
*
* The first requirement is not present in 11.3.5.5 which is virtually
* identical with 11.3.5.3, but we do implement it. Number 2 is also not
* reflected in 11.3.5.5 where the state transitions are the same as in
* 11.3.5.3 and 11.3.5.4 where the state transitions are the same as in
* 11.3.5.2 including f) "If a Reassociation Response frame is received
* with a status code other than SUCCESS [...] 1. [...] the state for
* the AP [...] shall be set to State 2 [...]"
*
* For the record here are the apparent differences between 802.11-2016
* 11.3.5.2 and 11.3.5.4 ignoring the s/Associate/Reassociate/ changes
* and the special case of Reassociation during a Fast Transition.
* o Points c) and d) are switched around.
* o On success, the STA is disassociated from all other APs in 11.3.5.2,
* and from the previous AP in 11.3.5.4 c). (Shouldn't make a
* difference as there seems to be no way for the STA to become
* associated with more than one AP)
* o After Association a 4-Way Handshake is always performed, after
* Reassociation it is only performed if STA was in State 3 according
* to 11.3.5.4 g). This is not reflected in 11.3.5.5 though.
* Additionally 11.3.5.4 and 11.3.5.5 require the STA and AP
* respectively to delete current PTKSA/GTKSA/IGTKSA at the beginning
* of the procedure independent of the STA state so without a 4-Way
* Handshake the two stations end up with no encryption keys.
*
* The main difference between 11.3.5.3 and 11.3.5.5 is presence of p).
*/
static void ap_assoc_reassoc(struct sta_state *sta, bool reassoc,
const struct mmpdu_field_capability *capability,
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
uint16_t listen_interval,
const uint8_t *ies, size_t ies_len,
const struct mmpdu_header *req)
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
{
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
struct ap_state *ap = sta->ap;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
const char *ssid = NULL;
const uint8_t *rsn = NULL;
ap: Drop a RSNE check After EAPOL logic was moved to eapol.c a check was added to ap_associate_sta_cb to bitwise compare the AP's RSNE to the RSNE received in the (Re)Association frame. There is as far as I know no reason for them to be the same (although they are in our autotest) and if there was a reason we'd rather validate the (Re)Association RSNE immediately when received. We also must set different RSNEs as the "own" (supplicant) and "ap" RSNEs in the handshake_state for validation of step 2/4 in eapol.c (fixes wpa_supplicant's and MS Windows connections being rejected)
2018-08-24 03:37:52 +02:00
size_t ssid_len = 0;
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
_auto_(l_uintset_free) struct l_uintset *rates = NULL;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
struct ie_rsn_info rsn_info;
int err;
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
struct ie_tlv_iter iter;
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
_auto_(l_free) uint8_t *wsc_data = NULL;
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
ssize_t wsc_data_len;
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
bool fils_ip_req = false;
struct ie_fils_ip_addr_request_info fils_ip_req_info;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
if (sta->assoc_resp_cmd_id)
return;
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
if (reassoc && !sta->associated) {
err = MMPDU_REASON_CODE_CLASS3_FRAME_FROM_NONASSOC_STA;
goto unsupported;
}
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
wsc_data = ie_tlv_extract_wsc_payload(ies, ies_len, &wsc_data_len);
ie_tlv_iter_init(&iter, ies, ies_len);
while (ie_tlv_iter_next(&iter))
switch (ie_tlv_iter_get_tag(&iter)) {
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
case IE_TYPE_SSID:
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
ssid = (const char *) ie_tlv_iter_get_data(&iter);
ssid_len = ie_tlv_iter_get_length(&iter);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
break;
case IE_TYPE_SUPPORTED_RATES:
case IE_TYPE_EXTENDED_SUPPORTED_RATES:
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
if (ap_parse_supported_rates(&iter, &rates) < 0) {
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
err = MMPDU_REASON_CODE_INVALID_IE;
goto bad_frame;
}
break;
case IE_TYPE_RSN:
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
/*
* WSC v2.0.5 Section 8.2:
* "Note that during the WSC association [...] the
* RSN IE and the WPA IE are irrelevant and shall be
* ignored by both the station and AP."
*/
if (wsc_data)
break;
if (ie_parse_rsne(&iter, &rsn_info) < 0) {
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
err = MMPDU_REASON_CODE_INVALID_IE;
goto bad_frame;
}
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
rsn = (const uint8_t *) ie_tlv_iter_get_data(&iter) - 2;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
break;
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
case IE_TYPE_FILS_IP_ADDRESS:
if (fils_ip_req || ie_parse_fils_ip_addr_request(&iter,
&fils_ip_req_info) < 0) {
l_debug("Can't parse FILS IP Address Assignment"
" IE, ignoring it");
break;
}
fils_ip_req = true;
break;
ap: build HT Capabilities/Operation elements If supported this will include the HT capabilities and HT operations elements in beacons/probes. Some shortcuts were taken here since not all the information is currently parsed from the hardware. Namely the HT operation element does not include the basic MCS set. Still, this will at least show stations that the AP is capable of more than just basic rates. The builders themselves are structured similar to the basic rates builder where they build only the contents and return the length. The caller must set the type/length manually. This is to support the two use cases of using with an IE builder vs direct pointer.
2022-12-28 23:32:31 +01:00
case IE_TYPE_HT_CAPABILITIES:
if (ie_tlv_iter_get_length(&iter) != 26) {
err = MMPDU_REASON_CODE_INVALID_IE;
goto bad_frame;
}
if (test_bit(ie_tlv_iter_get_data(&iter), 4))
sta->ht_greenfield = true;
sta->ht_support = true;
break;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
}
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
if (!rates || !ssid || (!wsc_data && !rsn) ||
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
ssid_len != strlen(ap->ssid) ||
memcmp(ssid, ap->ssid, ssid_len)) {
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
err = MMPDU_REASON_CODE_INVALID_IE;
goto bad_frame;
}
ap: Reorder ap_assoc_req_cb, fix comment Set all the new field values into struct sta_state only after all the error checks for better readabilty and fixing a possible issue if we did "sta->rates = rates" and then detected en error and freed "rates". Also update a comment which I think used the wording from 802.11-2012 instead of 802.11-2016.
2017-09-30 04:28:09 +02:00
if (!ap_common_rates(ap->rates, rates)) {
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
err = MMPDU_REASON_CODE_UNSPECIFIED;
goto unsupported;
}
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
/* Is the client requesting RSNA establishment or WSC registration */
if (!rsn) {
struct wsc_association_request wsc_req;
struct ap_event_registration_start_data event_data;
struct ap_wsc_pbc_probe_record *record;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
if (wsc_parse_association_request(wsc_data, wsc_data_len,
&wsc_req) < 0) {
err = MMPDU_REASON_CODE_INVALID_IE;
goto bad_frame;
}
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
if (wsc_req.request_type !=
WSC_REQUEST_TYPE_ENROLLEE_OPEN_8021X) {
err = MMPDU_REASON_CODE_INVALID_IE;
goto bad_frame;
}
if (!ap->wsc_pbc_timeout) {
l_debug("WSC association from %s but we're not in "
"PBC mode", util_address_to_string(sta->addr));
err = MMPDU_REASON_CODE_UNSPECIFIED;
goto bad_frame;
}
if (l_queue_isempty(ap->wsc_pbc_probes)) {
l_debug("%s tried to register as enrollee but we "
"don't have their Probe Request record",
util_address_to_string(sta->addr));
err = MMPDU_REASON_CODE_UNSPECIFIED;
goto bad_frame;
}
/*
* For PBC, the Enrollee must have sent the only PBC Probe
* Request within the monitor time and walk time.
*/
record = l_queue_peek_head(ap->wsc_pbc_probes);
if (memcmp(sta->addr, record->mac, 6)) {
l_debug("Session overlap during %s's attempt to "
"register as WSC enrollee",
util_address_to_string(sta->addr));
err = MMPDU_REASON_CODE_UNSPECIFIED;
goto bad_frame;
}
memcpy(sta->wsc_uuid_e, record->uuid_e, 16);
sta->wsc_v2 = wsc_req.version2;
event_data.mac = sta->addr;
ap: Pass frame IEs from clients to the ap_state user Pass the string IEs from the incoming STA association frames to the user in the AP event data. I drop ap_event_station_added_data.rsn_ie because that probably wasn't going to ever be useful and the RSN IE is included in the .assoc_ies array in any case.
2021-03-10 23:06:31 +01:00
event_data.assoc_ies = ies;
event_data.assoc_ies_len = ies_len;
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
if (ap_event(ap, AP_EVENT_REGISTRATION_START, &event_data))
return;
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
/*
* Since we're starting the PBC Registration Protocol
* we can now exit the "active PBC mode".
*/
ap_wsc_exit_pbc(ap);
} else {
if (rsn_info.mfpr && rsn_info.spp_a_msdu_required) {
err = MMPDU_REASON_CODE_UNSPECIFIED;
goto unsupported;
}
eapol,ap: Remove assumption of single cipher in authenticator IE Allow the user of the eapol_sm & handshake_state APIs to have multiple pairwise ciphers listed in the authenticator IE.
2021-02-01 14:40:42 +01:00
if (__builtin_popcount(rsn_info.pairwise_ciphers) != 1 ||
!(rsn_info.pairwise_ciphers & ap->ciphers)) {
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
err = MMPDU_REASON_CODE_INVALID_PAIRWISE_CIPHER;
goto unsupported;
}
if (rsn_info.akm_suites != IE_RSN_AKM_SUITE_PSK) {
err = MMPDU_REASON_CODE_INVALID_AKMP;
goto unsupported;
}
ap: validate group cipher Make sure to validate group_cipher from the STA similarly to how akm_suites and pairwise_ciphers are validated.
2021-09-20 21:25:34 +02:00
if (rsn_info.group_cipher != ap->group_cipher) {
err = MMPDU_REASON_CODE_INVALID_GROUP_CIPHER;
goto unsupported;
}
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
}
ap: Stop ongoing handshake on reassociation On a new association or re-association, in addition to forgetting a complete RSN Association, also stop the EAPoL SM to stop any ongoing handshake. Do this in a new function ap_stop_handshake that is now used in a few places that had copies of the same few lines. I'll be adding some more lines to this function for WSC support.
2020-08-28 14:46:45 +02:00
/* 802.11-2016 11.3.5.3 j) */
if (sta->rsna)
ap_drop_rsna(sta);
else if (sta->associated)
ap_stop_handshake(sta);
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
if (!sta->associated) {
/*
* Everything fine so far, assign an AID, send response.
* According to 802.11-2016 11.3.5.3 l) we will only go to
* State 3 (set sta->associated) once we receive the station's
* ACK or gave up on resends.
*/
sta->aid = ++ap->last_aid;
}
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
sta->capability = *capability;
sta->listen_interval = listen_interval;
ap: Reorder ap_assoc_req_cb, fix comment Set all the new field values into struct sta_state only after all the error checks for better readabilty and fixing a possible issue if we did "sta->rates = rates" and then detected en error and freed "rates". Also update a comment which I think used the wording from 802.11-2012 instead of 802.11-2016.
2017-09-30 04:28:09 +02:00
if (sta->rates)
l_uintset_free(sta->rates);
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
sta->rates = l_steal_ptr(rates);
ap: Reorder ap_assoc_req_cb, fix comment Set all the new field values into struct sta_state only after all the error checks for better readabilty and fixing a possible issue if we did "sta->rates = rates" and then detected en error and freed "rates". Also update a comment which I think used the wording from 802.11-2012 instead of 802.11-2016.
2017-09-30 04:28:09 +02:00
ap: Pass frame IEs from clients to the ap_state user Pass the string IEs from the incoming STA association frames to the user in the AP event data. I drop ap_event_station_added_data.rsn_ie because that probably wasn't going to ever be useful and the RSN IE is included in the .assoc_ies array in any case.
2021-03-10 23:06:31 +01:00
l_free(sta->assoc_ies);
ap: Reorder ap_assoc_req_cb, fix comment Set all the new field values into struct sta_state only after all the error checks for better readabilty and fixing a possible issue if we did "sta->rates = rates" and then detected en error and freed "rates". Also update a comment which I think used the wording from 802.11-2012 instead of 802.11-2016.
2017-09-30 04:28:09 +02:00
ap: Pass frame IEs from clients to the ap_state user Pass the string IEs from the incoming STA association frames to the user in the AP event data. I drop ap_event_station_added_data.rsn_ie because that probably wasn't going to ever be useful and the RSN IE is included in the .assoc_ies array in any case.
2021-03-10 23:06:31 +01:00
if (rsn) {
sta->assoc_ies = l_memdup(ies, ies_len);
sta->assoc_ies_len = ies_len;
sta->assoc_rsne = sta->assoc_ies + (rsn - ies);
} else {
sta->assoc_ies = NULL;
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
sta->assoc_rsne = NULL;
ap: Pass frame IEs from clients to the ap_state user Pass the string IEs from the incoming STA association frames to the user in the AP event data. I drop ap_event_station_added_data.rsn_ie because that probably wasn't going to ever be useful and the RSN IE is included in the .assoc_ies array in any case.
2021-03-10 23:06:31 +01:00
}
ap: Reorder ap_assoc_req_cb, fix comment Set all the new field values into struct sta_state only after all the error checks for better readabilty and fixing a possible issue if we did "sta->rates = rates" and then detected en error and freed "rates". Also update a comment which I think used the wording from 802.11-2012 instead of 802.11-2016.
2017-09-30 04:28:09 +02:00
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
sta->assoc_resp_cmd_id = ap_assoc_resp(ap, sta, sta->addr, 0, reassoc,
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
req, (void *) ies + ies_len -
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
(void *) req, fils_ip_req ?
&fils_ip_req_info : NULL,
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
ap_success_assoc_resp_cb);
if (!sta->assoc_resp_cmd_id)
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
l_error("Sending success (Re)Association Response failed");
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
return;
unsupported:
bad_frame:
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
/*
* TODO: MFP
*
ap: Reorder ap_assoc_req_cb, fix comment Set all the new field values into struct sta_state only after all the error checks for better readabilty and fixing a possible issue if we did "sta->rates = rates" and then detected en error and freed "rates". Also update a comment which I think used the wording from 802.11-2012 instead of 802.11-2016.
2017-09-30 04:28:09 +02:00
* 802.11-2016 11.3.5.3 m)
* "If the ResultCode in the MLME-ASSOCIATE.response primitive is
* not SUCCESS and management frame protection is in use the state
* for the STA shall be left unchanged. If the ResultCode is not
* SUCCESS and management frame protection is not in use the state
* for the STA shall be set to State 3 if it was State 4."
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
*
* For now, we need to drop the RSNA.
*/
ap: Stop ongoing handshake on reassociation On a new association or re-association, in addition to forgetting a complete RSN Association, also stop the EAPoL SM to stop any ongoing handshake. Do this in a new function ap_stop_handshake that is now used in a few places that had copies of the same few lines. I'll be adding some more lines to this function for WSC support.
2020-08-28 14:46:45 +02:00
if (sta->rsna)
ap: removed code that was moved to eapol All handshake packet handling has been removed from ap and moved to eapol. After association, the AP registers a new authenticator state machine which handles the AP side of the handshake. AP will receive a handshake event once the 4-way handshake is complete.
2018-06-27 19:06:31 +02:00
ap_drop_rsna(sta);
ap: Stop ongoing handshake on reassociation On a new association or re-association, in addition to forgetting a complete RSN Association, also stop the EAPoL SM to stop any ongoing handshake. Do this in a new function ap_stop_handshake that is now used in a few places that had copies of the same few lines. I'll be adding some more lines to this function for WSC support.
2020-08-28 14:46:45 +02:00
else if (sta->associated)
ap_stop_handshake(sta);
ap: Handle PTK handshake frame 4/4 Handle the 4-Way Handshake last frame, install keys, enable data frames on the controlled port.
2017-09-22 05:06:43 +02:00
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
if (!ap_assoc_resp(ap, sta, sta->addr, err, reassoc,
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
req, (void *) ies + ies_len - (void *) req,
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
NULL, ap_fail_assoc_resp_cb))
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
l_error("Sending error (Re)Association Response failed");
}
/* 802.11-2016 9.3.3.6 */
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
static void ap_assoc_req_cb(const struct mmpdu_header *hdr, const void *body,
size_t body_len, int rssi, void *user_data)
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
{
struct ap_state *ap = user_data;
struct sta_state *sta;
const uint8_t *from = hdr->address_2;
const struct mmpdu_association_request *req = body;
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
const uint8_t *bssid = netdev_get_address(ap->netdev);
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
l_info("AP Association Request from %s", util_address_to_string(from));
if (memcmp(hdr->address_1, bssid, 6) ||
memcmp(hdr->address_3, bssid, 6))
return;
sta = l_queue_find(ap->sta_states, ap_sta_match_addr, from);
if (!sta) {
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
if (!ap_assoc_resp(ap, NULL, from,
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
MMPDU_REASON_CODE_STA_REQ_ASSOC_WITHOUT_AUTH,
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
false, hdr, body + body_len - (void *) hdr,
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
NULL, ap_fail_assoc_resp_cb))
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
l_error("Sending error Association Response failed");
return;
}
ap_assoc_reassoc(sta, false, &req->capability,
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
L_LE16_TO_CPU(req->listen_interval),
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
req->ies, body_len - sizeof(*req), hdr);
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
}
/* 802.11-2016 9.3.3.8 */
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
static void ap_reassoc_req_cb(const struct mmpdu_header *hdr, const void *body,
size_t body_len, int rssi, void *user_data)
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
{
struct ap_state *ap = user_data;
struct sta_state *sta;
const uint8_t *from = hdr->address_2;
const struct mmpdu_reassociation_request *req = body;
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
const uint8_t *bssid = netdev_get_address(ap->netdev);
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
int err;
l_info("AP Reassociation Request from %s",
util_address_to_string(from));
if (memcmp(hdr->address_1, bssid, 6) ||
memcmp(hdr->address_3, bssid, 6))
return;
sta = l_queue_find(ap->sta_states, ap_sta_match_addr, from);
if (!sta) {
err = MMPDU_REASON_CODE_STA_REQ_ASSOC_WITHOUT_AUTH;
goto bad_frame;
}
if (memcmp(req->current_ap_address, bssid, 6)) {
err = MMPDU_REASON_CODE_UNSPECIFIED;
goto bad_frame;
}
ap_assoc_reassoc(sta, true, &req->capability,
ap: Parse WSC PBC association request and build response Check the conditions for PBC enrollee registration when we receive the Association Request with WSC IE and indicate to the enrollee whether we accept the association using a WSC IE in the Association Response. After this, a NULL sta->assoc_rsne indicates that the station is not establishing the RSNA and is a WSC enrollee.
2020-08-28 14:46:48 +02:00
L_LE16_TO_CPU(req->listen_interval),
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
req->ies, body_len - sizeof(*req), hdr);
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
return;
bad_frame:
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
if (!ap_assoc_resp(ap, NULL, from, err, true,
hdr, body + body_len - (void *) hdr,
ap: Support FILS IP Address Assignment IE Handle the 802.11ai FILS IP Address Assignment IEs in Association Request frames when netconfig is enabled. Only IPv4 is supported. Like the P2P IP Allocation mechanism, since the payload format and logic is independent from the rest of the FILS standard this is enabled unconditionally for clients who want to use it even though we don't actually do FILS in AP mode.
2021-08-23 16:14:27 +02:00
NULL, ap_fail_assoc_resp_cb))
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
l_error("Sending error Reassociation Response failed");
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
}
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
static void ap_probe_resp_cb(int err, void *user_data)
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
{
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
if (err == -ECOMM)
l_error("AP Probe Response received no ACK");
else if (err)
l_error("AP Probe Response not sent: %s (%i)",
strerror(-err), -err);
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
else
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
l_info("AP Probe Response delivered OK");
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
}
/*
* Parse Probe Request according to 802.11-2016 9.3.3.10 and act according
* to 802.11-2016 11.1.4.3
*/
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
static void ap_probe_req_cb(const struct mmpdu_header *hdr, const void *body,
size_t body_len, int rssi, void *user_data)
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
{
struct ap_state *ap = user_data;
const struct mmpdu_probe_request *req = body;
const char *ssid = NULL;
const uint8_t *ssid_list = NULL;
size_t ssid_len = 0, ssid_list_len = 0, len;
ap: update to use new TLV builder APIs AP still relies on the get_data/set_length semantics. Its more convenient to still use these since it avoids the need for extra temporary buffers when building the rates IE.
2019-04-23 19:47:26 +02:00
uint8_t dsss_channel = 0;
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
struct ie_tlv_iter iter;
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
const uint8_t *bssid = netdev_get_address(ap->netdev);
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
bool match = false;
ap: Fix invalid length argument to ap_build_beacon_pr_head Previously resp was a simple array of bytes allocated on the stack. This was changed to a dynamically allocated array, but the sizeof(resp) argument to ap_build_beacon_pr_head() was never changed appropriately. Fix this by introducing a new resp_len variable that holds the number of bytes allocated for resp. Also, move the allocation after the basic sanity checks have been performed to avoid allocating/freeing memory unnecessarily. Fixes: 18a63f91fd44 ("ap: Write extra frame IEs from the user")
2021-04-30 23:31:04 +02:00
uint32_t resp_len;
uint8_t *resp;
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
l_info("AP Probe Request from %s",
util_address_to_string(hdr->address_2));
ie_tlv_iter_init(&iter, req->ies, body_len - sizeof(*req));
while (ie_tlv_iter_next(&iter))
switch (ie_tlv_iter_get_tag(&iter)) {
case IE_TYPE_SSID:
ssid = (const char *) ie_tlv_iter_get_data(&iter);
ssid_len = ie_tlv_iter_get_length(&iter);
break;
case IE_TYPE_SSID_LIST:
ssid_list = ie_tlv_iter_get_data(&iter);
ssid_list_len = ie_tlv_iter_get_length(&iter);
break;
case IE_TYPE_DSSS_PARAMETER_SET:
if (ie_tlv_iter_get_length(&iter) != 1)
return;
dsss_channel = ie_tlv_iter_get_data(&iter)[0];
break;
}
/*
* Check if we should reply to this Probe Request according to
* 802.11-2016 section 11.1.4.3.2.
*/
if (memcmp(hdr->address_1, bssid, 6) &&
!util_is_broadcast_address(hdr->address_1))
ap: Fix incoming Probe Request BSSID check Setting 'match' false wouldn't do anything because it was already false. If the frame is addressed to some other non-broadcast address ignore it directly and exit ap_probe_req_cb.
2020-08-27 20:14:48 +02:00
return;
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
if (memcmp(hdr->address_3, bssid, 6) &&
!util_is_broadcast_address(hdr->address_3))
ap: Fix incoming Probe Request BSSID check Setting 'match' false wouldn't do anything because it was already false. If the frame is addressed to some other non-broadcast address ignore it directly and exit ap_probe_req_cb.
2020-08-27 20:14:48 +02:00
return;
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
if (!ssid || ssid_len == 0) /* Wildcard SSID */
match = true;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
else if (ssid && ssid_len == strlen(ap->ssid) && /* One SSID */
!memcmp(ssid, ap->ssid, ssid_len))
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
match = true;
ap: Accept P2P wildcard SSIDs in probe requests Add the special case "DIRECT-" SSID, called the P2P Wildcard SSID, in ap_probe_req_cb so as not to reject those Probe Requests on the basis of ssid mismatch. I'd have preferred to keep all the P2P-specific bits in p2p.c but in this case there's little point in adding a generic config setting for SSID-matching quirks.
2020-09-09 01:49:28 +02:00
else if (ssid && ssid_len == 7 && !memcmp(ssid, "DIRECT-", 7) &&
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
!memcmp(ssid, ap->ssid, 7)) /* P2P wildcard */
ap: Accept P2P wildcard SSIDs in probe requests Add the special case "DIRECT-" SSID, called the P2P Wildcard SSID, in ap_probe_req_cb so as not to reject those Probe Requests on the basis of ssid mismatch. I'd have preferred to keep all the P2P-specific bits in p2p.c but in this case there's little point in adding a generic config setting for SSID-matching quirks.
2020-09-09 01:49:28 +02:00
match = true;
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
else if (ssid_list) { /* SSID List */
ie_tlv_iter_init(&iter, ssid_list, ssid_list_len);
while (ie_tlv_iter_next(&iter)) {
if (ie_tlv_iter_get_tag(&iter) != IE_TYPE_SSID)
return;
ssid = (const char *) ie_tlv_iter_get_data(&iter);
ssid_len = ie_tlv_iter_get_length(&iter);
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
if (ssid_len == strlen(ap->ssid) &&
!memcmp(ssid, ap->ssid, ssid_len)) {
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
match = true;
break;
}
}
}
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
if (dsss_channel != 0 && dsss_channel != ap->channel)
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
match = false;
if (!match)
return;
ap: Fix invalid length argument to ap_build_beacon_pr_head Previously resp was a simple array of bytes allocated on the stack. This was changed to a dynamically allocated array, but the sizeof(resp) argument to ap_build_beacon_pr_head() was never changed appropriately. Fix this by introducing a new resp_len variable that holds the number of bytes allocated for resp. Also, move the allocation after the basic sanity checks have been performed to avoid allocating/freeing memory unnecessarily. Fixes: 18a63f91fd44 ("ap: Write extra frame IEs from the user")
2021-04-30 23:31:04 +02:00
resp_len = 512 + ap_get_extra_ies_len(ap,
MPDU_MANAGEMENT_SUBTYPE_PROBE_RESPONSE,
hdr, body + body_len - (void *) hdr);
resp = l_new(uint8_t, resp_len);
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
len = ap_build_beacon_pr_head(ap,
MPDU_MANAGEMENT_SUBTYPE_PROBE_RESPONSE,
ap: Fix invalid length argument to ap_build_beacon_pr_head Previously resp was a simple array of bytes allocated on the stack. This was changed to a dynamically allocated array, but the sizeof(resp) argument to ap_build_beacon_pr_head() was never changed appropriately. Fix this by introducing a new resp_len variable that holds the number of bytes allocated for resp. Also, move the allocation after the basic sanity checks have been performed to avoid allocating/freeing memory unnecessarily. Fixes: 18a63f91fd44 ("ap: Write extra frame IEs from the user")
2021-04-30 23:31:04 +02:00
hdr->address_2, resp, resp_len);
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
len += ap_build_beacon_pr_tail(ap,
MPDU_MANAGEMENT_SUBTYPE_PROBE_RESPONSE,
hdr, body + body_len - (void *) hdr,
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
resp + len, resp_len - len);
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
ap_send_mgmt_frame(ap, (struct mmpdu_header *) resp, len,
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
ap_probe_resp_cb, NULL);
ap: Fix invalid length argument to ap_build_beacon_pr_head Previously resp was a simple array of bytes allocated on the stack. This was changed to a dynamically allocated array, but the sizeof(resp) argument to ap_build_beacon_pr_head() was never changed appropriately. Fix this by introducing a new resp_len variable that holds the number of bytes allocated for resp. Also, move the allocation after the basic sanity checks have been performed to avoid allocating/freeing memory unnecessarily. Fixes: 18a63f91fd44 ("ap: Write extra frame IEs from the user")
2021-04-30 23:31:04 +02:00
l_free(resp);
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
}
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
/* 802.11-2016 9.3.3.5 (frame format), 802.11-2016 11.3.5.9 (MLME/SME) */
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
static void ap_disassoc_cb(const struct mmpdu_header *hdr, const void *body,
size_t body_len, int rssi, void *user_data)
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
{
struct ap_state *ap = user_data;
struct sta_state *sta;
const struct mmpdu_disassociation *disassoc = body;
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
const uint8_t *bssid = netdev_get_address(ap->netdev);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
l_info("AP Disassociation from %s, reason %i",
util_address_to_string(hdr->address_2),
(int) L_LE16_TO_CPU(disassoc->reason_code));
if (memcmp(hdr->address_1, bssid, 6) ||
memcmp(hdr->address_3, bssid, 6))
return;
sta = l_queue_find(ap->sta_states, ap_sta_match_addr, hdr->address_2);
if (sta && sta->assoc_resp_cmd_id) {
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
l_genl_family_cancel(ap->nl80211, sta->assoc_resp_cmd_id);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
sta->assoc_resp_cmd_id = 0;
}
if (!sta || !sta->associated)
return;
ap: use netdev_del_station API Removes del station code from AP in favor of netdev_del_station.
2018-07-03 23:36:35 +02:00
ap_del_station(sta, L_LE16_TO_CPU(disassoc->reason_code), true);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
}
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
static void ap_auth_reply_cb(int err, void *user_data)
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
{
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
if (err == -ECOMM)
l_error("AP Authentication frame 2 received no ACK");
else if (err)
l_error("AP Authentication frame 2 not sent: %s (%i)",
strerror(-err), -err);
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
else
l_info("AP Authentication frame 2 ACKed by STA");
}
static void ap_auth_reply(struct ap_state *ap, const uint8_t *dest,
enum mmpdu_reason_code status_code)
{
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
const uint8_t *addr = netdev_get_address(ap->netdev);
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
uint8_t mpdu_buf[64];
struct mmpdu_header *mpdu = (struct mmpdu_header *) mpdu_buf;
struct mmpdu_authentication *auth;
memset(mpdu, 0, sizeof(*mpdu));
/* Header */
mpdu->fc.protocol_version = 0;
mpdu->fc.type = MPDU_TYPE_MANAGEMENT;
mpdu->fc.subtype = MPDU_MANAGEMENT_SUBTYPE_AUTHENTICATION;
memcpy(mpdu->address_1, dest, 6); /* DA */
memcpy(mpdu->address_2, addr, 6); /* SA */
memcpy(mpdu->address_3, addr, 6); /* BSSID */
/* Authentication body */
auth = (void *) mmpdu_body(mpdu);
auth->algorithm = L_CPU_TO_LE16(MMPDU_AUTH_ALGO_OPEN_SYSTEM);
auth->transaction_sequence = L_CPU_TO_LE16(2);
auth->status = L_CPU_TO_LE16(status_code);
ap: Use frame-xchg when sending frames Convert ap_send_mgmt_frame() to use frame_xchg_start for sending frames, this fixes among other things the ACK-received checks. One side effect is that we're no longer sending Probe Responses with the don't-wait-for-ack flag because frame-xchg doesn't support it, but other AP implementations don't use that flag either. Another side-effect is that we do use the no-cck-rate flag unconditionally, something we may want to fix but would need to add another parameter to frame-xchg.
2020-09-21 21:04:25 +02:00
ap_send_mgmt_frame(ap, mpdu, (uint8_t *) auth + 6 - mpdu_buf,
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
ap_auth_reply_cb, NULL);
}
/*
* 802.11-2016 9.3.3.12 (frame format), 802.11-2016 11.3.4.3 and
* 802.11-2016 12.3.3.2 (MLME/SME)
*/
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
static void ap_auth_cb(const struct mmpdu_header *hdr, const void *body,
size_t body_len, int rssi, void *user_data)
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
{
struct ap_state *ap = user_data;
const struct mmpdu_authentication *auth = body;
const uint8_t *from = hdr->address_2;
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
const uint8_t *bssid = netdev_get_address(ap->netdev);
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
struct sta_state *sta;
l_info("AP Authentication from %s", util_address_to_string(from));
if (memcmp(hdr->address_1, bssid, 6) ||
memcmp(hdr->address_3, bssid, 6))
return;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
if (ap->authorized_macs_num) {
ap: Move AP parameters to a struct To limit the number of ap_start parameters, group basic AP config parameters in the ap_config struct that is passed as a pointer and owned by the ap_state.
2020-08-27 20:14:46 +02:00
unsigned int i;
ap: Add authorized_macs parameter
2020-08-04 16:53:03 +02:00
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
for (i = 0; i < ap->authorized_macs_num; i++)
if (!memcmp(from, ap->authorized_macs + i * 6,
ap: Move AP parameters to a struct To limit the number of ap_start parameters, group basic AP config parameters in the ap_config struct that is passed as a pointer and owned by the ap_state.
2020-08-27 20:14:46 +02:00
6))
ap: Add authorized_macs parameter
2020-08-04 16:53:03 +02:00
break;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
if (i == ap->authorized_macs_num) {
ap: Add authorized_macs parameter
2020-08-04 16:53:03 +02:00
ap_auth_reply(ap, from, MMPDU_REASON_CODE_UNSPECIFIED);
return;
}
}
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
/* Only Open System authentication implemented here */
if (L_LE16_TO_CPU(auth->algorithm) !=
MMPDU_AUTH_ALGO_OPEN_SYSTEM) {
ap_auth_reply(ap, from, MMPDU_REASON_CODE_UNSPECIFIED);
return;
}
if (L_LE16_TO_CPU(auth->transaction_sequence) != 1) {
ap_auth_reply(ap, from, MMPDU_REASON_CODE_UNSPECIFIED);
return;
}
sta = l_queue_find(ap->sta_states, ap_sta_match_addr, from);
/*
* Figure 11-13 in 802.11-2016 11.3.2 shows a transition from
* States 3 / 4 to State 2 on "Successful 802.11 Authentication"
* however 11.3.4.2 and 11.3.4.3 clearly say the connection goes to
* State 2 only if it was in State 1:
*
* "c) [...] the state for the indicated STA shall be set to State 2
* if it was State 1; the state shall remain unchanged if it was other
* than State 1."
*/
if (sta)
goto done;
/*
* Per 12.3.3.2.3 with Open System the state change is immediate,
* no waiting for the response to be ACKed as with the association
* frames.
*/
sta = l_new(struct sta_state, 1);
memcpy(sta->addr, from, 6);
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
sta->ap = ap;
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
if (!ap->sta_states)
ap->sta_states = l_queue_new();
l_queue_push_tail(ap->sta_states, sta);
/*
* Nothing to do here netlink-wise as we can't receive any data
* frames until after association anyway. We do need to add a
* timeout for the authentication and possibly the kernel could
* handle that if we registered the STA with NEW_STATION now (TODO)
*/
done:
ap_auth_reply(ap, from, 0);
}
/* 802.11-2016 9.3.3.13 (frame format), 802.11-2016 11.3.4.5 (MLME/SME) */
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
static void ap_deauth_cb(const struct mmpdu_header *hdr, const void *body,
size_t body_len, int rssi, void *user_data)
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
{
struct ap_state *ap = user_data;
const struct mmpdu_deauthentication *deauth = body;
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
const uint8_t *bssid = netdev_get_address(ap->netdev);
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
l_info("AP Deauthentication from %s, reason %i",
util_address_to_string(hdr->address_2),
(int) L_LE16_TO_CPU(deauth->reason_code));
if (memcmp(hdr->address_1, bssid, 6) ||
memcmp(hdr->address_3, bssid, 6))
return;
ap: Make station removal safer Replace instances of the ap_del_station() + ap_sta_free()/ap_remove_sta() with calls to ap_station_disconnect to make sure we consistently remove the station from the ap->sta_states queue before using ap_del_station(). ap_del_station() may generate an event to the ap.h API user (e.g. P2P) and this may end up tearing down the AP completely. For that scenario we also don't want ap_sta_free() to access sta->ap so we make sure ap_del_station() performs these cleanup steps so that ap_sta_free() has nothing to do that accesses sta->ap.
2021-08-07 04:10:42 +02:00
ap_station_disconnect(ap, hdr->address_2,
L_LE16_TO_CPU(deauth->reason_code));
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
}
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
static void do_debug(const char *str, void *user_data)
{
const char *prefix = user_data;
l_info("%s%s", prefix, str);
}
ap: Send a specific error message on async AP start failure We generate the DBus error reply type from the errno only when ap_start() was failing synchronously, now also send the errno through the callbacks so that we can also return a specific DBus reply when failing asynchronously. Thea AP autotest relies on receiving the AlreadyExists DBus error.
2021-05-28 02:47:34 +02:00
static void ap_start_failed(struct ap_state *ap, int err)
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
{
ap: Send a specific error message on async AP start failure We generate the DBus error reply type from the errno only when ap_start() was failing synchronously, now also send the errno through the callbacks so that we can also return a specific DBus reply when failing asynchronously. Thea AP autotest relies on receiving the AlreadyExists DBus error.
2021-05-28 02:47:34 +02:00
struct ap_event_start_failed_data data = { err };
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
ap->in_event = true;
ap: Send a specific error message on async AP start failure We generate the DBus error reply type from the errno only when ap_start() was failing synchronously, now also send the errno through the callbacks so that we can also return a specific DBus reply when failing asynchronously. Thea AP autotest relies on receiving the AlreadyExists DBus error.
2021-05-28 02:47:34 +02:00
ap->ops->handle_event(AP_EVENT_START_FAILED, &data, ap->user_data);
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
ap_reset(ap);
l_genl_family_free(ap->nl80211);
l_free(ap);
}
ap: Forward DHCP events to AP event handler
2021-06-04 03:50:45 +02:00
static void ap_dhcp_event_cb(struct l_dhcp_server *server,
enum l_dhcp_server_event event, void *user_data,
const struct l_dhcp_lease *lease)
{
struct ap_state *ap = user_data;
switch (event) {
case L_DHCP_SERVER_EVENT_NEW_LEASE:
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
ap_event(ap, AP_EVENT_DHCP_NEW_LEASE, lease);
ap: Forward DHCP events to AP event handler
2021-06-04 03:50:45 +02:00
break;
case L_DHCP_SERVER_EVENT_LEASE_EXPIRED:
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
ap_event(ap, AP_EVENT_DHCP_LEASE_EXPIRED, lease);
ap: Forward DHCP events to AP event handler
2021-06-04 03:50:45 +02:00
break;
default:
break;
}
}
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
static void ap_start_cb(struct l_genl_msg *msg, void *user_data)
{
struct ap_state *ap = user_data;
ap->start_stop_cmd_id = 0;
if (l_genl_msg_get_error(msg) < 0) {
l_error("START_AP failed: %i", l_genl_msg_get_error(msg));
ap: Send a specific error message on async AP start failure We generate the DBus error reply type from the errno only when ap_start() was failing synchronously, now also send the errno through the callbacks so that we can also return a specific DBus reply when failing asynchronously. Thea AP autotest relies on receiving the AlreadyExists DBus error.
2021-05-28 02:47:34 +02:00
ap_start_failed(ap, l_genl_msg_get_error(msg));
return;
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
}
ap: Forward DHCP events to AP event handler
2021-06-04 03:50:45 +02:00
if (ap->netconfig_dhcp) {
if (!l_dhcp_server_start(ap->netconfig_dhcp)) {
l_error("DHCP server failed to start");
ap_start_failed(ap, -EINVAL);
return;
}
if (!l_dhcp_server_set_event_handler(ap->netconfig_dhcp,
ap_dhcp_event_cb,
ap, NULL)) {
l_error("l_dhcp_server_set_event_handler failed");
ap_start_failed(ap, -EIO);
return;
}
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
}
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap->started = true;
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
ap_event(ap, AP_EVENT_STARTED, NULL);
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
}
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
static struct l_genl_msg *ap_build_cmd_start_ap(struct ap_state *ap)
{
struct l_genl_msg *cmd;
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
uint8_t head[256];
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
size_t tail_len = 256 + ap_get_extra_ies_len(ap,
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
MPDU_MANAGEMENT_SUBTYPE_BEACON,
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
NULL, 0);
L_AUTO_FREE_VAR(uint8_t *, tail) = l_malloc(tail_len);
size_t head_len;
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
uint32_t dtim_period = 3;
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
uint32_t ifindex = netdev_get_ifindex(ap->netdev);
struct wiphy *wiphy = netdev_get_wiphy(ap->netdev);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
uint32_t hidden_ssid = NL80211_HIDDEN_SSID_NOT_IN_USE;
eapol,ap: Remove assumption of single cipher in authenticator IE Allow the user of the eapol_sm & handshake_state APIs to have multiple pairwise ciphers listed in the authenticator IE.
2021-02-01 14:40:42 +01:00
unsigned int nl_ciphers_cnt = __builtin_popcount(ap->ciphers);
uint32_t nl_ciphers[nl_ciphers_cnt];
ap: Set the group cipher when sending START_AP Seems this was overlooked because an initial version of ap.c didn't have group traffic support.
2021-02-01 14:40:41 +01:00
uint32_t group_nl_cipher =
ie_rsn_cipher_suite_to_cipher(ap->group_cipher);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
uint32_t nl_akm = CRYPTO_AKM_PSK;
uint32_t wpa_version = NL80211_WPA_VERSION_2;
uint32_t auth_type = NL80211_AUTHTYPE_OPEN_SYSTEM;
eapol,ap: Remove assumption of single cipher in authenticator IE Allow the user of the eapol_sm & handshake_state APIs to have multiple pairwise ciphers listed in the authenticator IE.
2021-02-01 14:40:42 +01:00
unsigned int i;
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
static const uint8_t bcast_addr[6] = {
0xff, 0xff, 0xff, 0xff, 0xff, 0xff
};
ap: include PROBE_RESP attribute if required by driver Certain drivers support/require probe response offloading which IWD did not check for or properly handle. If probe response offloading is required the probe response frame watch will not be added and instead the ATTR_PROBE_RESP will be included with START_AP. The head/tail builders were reused but slightly modified to check if the probe request frame is NULL, since it will be for use with START_AP.
2022-09-28 21:48:52 +02:00
static const uint8_t zero_addr[6] = { 0 };
eapol,ap: Remove assumption of single cipher in authenticator IE Allow the user of the eapol_sm & handshake_state APIs to have multiple pairwise ciphers listed in the authenticator IE.
2021-02-01 14:40:42 +01:00
for (i = 0, nl_ciphers_cnt = 0; i < 8; i++)
if (ap->ciphers & (1 << i))
nl_ciphers[nl_ciphers_cnt++] =
ie_rsn_cipher_suite_to_cipher(1 << i);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
head_len = ap_build_beacon_pr_head(ap, MPDU_MANAGEMENT_SUBTYPE_BEACON,
ap: update to use new TLV builder APIs AP still relies on the get_data/set_length semantics. Its more convenient to still use these since it avoids the need for extra temporary buffers when building the rates IE.
2019-04-23 19:47:26 +02:00
bcast_addr, head, sizeof(head));
ap: Write extra frame IEs from the user Add an API for the ap.h users to add extra IEs to outgoing management frames: beacons, etc.
2021-03-11 14:07:23 +01:00
tail_len = ap_build_beacon_pr_tail(ap, MPDU_MANAGEMENT_SUBTYPE_BEACON,
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
NULL, 0, tail, tail_len);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
if (!head_len || !tail_len)
return NULL;
ap: Use EAPoL over NL80211 if available
2018-05-01 17:47:10 +02:00
cmd = l_genl_msg_new_sized(NL80211_CMD_START_AP, 256 + head_len +
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
tail_len + strlen(ap->ssid));
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
/* SET_BEACON attrs */
l_genl_msg_append_attr(cmd, NL80211_ATTR_BEACON_HEAD, head_len, head);
l_genl_msg_append_attr(cmd, NL80211_ATTR_BEACON_TAIL, tail_len, tail);
l_genl_msg_append_attr(cmd, NL80211_ATTR_IE, 0, "");
l_genl_msg_append_attr(cmd, NL80211_ATTR_IE_PROBE_RESP, 0, "");
l_genl_msg_append_attr(cmd, NL80211_ATTR_IE_ASSOC_RESP, 0, "");
/* START_AP attrs */
l_genl_msg_append_attr(cmd, NL80211_ATTR_BEACON_INTERVAL, 4,
&ap->beacon_interval);
l_genl_msg_append_attr(cmd, NL80211_ATTR_DTIM_PERIOD, 4, &dtim_period);
l_genl_msg_append_attr(cmd, NL80211_ATTR_IFINDEX, 4, &ifindex);
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
l_genl_msg_append_attr(cmd, NL80211_ATTR_SSID, strlen(ap->ssid),
ap->ssid);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
l_genl_msg_append_attr(cmd, NL80211_ATTR_HIDDEN_SSID, 4,
&hidden_ssid);
eapol,ap: Remove assumption of single cipher in authenticator IE Allow the user of the eapol_sm & handshake_state APIs to have multiple pairwise ciphers listed in the authenticator IE.
2021-02-01 14:40:42 +01:00
l_genl_msg_append_attr(cmd, NL80211_ATTR_CIPHER_SUITES_PAIRWISE,
nl_ciphers_cnt * 4, nl_ciphers);
ap: Set the group cipher when sending START_AP Seems this was overlooked because an initial version of ap.c didn't have group traffic support.
2021-02-01 14:40:41 +01:00
l_genl_msg_append_attr(cmd, NL80211_ATTR_CIPHER_SUITE_GROUP, 4,
&group_nl_cipher);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
l_genl_msg_append_attr(cmd, NL80211_ATTR_WPA_VERSIONS, 4, &wpa_version);
l_genl_msg_append_attr(cmd, NL80211_ATTR_AKM_SUITES, 4, &nl_akm);
l_genl_msg_append_attr(cmd, NL80211_ATTR_AUTH_TYPE, 4, &auth_type);
ap: generate chandef for starting AP To include HT support a chandef needs to be created for whatever frequency is being used. This allows IWD to provide a secondary channel to the kernel in the case of 40MHz operation. Now the AP will generate a chandef when starting based on the channel set in the user profile (or default). If HT is not supported the chandef width is set to 20MHz no-HT, otherwise band_freq_to_ht_chandef is used.
2022-12-28 23:32:30 +01:00
l_genl_msg_append_attr(cmd, NL80211_ATTR_WIPHY_FREQ, 4,
&ap->chandef.frequency);
l_genl_msg_append_attr(cmd, NL80211_ATTR_CHANNEL_WIDTH, 4,
&ap->chandef.channel_width);
if (ap->chandef.center1_frequency)
l_genl_msg_append_attr(cmd, NL80211_ATTR_CENTER_FREQ1, 4,
&ap->chandef.center1_frequency);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
ap: include PROBE_RESP attribute if required by driver Certain drivers support/require probe response offloading which IWD did not check for or properly handle. If probe response offloading is required the probe response frame watch will not be added and instead the ATTR_PROBE_RESP will be included with START_AP. The head/tail builders were reused but slightly modified to check if the probe request frame is NULL, since it will be for use with START_AP.
2022-09-28 21:48:52 +02:00
if (wiphy_supports_probe_resp_offload(wiphy)) {
uint8_t probe_resp[head_len + tail_len];
uint8_t *ptr = probe_resp;
ptr += ap_build_beacon_pr_head(ap,
MPDU_MANAGEMENT_SUBTYPE_PROBE_RESPONSE,
zero_addr, ptr, sizeof(probe_resp));
ptr += ap_build_beacon_pr_tail(ap,
MPDU_MANAGEMENT_SUBTYPE_PROBE_RESPONSE,
ap: support setting country IE This adds a builder which sets the country IE in probes/beacons. The IE will use the 'single subband triplet sequence' meaning dot11OperatingClassesRequired is false. This is much easier to build and doesn't require knowing an operating class. The IE itself is variable in length and potentially could grow large if the hardware has a weird configuration (many different power levels or segmentation in supported channels) so the overall builder was changed to take the length of the buffer and warnings will be printed if any space issues are encountered.
2022-12-30 23:12:16 +01:00
NULL, 0, ptr, sizeof(probe_resp) -
(ptr - probe_resp));
ap: include PROBE_RESP attribute if required by driver Certain drivers support/require probe response offloading which IWD did not check for or properly handle. If probe response offloading is required the probe response frame watch will not be added and instead the ATTR_PROBE_RESP will be included with START_AP. The head/tail builders were reused but slightly modified to check if the probe request frame is NULL, since it will be for use with START_AP.
2022-09-28 21:48:52 +02:00
l_genl_msg_append_attr(cmd, NL80211_ATTR_PROBE_RESP,
ptr - probe_resp, probe_resp);
}
wiphy: Rename get_ext_feature API to has_ext_feature
2018-05-24 20:12:36 +02:00
if (wiphy_has_ext_feature(wiphy,
ap: Use EAPoL over NL80211 if available
2018-05-01 17:47:10 +02:00
NL80211_EXT_FEATURE_CONTROL_PORT_OVER_NL80211)) {
l_genl_msg_append_attr(cmd, NL80211_ATTR_SOCKET_OWNER, 0, NULL);
l_genl_msg_append_attr(cmd,
NL80211_ATTR_CONTROL_PORT_OVER_NL80211,
0, NULL);
}
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
return cmd;
}
ap: Move sending CMD_START_AP to common function
2021-05-10 10:57:56 +02:00
static bool ap_start_send(struct ap_state *ap)
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
{
ap: Move sending CMD_START_AP to common function
2021-05-10 10:57:56 +02:00
struct l_genl_msg *cmd = ap_build_cmd_start_ap(ap);
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
ap: Move sending CMD_START_AP to common function
2021-05-10 10:57:56 +02:00
if (!cmd) {
l_error("ap_build_cmd_start_ap failed");
return false;
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
}
ap->start_stop_cmd_id = l_genl_family_send(ap->nl80211, cmd,
ap_start_cb, ap, NULL);
if (!ap->start_stop_cmd_id) {
ap: Move sending CMD_START_AP to common function
2021-05-10 10:57:56 +02:00
l_error("AP_START l_genl_family_send failed");
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
l_genl_msg_unref(cmd);
ap: Move sending CMD_START_AP to common function
2021-05-10 10:57:56 +02:00
return false;
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
}
ap: Move sending CMD_START_AP to common function
2021-05-10 10:57:56 +02:00
return true;
}
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
ap: Move sending CMD_START_AP to common function
2021-05-10 10:57:56 +02:00
static void ap_ifaddr4_added_cb(int error, uint16_t type, const void *data,
uint32_t len, void *user_data)
{
struct ap_state *ap = user_data;
ap->rtnl_add_cmd = 0;
if (error) {
l_error("Failed to set IP address");
ap: Send a specific error message on async AP start failure We generate the DBus error reply type from the errno only when ap_start() was failing synchronously, now also send the errno through the callbacks so that we can also return a specific DBus reply when failing asynchronously. Thea AP autotest relies on receiving the AlreadyExists DBus error.
2021-05-28 02:47:34 +02:00
ap_start_failed(ap, error);
ap: Move sending CMD_START_AP to common function
2021-05-10 10:57:56 +02:00
return;
}
if (!ap_start_send(ap))
ap: Send a specific error message on async AP start failure We generate the DBus error reply type from the errno only when ap_start() was failing synchronously, now also send the errno through the callbacks so that we can also return a specific DBus reply when failing asynchronously. Thea AP autotest relies on receiving the AlreadyExists DBus error.
2021-05-28 02:47:34 +02:00
ap_start_failed(ap, -EIO);
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
}
ap: rename ap_parse_ie to be more descriptive This is only used for NEW_STATION parsing so name appropriately
2021-01-29 19:47:24 +01:00
static bool ap_parse_new_station_ies(const void *data, uint16_t len,
uint8_t **rsn_out,
struct l_uintset **rates_out)
ap: handle NEW_STATION for fullmac cards Since fullmac cards handle auth/assoc in firmware IWD must react differently while in AP mode just as it does in station. For fullmac cards a NEW_STATION event is emitted post association and from here the 4-way handshake can begin. In this NEW_STATION handler a new sta_state is created and the needed members are set in order to inject us back into the normal code execution for softmac post association (i.e. creating group keys and starting the 4-way handshake). From here everything works the same as softmac.
2021-01-28 19:33:51 +01:00
{
struct ie_tlv_iter iter;
uint8_t *rsn = NULL;
struct l_uintset *rates = NULL;
ie_tlv_iter_init(&iter, data, len);
while (ie_tlv_iter_next(&iter)) {
switch (ie_tlv_iter_get_tag(&iter)) {
case IE_TYPE_RSN:
ap: Fix cleanup on ap_parse_new_station_ies errors
2021-01-30 00:27:37 +01:00
if (rsn || ie_parse_rsne(&iter, NULL) < 0)
ap: handle NEW_STATION for fullmac cards Since fullmac cards handle auth/assoc in firmware IWD must react differently while in AP mode just as it does in station. For fullmac cards a NEW_STATION event is emitted post association and from here the 4-way handshake can begin. In this NEW_STATION handler a new sta_state is created and the needed members are set in order to inject us back into the normal code execution for softmac post association (i.e. creating group keys and starting the 4-way handshake). From here everything works the same as softmac.
2021-01-28 19:33:51 +01:00
goto parse_error;
rsn = l_memdup(ie_tlv_iter_get_data(&iter) - 2,
ap: fix off by one error The RSN element was not being allocated properly which caused an invalid read in some cases.
2021-01-29 21:59:50 +01:00
ie_tlv_iter_get_length(&iter) + 2);
ap: handle NEW_STATION for fullmac cards Since fullmac cards handle auth/assoc in firmware IWD must react differently while in AP mode just as it does in station. For fullmac cards a NEW_STATION event is emitted post association and from here the 4-way handshake can begin. In this NEW_STATION handler a new sta_state is created and the needed members are set in order to inject us back into the normal code execution for softmac post association (i.e. creating group keys and starting the 4-way handshake). From here everything works the same as softmac.
2021-01-28 19:33:51 +01:00
break;
case IE_TYPE_EXTENDED_SUPPORTED_RATES:
ap: Fix cleanup on ap_parse_new_station_ies errors
2021-01-30 00:27:37 +01:00
if (rates || ap_parse_supported_rates(&iter, &rates) <
0)
ap: handle NEW_STATION for fullmac cards Since fullmac cards handle auth/assoc in firmware IWD must react differently while in AP mode just as it does in station. For fullmac cards a NEW_STATION event is emitted post association and from here the 4-way handshake can begin. In this NEW_STATION handler a new sta_state is created and the needed members are set in order to inject us back into the normal code execution for softmac post association (i.e. creating group keys and starting the 4-way handshake). From here everything works the same as softmac.
2021-01-28 19:33:51 +01:00
goto parse_error;
break;
}
}
*rsn_out = rsn;
ap: remove rates requirement for fmac cards It was seen that some full mac cards/drivers do not include any rate information with the NEW_STATION event. This was causing the NEW_STATION event to be ignored, preventing AP mode from working on these cards. Since the full mac path does not even require sta->rates the parsing can be removed completely.
2021-04-12 20:22:45 +02:00
if (rates_out)
*rates_out = rates;
ap: Fix possible memleak of 'rates'
2021-04-30 18:38:04 +02:00
else
l_uintset_free(rates);
ap: handle NEW_STATION for fullmac cards Since fullmac cards handle auth/assoc in firmware IWD must react differently while in AP mode just as it does in station. For fullmac cards a NEW_STATION event is emitted post association and from here the 4-way handshake can begin. In this NEW_STATION handler a new sta_state is created and the needed members are set in order to inject us back into the normal code execution for softmac post association (i.e. creating group keys and starting the 4-way handshake). From here everything works the same as softmac.
2021-01-28 19:33:51 +01:00
return true;
parse_error:
if (rsn)
l_free(rsn);
if (rates)
l_uintset_free(rates);
return false;
}
ap: add DEL_STATION path to fullmac AP handling This is how a fullmac card tells userspace that a station has left. This fixes the issue where the same client cannot re-connect to the same AP multiple times. ap_new_station was renamed to ap_handle_new_station for consistency.
2021-01-29 18:28:18 +01:00
static void ap_handle_new_station(struct ap_state *ap, struct l_genl_msg *msg)
ap: handle NEW_STATION for fullmac cards Since fullmac cards handle auth/assoc in firmware IWD must react differently while in AP mode just as it does in station. For fullmac cards a NEW_STATION event is emitted post association and from here the 4-way handshake can begin. In this NEW_STATION handler a new sta_state is created and the needed members are set in order to inject us back into the normal code execution for softmac post association (i.e. creating group keys and starting the 4-way handshake). From here everything works the same as softmac.
2021-01-28 19:33:51 +01:00
{
struct sta_state *sta;
struct l_genl_attr attr;
uint16_t type;
uint16_t len;
const void *data;
uint8_t mac[6];
uint8_t *assoc_rsne = NULL;
if (!l_genl_attr_init(&attr, msg))
return;
while (l_genl_attr_next(&attr, &type, &len, &data)) {
switch (type) {
case NL80211_ATTR_IE:
ap: remove rates requirement for fmac cards It was seen that some full mac cards/drivers do not include any rate information with the NEW_STATION event. This was causing the NEW_STATION event to be ignored, preventing AP mode from working on these cards. Since the full mac path does not even require sta->rates the parsing can be removed completely.
2021-04-12 20:22:45 +02:00
if (assoc_rsne)
ap: Fix cleanup on ap_parse_new_station_ies errors
2021-01-30 00:27:37 +01:00
goto cleanup;
ap: rename ap_parse_ie to be more descriptive This is only used for NEW_STATION parsing so name appropriately
2021-01-29 19:47:24 +01:00
if (!ap_parse_new_station_ies(data, len, &assoc_rsne,
ap: remove rates requirement for fmac cards It was seen that some full mac cards/drivers do not include any rate information with the NEW_STATION event. This was causing the NEW_STATION event to be ignored, preventing AP mode from working on these cards. Since the full mac path does not even require sta->rates the parsing can be removed completely.
2021-04-12 20:22:45 +02:00
NULL))
ap: handle NEW_STATION for fullmac cards Since fullmac cards handle auth/assoc in firmware IWD must react differently while in AP mode just as it does in station. For fullmac cards a NEW_STATION event is emitted post association and from here the 4-way handshake can begin. In this NEW_STATION handler a new sta_state is created and the needed members are set in order to inject us back into the normal code execution for softmac post association (i.e. creating group keys and starting the 4-way handshake). From here everything works the same as softmac.
2021-01-28 19:33:51 +01:00
return;
break;
case NL80211_ATTR_MAC:
if (len != 6)
ap: Fix cleanup on ap_parse_new_station_ies errors
2021-01-30 00:27:37 +01:00
goto cleanup;
ap: handle NEW_STATION for fullmac cards Since fullmac cards handle auth/assoc in firmware IWD must react differently while in AP mode just as it does in station. For fullmac cards a NEW_STATION event is emitted post association and from here the 4-way handshake can begin. In this NEW_STATION handler a new sta_state is created and the needed members are set in order to inject us back into the normal code execution for softmac post association (i.e. creating group keys and starting the 4-way handshake). From here everything works the same as softmac.
2021-01-28 19:33:51 +01:00
memcpy(mac, data, 6);
break;
}
}
ap: remove rates requirement for fmac cards It was seen that some full mac cards/drivers do not include any rate information with the NEW_STATION event. This was causing the NEW_STATION event to be ignored, preventing AP mode from working on these cards. Since the full mac path does not even require sta->rates the parsing can be removed completely.
2021-04-12 20:22:45 +02:00
if (!assoc_rsne)
ap: Fix cleanup on ap_parse_new_station_ies errors
2021-01-30 00:27:37 +01:00
goto cleanup;
ap: handle NEW_STATION for fullmac cards Since fullmac cards handle auth/assoc in firmware IWD must react differently while in AP mode just as it does in station. For fullmac cards a NEW_STATION event is emitted post association and from here the 4-way handshake can begin. In this NEW_STATION handler a new sta_state is created and the needed members are set in order to inject us back into the normal code execution for softmac post association (i.e. creating group keys and starting the 4-way handshake). From here everything works the same as softmac.
2021-01-28 19:33:51 +01:00
/*
* Softmac's should already have a station created. The above check
* may also fail for softmac cards.
*/
sta = l_queue_find(ap->sta_states, ap_sta_match_addr, mac);
ap: Fix cleanup on ap_parse_new_station_ies errors
2021-01-30 00:27:37 +01:00
if (sta)
goto cleanup;
ap: handle NEW_STATION for fullmac cards Since fullmac cards handle auth/assoc in firmware IWD must react differently while in AP mode just as it does in station. For fullmac cards a NEW_STATION event is emitted post association and from here the 4-way handshake can begin. In this NEW_STATION handler a new sta_state is created and the needed members are set in order to inject us back into the normal code execution for softmac post association (i.e. creating group keys and starting the 4-way handshake). From here everything works the same as softmac.
2021-01-28 19:33:51 +01:00
sta = l_new(struct sta_state, 1);
memcpy(sta->addr, mac, 6);
sta->ap = ap;
sta->assoc_rsne = assoc_rsne;
sta->aid = ++ap->last_aid;
sta->associated = true;
if (!ap->sta_states)
ap->sta_states = l_queue_new();
l_queue_push_tail(ap->sta_states, sta);
ap: build HT Capabilities/Operation elements If supported this will include the HT capabilities and HT operations elements in beacons/probes. Some shortcuts were taken here since not all the information is currently parsed from the hardware. Namely the HT operation element does not include the basic MCS set. Still, this will at least show stations that the AP is capable of more than just basic rates. The builders themselves are structured similar to the basic rates builder where they build only the contents and return the length. The caller must set the type/length manually. This is to support the two use cases of using with an IE builder vs direct pointer.
2022-12-28 23:32:31 +01:00
if (ap->supports_ht)
ap_update_beacon(ap);
ap: handle NEW_STATION for fullmac cards Since fullmac cards handle auth/assoc in firmware IWD must react differently while in AP mode just as it does in station. For fullmac cards a NEW_STATION event is emitted post association and from here the 4-way handshake can begin. In this NEW_STATION handler a new sta_state is created and the needed members are set in order to inject us back into the normal code execution for softmac post association (i.e. creating group keys and starting the 4-way handshake). From here everything works the same as softmac.
2021-01-28 19:33:51 +01:00
msg = nl80211_build_set_station_unauthorized(
netdev_get_ifindex(ap->netdev), mac);
if (!l_genl_family_send(ap->nl80211, msg, ap_associate_sta_cb,
sta, NULL)) {
l_genl_msg_unref(msg);
l_error("Issuing SET_STATION failed");
ap_del_station(sta, MMPDU_REASON_CODE_UNSPECIFIED, true);
}
ap: Fix cleanup on ap_parse_new_station_ies errors
2021-01-30 00:27:37 +01:00
return;
cleanup:
l_free(assoc_rsne);
ap: handle NEW_STATION for fullmac cards Since fullmac cards handle auth/assoc in firmware IWD must react differently while in AP mode just as it does in station. For fullmac cards a NEW_STATION event is emitted post association and from here the 4-way handshake can begin. In this NEW_STATION handler a new sta_state is created and the needed members are set in order to inject us back into the normal code execution for softmac post association (i.e. creating group keys and starting the 4-way handshake). From here everything works the same as softmac.
2021-01-28 19:33:51 +01:00
}
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
ap: add DEL_STATION path to fullmac AP handling This is how a fullmac card tells userspace that a station has left. This fixes the issue where the same client cannot re-connect to the same AP multiple times. ap_new_station was renamed to ap_handle_new_station for consistency.
2021-01-29 18:28:18 +01:00
static void ap_handle_del_station(struct ap_state *ap, struct l_genl_msg *msg)
{
struct l_genl_attr attr;
uint16_t type;
uint16_t len;
const void *data;
uint8_t mac[6];
uint16_t reason = MMPDU_REASON_CODE_UNSPECIFIED;
if (!l_genl_attr_init(&attr, msg))
return;
while (l_genl_attr_next(&attr, &type, &len, &data)) {
switch (type) {
case NL80211_ATTR_MAC:
if (len != 6)
return;
memcpy(mac, data, 6);
break;
case NL80211_ATTR_REASON_CODE:
if (len != 2)
return;
reason = l_get_u16(data);
}
}
ap: Make station removal safer Replace instances of the ap_del_station() + ap_sta_free()/ap_remove_sta() with calls to ap_station_disconnect to make sure we consistently remove the station from the ap->sta_states queue before using ap_del_station(). ap_del_station() may generate an event to the ap.h API user (e.g. P2P) and this may end up tearing down the AP completely. For that scenario we also don't want ap_sta_free() to access sta->ap so we make sure ap_del_station() performs these cleanup steps so that ap_sta_free() has nothing to do that accesses sta->ap.
2021-08-07 04:10:42 +02:00
ap_station_disconnect(ap, mac, reason);
ap: add DEL_STATION path to fullmac AP handling This is how a fullmac card tells userspace that a station has left. This fixes the issue where the same client cannot re-connect to the same AP multiple times. ap_new_station was renamed to ap_handle_new_station for consistency.
2021-01-29 18:28:18 +01:00
}
ap: React to NL80211_CMD_STOP_AP events These events will tell use when our AP gets stopped without our request, for example due to suspend/resume.
2020-02-15 01:17:46 +01:00
static void ap_mlme_notify(struct l_genl_msg *msg, void *user_data)
{
struct ap_state *ap = user_data;
uint32_t ifindex;
if (nl80211_parse_attrs(msg, NL80211_ATTR_IFINDEX, &ifindex,
NL80211_ATTR_UNSPEC) < 0 ||
ifindex != netdev_get_ifindex(ap->netdev))
return;
switch (l_genl_msg_get_command(msg)) {
case NL80211_CMD_STOP_AP:
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
ap->in_event = true;
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
if (ap->start_stop_cmd_id) {
ap: Send a specific error message on async AP start failure We generate the DBus error reply type from the errno only when ap_start() was failing synchronously, now also send the errno through the callbacks so that we can also return a specific DBus reply when failing asynchronously. Thea AP autotest relies on receiving the AlreadyExists DBus error.
2021-05-28 02:47:34 +02:00
struct ap_event_start_failed_data data = { -ECANCELED };
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
l_genl_family_cancel(ap->nl80211,
ap->start_stop_cmd_id);
ap->start_stop_cmd_id = 0;
ap: Send a specific error message on async AP start failure We generate the DBus error reply type from the errno only when ap_start() was failing synchronously, now also send the errno through the callbacks so that we can also return a specific DBus reply when failing asynchronously. Thea AP autotest relies on receiving the AlreadyExists DBus error.
2021-05-28 02:47:34 +02:00
ap->ops->handle_event(AP_EVENT_START_FAILED, &data,
ap: Pass "ops" struct to ap_start() Pass the event callback function pointer in a "struct ap_ops" instead of as individual ap_start() argument to make adding new callbacks easier.
2020-09-16 11:17:49 +02:00
ap->user_data);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
} else if (ap->started) {
ap->started = false;
ap: Pass "ops" struct to ap_start() Pass the event callback function pointer in a "struct ap_ops" instead of as individual ap_start() argument to make adding new callbacks easier.
2020-09-16 11:17:49 +02:00
ap->ops->handle_event(AP_EVENT_STOPPING, NULL,
ap->user_data);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
}
ap: React to NL80211_CMD_STOP_AP events These events will tell use when our AP gets stopped without our request, for example due to suspend/resume.
2020-02-15 01:17:46 +01:00
ap_reset(ap);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
l_genl_family_free(ap->nl80211);
l_free(ap);
ap: React to NL80211_CMD_STOP_AP events These events will tell use when our AP gets stopped without our request, for example due to suspend/resume.
2020-02-15 01:17:46 +01:00
break;
ap: handle NEW_STATION for fullmac cards Since fullmac cards handle auth/assoc in firmware IWD must react differently while in AP mode just as it does in station. For fullmac cards a NEW_STATION event is emitted post association and from here the 4-way handshake can begin. In this NEW_STATION handler a new sta_state is created and the needed members are set in order to inject us back into the normal code execution for softmac post association (i.e. creating group keys and starting the 4-way handshake). From here everything works the same as softmac.
2021-01-28 19:33:51 +01:00
case NL80211_CMD_NEW_STATION:
ap: add DEL_STATION path to fullmac AP handling This is how a fullmac card tells userspace that a station has left. This fixes the issue where the same client cannot re-connect to the same AP multiple times. ap_new_station was renamed to ap_handle_new_station for consistency.
2021-01-29 18:28:18 +01:00
ap_handle_new_station(ap, msg);
break;
case NL80211_CMD_DEL_STATION:
ap_handle_del_station(ap, msg);
ap: handle NEW_STATION for fullmac cards Since fullmac cards handle auth/assoc in firmware IWD must react differently while in AP mode just as it does in station. For fullmac cards a NEW_STATION event is emitted post association and from here the 4-way handshake can begin. In this NEW_STATION handler a new sta_state is created and the needed members are set in order to inject us back into the normal code execution for softmac post association (i.e. creating group keys and starting the 4-way handshake). From here everything works the same as softmac.
2021-01-28 19:33:51 +01:00
break;
ap: React to NL80211_CMD_STOP_AP events These events will tell use when our AP gets stopped without our request, for example due to suspend/resume.
2020-02-15 01:17:46 +01:00
}
}
ap: Add MACs to FILS IP Assignment responses Try to include the gateway and DNS MAC addresses in the corresponding fields in the FILS IP Address Assignment IEs we send to the clients.
2021-08-28 05:23:05 +02:00
static void ap_get_gateway4_mac_cb(int error, const uint8_t *hwaddr,
size_t hwaddr_len, void *user_data)
{
struct ap_state *ap = user_data;
ap->rtnl_get_gateway4_mac_cmd = 0;
if (error) {
l_debug("Error: %s (%i)", strerror(-error), -error);
return;
}
if (L_WARN_ON(unlikely(hwaddr_len != 6)))
return;
l_debug("Resolved mac to " MAC, MAC_STR(hwaddr));
memcpy(ap->netconfig_gateway4_mac, hwaddr, 6);
}
static void ap_get_dns4_mac_cb(int error, const uint8_t *hwaddr,
size_t hwaddr_len, void *user_data)
{
struct ap_state *ap = user_data;
ap->rtnl_get_dns4_mac_cmd = 0;
if (error) {
l_debug("Error: %s (%i)", strerror(-error), -error);
return;
}
if (L_WARN_ON(unlikely(hwaddr_len != 6)))
return;
l_debug("Resolved mac to " MAC, MAC_STR(hwaddr));
memcpy(ap->netconfig_dns4_mac, hwaddr, 6);
}
static void ap_query_macs(struct ap_state *ap, const char *addr_str,
uint8_t prefix_len, const char *gateway_str,
const char **dns_str_list)
{
uint32_t local = IP4_FROM_STR(addr_str);
uint32_t gateway = 0;
uint32_t dns = 0;
uint32_t ifindex = netdev_get_ifindex(ap->netdev);
/*
* For simplicity only check the ARP/NDP tables to see if we already
* have the MACs that we need. There doesn't seem to be an API to
* actually resolve addresses that are not in these tables other than
* by triggering IP traffic to those hosts, such as a ping. In a PC
* or mobile device scenario we're likely to have these MACs already,
* otherwise we give up as this is a pretty low-priority feature.
*/
if (gateway_str) {
gateway = IP4_FROM_STR(gateway_str);
if (L_WARN_ON(unlikely(!gateway)))
return;
if (gateway == local)
memcpy(ap->netconfig_gateway4_mac,
netdev_get_address(ap->netdev), 6);
else {
ap->rtnl_get_gateway4_mac_cmd =
l_rtnl_neighbor_get_hwaddr(rtnl, ifindex,
AF_INET, &gateway,
ap_get_gateway4_mac_cb,
ap, NULL);
if (!ap->rtnl_get_gateway4_mac_cmd)
l_debug("l_rtnl_neighbor_get_hwaddr() failed "
"for the gateway IP");
}
}
if (dns_str_list) {
dns = IP4_FROM_STR(dns_str_list[0]);
if (L_WARN_ON(unlikely(!dns)))
return;
/* TODO: can also skip query if dns == gateway */
if (dns == local)
memcpy(ap->netconfig_dns4_mac,
netdev_get_address(ap->netdev), 6);
else if (util_ip_subnet_match(prefix_len, &dns, &local)) {
ap->rtnl_get_dns4_mac_cmd =
l_rtnl_neighbor_get_hwaddr(rtnl, ifindex,
AF_INET, &dns,
ap_get_dns4_mac_cb,
ap, NULL);
if (!ap->rtnl_get_dns4_mac_cmd)
l_debug("l_rtnl_neighbor_get_hwaddr() failed "
"for the DNS IP");
}
}
}
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
#define AP_DEFAULT_IPV4_PREFIX_LEN 28
static int ap_setup_netconfig4(struct ap_state *ap, const char **addr_str_list,
uint8_t prefix_len, const char *gateway_str,
const char **ip_range,
const char **dns_str_list,
unsigned int lease_time)
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
{
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
uint32_t ifindex = netdev_get_ifindex(ap->netdev);
struct l_rtnl_address *existing_addr = ip_pool_get_addr4(ifindex);
struct l_rtnl_address *new_addr = NULL;
int ret;
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
struct in_addr ia;
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
struct l_dhcp_server *dhcp = NULL;
bool r;
char addr_str_buf[INET_ADDRSTRLEN];
dhcp = l_dhcp_server_new(ifindex);
if (!dhcp) {
l_error("Failed to create DHCP server on ifindex %u", ifindex);
ret = -EIO;
goto cleanup;
}
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (getenv("IWD_DHCP_DEBUG"))
l_dhcp_server_set_debug(dhcp, do_debug,
"[DHCPv4 SERV] ", NULL);
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
/*
* The address pool specified for this AP (if any) has the priority,
* next is the address currently set on the interface (if any) and
* last is the global AP address pool (APRanges setting).
*/
if (addr_str_list) {
if (!prefix_len)
prefix_len = AP_DEFAULT_IPV4_PREFIX_LEN;
ret = ip_pool_select_addr4(addr_str_list, prefix_len,
&new_addr);
} else if (existing_addr &&
l_rtnl_address_get_prefix_length(existing_addr) <
31) {
if (!prefix_len)
prefix_len = l_rtnl_address_get_prefix_length(
existing_addr);
if (!l_rtnl_address_get_address(existing_addr, addr_str_buf)) {
ret = -EIO;
goto cleanup;
}
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
new_addr = l_rtnl_address_new(addr_str_buf, prefix_len);
ret = 0;
} else {
ap: Refactor global address pool loading Deprecate the global [General].APRanges setting in favour of [IPv4].APAddressPool with an extended (but backwards-compatible) syntax. Drop the existing address pool creation code. The new APAddressPool setting has the same syntax as the profile-local [IPv4].Address setting and the subnet selection code will fall back to the global setting if it's missing, this way we use common code to handle both settings.
2021-05-28 02:47:33 +02:00
if (!prefix_len)
prefix_len = AP_DEFAULT_IPV4_PREFIX_LEN;
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
ap: Refactor global address pool loading Deprecate the global [General].APRanges setting in favour of [IPv4].APAddressPool with an extended (but backwards-compatible) syntax. Drop the existing address pool creation code. The new APAddressPool setting has the same syntax as the profile-local [IPv4].Address setting and the subnet selection code will fall back to the global setting if it's missing, this way we use common code to handle both settings.
2021-05-28 02:47:33 +02:00
ret = ip_pool_select_addr4((const char **) global_addr4_strs,
prefix_len, &new_addr);
ap: Print error messages in dhcp_load_settings
2021-02-26 17:17:24 +01:00
}
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (ret)
goto cleanup;
ret = -EIO;
/*
* l_dhcp_server_start() would retrieve the current IPv4 from
* the interface but set it anyway in case there are multiple
* addresses, saves one ioctl too.
*/
if (!l_rtnl_address_get_address(new_addr, addr_str_buf)) {
l_error("l_rtnl_address_get_address failed");
goto cleanup;
ap: Print error messages in dhcp_load_settings
2021-02-26 17:17:24 +01:00
}
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (!l_dhcp_server_set_ip_address(dhcp, addr_str_buf)) {
l_error("l_dhcp_server_set_ip_address failed");
goto cleanup;
}
ia.s_addr = htonl(util_netmask_from_prefix(prefix_len));
if (!l_dhcp_server_set_netmask(dhcp, inet_ntoa(ia))) {
l_error("l_dhcp_server_set_netmask failed");
goto cleanup;
}
if (gateway_str && !l_dhcp_server_set_gateway(dhcp, gateway_str)) {
l_error("l_dhcp_server_set_gateway failed");
goto cleanup;
ap: Print error messages in dhcp_load_settings
2021-02-26 17:17:24 +01:00
}
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Print error messages in dhcp_load_settings
2021-02-26 17:17:24 +01:00
if (ip_range) {
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
r = l_dhcp_server_set_ip_range(dhcp, ip_range[0], ip_range[1]);
if (!r) {
l_error("l_dhcp_server_set_ip_range failed");
goto cleanup;
ap: Print error messages in dhcp_load_settings
2021-02-26 17:17:24 +01:00
}
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
}
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (dns_str_list) {
r = l_dhcp_server_set_dns(dhcp, (char **) dns_str_list);
if (!r) {
l_error("l_dhcp_server_set_dns failed");
goto cleanup;
ap: Print error messages in dhcp_load_settings
2021-02-26 17:17:24 +01:00
}
}
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (lease_time && !l_dhcp_server_set_lease_time(dhcp, lease_time)) {
l_error("l_dhcp_server_set_lease_time failed");
goto cleanup;
ap: Print error messages in dhcp_load_settings
2021-02-26 17:17:24 +01:00
}
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
ap->netconfig_set_addr4 = true;
ap: Save AP address as l_rtnl_address Change the char *addr_str and uint8_t prefix_len pair to an l_rtnl_address object and use ell/rtnl.h utilities that use that directly. Extend broadcast_from_ip to handle prefix_len.
2021-05-28 02:47:35 +02:00
ap->netconfig_addr4 = l_steal_ptr(new_addr);
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
ap->netconfig_dhcp = l_steal_ptr(dhcp);
ret = 0;
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (existing_addr && l_rtnl_address_get_prefix_length(existing_addr) >
prefix_len) {
char addr_str_buf2[INET_ADDRSTRLEN];
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (l_rtnl_address_get_address(existing_addr, addr_str_buf2) &&
!strcmp(addr_str_buf, addr_str_buf2))
ap->netconfig_set_addr4 = false;
}
ap: Add MACs to FILS IP Assignment responses Try to include the gateway and DNS MAC addresses in the corresponding fields in the FILS IP Address Assignment IEs we send to the clients.
2021-08-28 05:23:05 +02:00
ap_query_macs(ap, addr_str_buf, prefix_len, gateway_str, dns_str_list);
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
cleanup:
l_dhcp_server_destroy(dhcp);
l_rtnl_address_free(new_addr);
l_rtnl_address_free(existing_addr);
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
return ret;
}
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
static int ap_load_ipv4(struct ap_state *ap, const struct l_settings *config)
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
{
int ret = -EINVAL;
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
char **addr_str_list = NULL;
uint32_t static_addr = 0;
uint8_t prefix_len = 0;
char *gateway_str = NULL;
char **ip_range = NULL;
char **dns_str_list = NULL;
unsigned int lease_time = 0;
struct in_addr ia;
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
treewide: Parse EnableNetworkConfiguration in one place Add netconfig_enabled() and use that in all places that want to know whether network configuration is enabled. Drop the enable_network_config deprecated setting, which was only being handled in one of these 5 or so places.
2021-10-21 10:50:30 +02:00
if (!l_settings_has_group(config, "IPv4") || !netconfig_enabled())
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
return 0;
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (l_settings_has_key(config, "IPv4", "Address")) {
addr_str_list = l_settings_get_string_list(config, "IPv4",
"Address", ',');
if (!addr_str_list || !*addr_str_list) {
l_error("Can't parse the profile [IPv4].Address "
"setting as a string list");
goto done;
}
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
/* Check for the static IP syntax: Address=<IP> */
if (l_strv_length(addr_str_list) == 1 &&
inet_pton(AF_INET, *addr_str_list, &ia) == 1)
static_addr = ntohl(ia.s_addr);
}
if (l_settings_has_key(config, "IPv4", "Netmask")) {
L_AUTO_FREE_VAR(char *, netmask_str) =
l_settings_get_string(config, "IPv4", "Netmask");
if (inet_pton(AF_INET, netmask_str, &ia) != 1) {
l_error("Can't parse the profile [IPv4].Netmask "
"setting");
goto done;
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
}
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
prefix_len = __builtin_popcount(ia.s_addr);
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (ntohl(ia.s_addr) != util_netmask_from_prefix(prefix_len)) {
l_error("Invalid profile [IPv4].Netmask value");
goto done;
}
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
}
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (l_settings_has_key(config, "IPv4", "Gateway")) {
gateway_str = l_settings_get_string(config, "IPv4", "Gateway");
if (!gateway_str) {
l_error("Invalid profile [IPv4].Gateway value");
goto done;
}
}
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (l_settings_get_value(config, "IPv4", "IPRange")) {
int i;
uint32_t netmask;
uint8_t tmp_len = prefix_len ?: AP_DEFAULT_IPV4_PREFIX_LEN;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
ip_range = l_settings_get_string_list(config, "IPv4",
"IPRange", ',');
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (!static_addr) {
l_error("[IPv4].IPRange only makes sense in an AP "
"profile if a static local address has also "
"been specified");
goto done;
}
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (!ip_range || l_strv_length(ip_range) != 2) {
l_error("Can't parse the profile [IPv4].IPRange "
"setting as two address strings");
goto done;
}
ap: add StartProfile DBus method Users can now start an AP from settings based on a profile on disk. The only argument is the SSID which will be used to lookup the profile. If no profile is found a NotFound error will be returned. Any invalid profiles will result in an Invalid return.
2020-11-02 19:28:49 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
netmask = util_netmask_from_prefix(tmp_len);
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
for (i = 0; i < 2; i++) {
struct in_addr range_addr;
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Do not use inet_aton
2021-06-04 19:31:36 +02:00
if (inet_pton(AF_INET, ip_range[i], &range_addr) != 1) {
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
l_error("Can't parse address in "
"[IPv4].IPRange[%i]", i + 1);
goto done;
}
if ((static_addr ^ ntohl(range_addr.s_addr)) &
netmask) {
ia.s_addr = htonl(static_addr);
l_error("[IPv4].IPRange[%i] is not in the "
"%s/%i subnet", i + 1, inet_ntoa(ia),
tmp_len);
goto done;
}
ap: add StartProfile DBus method Users can now start an AP from settings based on a profile on disk. The only argument is the SSID which will be used to lookup the profile. If no profile is found a NotFound error will be returned. Any invalid profiles will result in an Invalid return.
2020-11-02 19:28:49 +01:00
}
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
}
ap: add StartProfile DBus method Users can now start an AP from settings based on a profile on disk. The only argument is the SSID which will be used to lookup the profile. If no profile is found a NotFound error will be returned. Any invalid profiles will result in an Invalid return.
2020-11-02 19:28:49 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (l_settings_has_key(config, "IPv4", "DNSList")) {
dns_str_list = l_settings_get_string_list(config, "IPv4",
"DNSList", ',');
if (!dns_str_list || !*dns_str_list) {
l_error("Can't parse the profile [IPv4].DNSList "
"setting as a string list");
goto done;
}
}
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (l_settings_has_key(config, "IPv4", "LeaseTime")) {
if (!l_settings_get_uint(config, "IPv4", "LeaseTime",
&lease_time) ||
lease_time < 1) {
l_error("Error parsing [IPv4].LeaseTime as a positive "
"integer");
goto done;
}
ap: add StartProfile DBus method Users can now start an AP from settings based on a profile on disk. The only argument is the SSID which will be used to lookup the profile. If no profile is found a NotFound error will be returned. Any invalid profiles will result in an Invalid return.
2020-11-02 19:28:49 +01:00
}
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
ret = ap_setup_netconfig4(ap, (const char **) addr_str_list, prefix_len,
gateway_str, (const char **) ip_range,
(const char **) dns_str_list,
lease_time);
done:
l_strv_free(addr_str_list);
l_free(gateway_str);
l_strv_free(ip_range);
l_strv_free(dns_str_list);
return ret;
ap: add StartProfile DBus method Users can now start an AP from settings based on a profile on disk. The only argument is the SSID which will be used to lookup the profile. If no profile is found a NotFound error will be returned. Any invalid profiles will result in an Invalid return.
2020-11-02 19:28:49 +01:00
}
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
static bool ap_load_psk(struct ap_state *ap, const struct l_settings *config)
{
L_AUTO_FREE_VAR(char *, passphrase) =
l_settings_get_string(config, "Security", "Passphrase");
int err;
if (passphrase) {
if (strlen(passphrase) > 63) {
l_error("AP [Security].Passphrase must not exceed "
"63 characters");
return false;
}
strcpy(ap->passphrase, passphrase);
}
if (l_settings_has_key(config, "Security", "PreSharedKey")) {
size_t psk_len;
L_AUTO_FREE_VAR(uint8_t *, psk) = l_settings_get_bytes(config,
"Security",
"PreSharedKey",
&psk_len);
if (!psk || psk_len != 32) {
l_error("AP [Security].PreSharedKey must be a 32-byte "
"hexstring");
return false;
}
memcpy(ap->psk, psk, 32);
return true;
}
if (!passphrase) {
l_error("AP requires at least one of [Security].PreSharedKey, "
"[Security].Passphrase to be present");
return false;
}
err = crypto_psk_from_passphrase(passphrase, (uint8_t *) ap->ssid,
strlen(ap->ssid), ap->psk);
if (err < 0) {
l_error("AP couldn't generate the PSK from given "
"[Security].Passphrase value: %s (%i)",
strerror(-err), -err);
return false;
}
return true;
}
ap: add profile settings PairwiseCiphers/GroupCipher These can now be optionally provided in an AP profile and provide a way to limit what ciphers can be chosen. This still is dependent on what the hardware supports.
2022-11-02 23:51:08 +01:00
/*
* Note: only PTK/GTK ciphers are supported here since this is all these are
* used for.
*/
static enum ie_rsn_cipher_suite ap_string_to_cipher(const char *str)
{
if (!strcmp(str, "UseGroupCipher"))
return IE_RSN_CIPHER_SUITE_USE_GROUP_CIPHER;
else if (!strcmp(str, "TKIP"))
return IE_RSN_CIPHER_SUITE_TKIP;
else if (!strcmp(str, "CCMP-128") || !strcmp(str, "CCMP"))
return IE_RSN_CIPHER_SUITE_CCMP;
else if (!strcmp(str, "GCMP-128") || !strcmp(str, "GCMP"))
return IE_RSN_CIPHER_SUITE_GCMP;
else if (!strcmp(str, "GCMP-256"))
return IE_RSN_CIPHER_SUITE_GCMP_256;
else if (!strcmp(str, "CCMP-256"))
return IE_RSN_CIPHER_SUITE_CCMP_256;
else
return 0;
}
ap: change PairwiseCiphers to string array A comma separated list as a string was ok for pure display purposes but if any processing needed to be done on these values by external consumers it really makes more sense to use a DBus array.
2022-11-04 17:37:58 +01:00
static char **ap_ciphers_to_strv(uint16_t ciphers)
ap: add PairwiseCiphers/GroupCipher to dbus interface Expose these values on the DBus interface so clients can view them.
2022-11-02 23:51:11 +01:00
{
uint16_t i;
char **list = l_strv_new();
for (i = 0; i < 16; i++) {
if (!(ciphers & (1 << i)))
continue;
list = l_strv_append(list,
ie_rsn_cipher_suite_to_string(1 << i));
}
ap: change PairwiseCiphers to string array A comma separated list as a string was ok for pure display purposes but if any processing needed to be done on these values by external consumers it really makes more sense to use a DBus array.
2022-11-04 17:37:58 +01:00
return list;
ap: add PairwiseCiphers/GroupCipher to dbus interface Expose these values on the DBus interface so clients can view them.
2022-11-02 23:51:11 +01:00
}
ap: add support for 5GHz frequencies in AP mode This enables IWD to use 5GHz frequencies in AP mode. Currently 6GHz is not supported so we can assume a [General].Channel value 36 or above indicates the 5GHz band. It should be noted that the system will probably need a regulatory domain set in order for 5GHz to be allowed in AP mode. This is due to world roaming (00) restricting any/all 5GHz frequencies. This can be accomplished by setting main.conf [General].Country=CC to the country this AP will operate in.
2022-12-09 20:39:28 +01:00
static bool ap_validate_band_channel(struct ap_state *ap)
{
struct wiphy *wiphy = netdev_get_wiphy(ap->netdev);
uint32_t freq;
ap: use wiphy_get_frequency_info Replace disabled/supported frequency list with the new wiphy_get_frequency_info()
2022-12-16 22:27:40 +01:00
const struct band_freq_attrs *attr;
ap: add support for 5GHz frequencies in AP mode This enables IWD to use 5GHz frequencies in AP mode. Currently 6GHz is not supported so we can assume a [General].Channel value 36 or above indicates the 5GHz band. It should be noted that the system will probably need a regulatory domain set in order for 5GHz to be allowed in AP mode. This is due to world roaming (00) restricting any/all 5GHz frequencies. This can be accomplished by setting main.conf [General].Country=CC to the country this AP will operate in.
2022-12-09 20:39:28 +01:00
if (!(wiphy_get_supported_bands(wiphy) & ap->band)) {
l_error("AP hardware does not support band");
return -EINVAL;
}
freq = band_channel_to_freq(ap->channel, ap->band);
if (!freq) {
l_error("AP invalid band (%s) and channel (%u) combination",
tree-wide: correct the spelling Ghz -> GHz This mispelling was present in the configuration, so I retained parsing of the legacy BandModifier*Ghz options for compatibility. Without this change anyone spelling GHz correctly in their configs would be very confused.
2023-11-08 02:33:54 +01:00
(ap->band & BAND_FREQ_5_GHZ) ? "5GHz" : "2.4GHz",
ap: add support for 5GHz frequencies in AP mode This enables IWD to use 5GHz frequencies in AP mode. Currently 6GHz is not supported so we can assume a [General].Channel value 36 or above indicates the 5GHz band. It should be noted that the system will probably need a regulatory domain set in order for 5GHz to be allowed in AP mode. This is due to world roaming (00) restricting any/all 5GHz frequencies. This can be accomplished by setting main.conf [General].Country=CC to the country this AP will operate in.
2022-12-09 20:39:28 +01:00
ap->channel);
return false;
}
ap: use wiphy_get_frequency_info Replace disabled/supported frequency list with the new wiphy_get_frequency_info()
2022-12-16 22:27:40 +01:00
attr = wiphy_get_frequency_info(wiphy, freq);
if (!attr || attr->disabled) {
l_error("AP frequency %u disabled or unsupported", freq);
ap: add support for 5GHz frequencies in AP mode This enables IWD to use 5GHz frequencies in AP mode. Currently 6GHz is not supported so we can assume a [General].Channel value 36 or above indicates the 5GHz band. It should be noted that the system will probably need a regulatory domain set in order for 5GHz to be allowed in AP mode. This is due to world roaming (00) restricting any/all 5GHz frequencies. This can be accomplished by setting main.conf [General].Country=CC to the country this AP will operate in.
2022-12-09 20:39:28 +01:00
return false;
}
ap: generate chandef for starting AP To include HT support a chandef needs to be created for whatever frequency is being used. This allows IWD to provide a secondary channel to the kernel in the case of 40MHz operation. Now the AP will generate a chandef when starting based on the channel set in the user profile (or default). If HT is not supported the chandef width is set to 20MHz no-HT, otherwise band_freq_to_ht_chandef is used.
2022-12-28 23:32:30 +01:00
if (ap->supports_ht) {
if (band_freq_to_ht_chandef(freq, attr, &ap->chandef) < 0) {
/*
* This is unlikely ever to fail since there are no
* 20Mhz restrictions, but just in case fall back to
* non-HT.
*/
ap->supports_ht = false;
l_warn("AP could not find HT chandef for frequency %u"
" using 20Mhz no-HT", freq);
goto no_ht;
}
} else {
no_ht:
ap->chandef.frequency = freq;
ap->chandef.channel_width = BAND_CHANDEF_WIDTH_20NOHT;
}
l_debug("AP using frequency %u and channel width %s",
ap->chandef.frequency,
band_chandef_width_to_string(
ap->chandef.channel_width));
ap: add support for 5GHz frequencies in AP mode This enables IWD to use 5GHz frequencies in AP mode. Currently 6GHz is not supported so we can assume a [General].Channel value 36 or above indicates the 5GHz band. It should be noted that the system will probably need a regulatory domain set in order for 5GHz to be allowed in AP mode. This is due to world roaming (00) restricting any/all 5GHz frequencies. This can be accomplished by setting main.conf [General].Country=CC to the country this AP will operate in.
2022-12-09 20:39:28 +01:00
return true;
}
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
static int ap_load_config(struct ap_state *ap, const struct l_settings *config,
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
bool *out_cck_rates)
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
{
ap: add profile settings PairwiseCiphers/GroupCipher These can now be optionally provided in an AP profile and provide a way to limit what ciphers can be chosen. This still is dependent on what the hardware supports.
2022-11-02 23:51:08 +01:00
struct wiphy *wiphy = netdev_get_wiphy(ap->netdev);
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
size_t len;
L_AUTO_FREE_VAR(char *, strval) = NULL;
ap: add profile settings PairwiseCiphers/GroupCipher These can now be optionally provided in an AP profile and provide a way to limit what ciphers can be chosen. This still is dependent on what the hardware supports.
2022-11-02 23:51:08 +01:00
_auto_(l_strv_free) char **ciphers_str = NULL;
uint16_t cipher_mask;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
int err;
ap: add profile settings PairwiseCiphers/GroupCipher These can now be optionally provided in an AP profile and provide a way to limit what ciphers can be chosen. This still is dependent on what the hardware supports.
2022-11-02 23:51:08 +01:00
int i;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
strval = l_settings_get_string(config, "General", "SSID");
if (L_WARN_ON(!strval))
return -ENOMSG;
len = strlen(strval);
if (len < 1 || len > 32) {
l_error("AP SSID length outside the [1, 32] range");
return -EINVAL;
}
if (L_WARN_ON(!l_utf8_validate(strval, len, NULL)))
return -EINVAL;
strcpy(ap->ssid, strval);
l_free(l_steal_ptr(strval));
if (!ap_load_psk(ap, config))
return -EINVAL;
/*
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
* This looks at the network configuration settings in @config and
* relevant global settings and if it determines that netconfig is to
* be enabled for the AP, it both creates the DHCP server object and
* processes IP settings, applying the defaults where needed.
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
*/
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
err = ap_load_ipv4(ap, config);
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
if (err)
return err;
if (l_settings_has_key(config, "General", "Channel")) {
unsigned int uintval;
if (!l_settings_get_uint(config, "General", "Channel",
ap: add support for 5GHz frequencies in AP mode This enables IWD to use 5GHz frequencies in AP mode. Currently 6GHz is not supported so we can assume a [General].Channel value 36 or above indicates the 5GHz band. It should be noted that the system will probably need a regulatory domain set in order for 5GHz to be allowed in AP mode. This is due to world roaming (00) restricting any/all 5GHz frequencies. This can be accomplished by setting main.conf [General].Country=CC to the country this AP will operate in.
2022-12-09 20:39:28 +01:00
&uintval)) {
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
l_error("AP Channel value unsupported");
return -EINVAL;
}
ap->channel = uintval;
ap: add support for 5GHz frequencies in AP mode This enables IWD to use 5GHz frequencies in AP mode. Currently 6GHz is not supported so we can assume a [General].Channel value 36 or above indicates the 5GHz band. It should be noted that the system will probably need a regulatory domain set in order for 5GHz to be allowed in AP mode. This is due to world roaming (00) restricting any/all 5GHz frequencies. This can be accomplished by setting main.conf [General].Country=CC to the country this AP will operate in.
2022-12-09 20:39:28 +01:00
/*
* 6GHz is not supported so we can use only a channel number to
* distinguish between 2.4 and 5GHz.
*/
if (ap->channel >= 36)
ap->band = BAND_FREQ_5_GHZ;
else
ap->band = BAND_FREQ_2_4_GHZ;
} else {
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
/* TODO: Start a Get Survey to decide the channel */
ap->channel = 6;
ap: add support for 5GHz frequencies in AP mode This enables IWD to use 5GHz frequencies in AP mode. Currently 6GHz is not supported so we can assume a [General].Channel value 36 or above indicates the 5GHz band. It should be noted that the system will probably need a regulatory domain set in order for 5GHz to be allowed in AP mode. This is due to world roaming (00) restricting any/all 5GHz frequencies. This can be accomplished by setting main.conf [General].Country=CC to the country this AP will operate in.
2022-12-09 20:39:28 +01:00
ap->band = BAND_FREQ_2_4_GHZ;
}
ap: add DisableHT setting This allows the user to disable HT explicitly in the AP profile
2023-02-24 01:57:23 +01:00
if (l_settings_has_key(config, "General", "DisableHT")) {
bool boolval;
if (!l_settings_get_bool(config, "General", "DisableHT",
&boolval)) {
l_error("AP [General].DisableHT not a valid boolean");
return -EINVAL;
}
ap->supports_ht = !boolval;
} else
ap->supports_ht = wiphy_get_ht_capabilities(wiphy, ap->band,
ap: include WMM parameter IE The WMM parameter IE is expected by the linux kernel for any AP supporting HT/VHT etc. IWD won't actually use WMM and its not clear exactly why the kernel uses this restriction, but regardless it must be included to support HT.
2022-12-28 23:32:29 +01:00
NULL) != NULL;
ap: add support for 5GHz frequencies in AP mode This enables IWD to use 5GHz frequencies in AP mode. Currently 6GHz is not supported so we can assume a [General].Channel value 36 or above indicates the 5GHz band. It should be noted that the system will probably need a regulatory domain set in order for 5GHz to be allowed in AP mode. This is due to world roaming (00) restricting any/all 5GHz frequencies. This can be accomplished by setting main.conf [General].Country=CC to the country this AP will operate in.
2022-12-09 20:39:28 +01:00
if (!ap_validate_band_channel(ap)) {
l_error("AP Band and Channel combination invalid");
return -EINVAL;
}
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
strval = l_settings_get_string(config, "WSC", "DeviceName");
if (strval) {
len = strlen(strval);
if (len > 32) {
l_error("AP WSC name length outside the [1, 32] range");
return -EINVAL;
}
if (!l_utf8_validate(strval, len, NULL)) {
l_error("AP WSC name doesn't validate as UTF-8");
return -EINVAL;
}
strcpy(ap->wsc_name, strval);
l_free(l_steal_ptr(strval));
} else
memcpy(ap->wsc_name, ap->ssid, 33);
strval = l_settings_get_string(config, "WSC", "PrimaryDeviceType");
if (strval) {
bool ok = wsc_device_type_from_setting_str(strval,
&ap->wsc_primary_device_type);
if (!ok) {
l_error("AP [WSC].PrimaryDeviceType format unknown");
return -EINVAL;
}
ap: add profile settings PairwiseCiphers/GroupCipher These can now be optionally provided in an AP profile and provide a way to limit what ciphers can be chosen. This still is dependent on what the hardware supports.
2022-11-02 23:51:08 +01:00
l_free(l_steal_ptr(strval));
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
} else {
/* Make ourselves a WFA standard PC by default */
ap->wsc_primary_device_type.category = 1;
memcpy(ap->wsc_primary_device_type.oui, wsc_wfa_oui, 3);
ap->wsc_primary_device_type.oui_type = 0x04;
ap->wsc_primary_device_type.subcategory = 1;
}
if (l_settings_get_value(config, "WSC", "AuthorizedMACs")) {
char **strvval;
unsigned int i;
strvval = l_settings_get_string_list(config, "WSC",
"AuthorizedMACs", ',');
if (!strvval) {
l_error("AP Authorized MACs list format wrong");
return -EINVAL;
}
ap->authorized_macs_num = l_strv_length(strvval);
ap->authorized_macs = l_malloc(ap->authorized_macs_num * 6);
for (i = 0; strvval[i]; i++)
if (!util_string_to_address(strvval[i],
ap->authorized_macs + i * 6)) {
l_error("Bad authorized MAC format: %s",
strvval[i]);
l_strfreev(strvval);
return -EINVAL;
}
l_strfreev(strvval);
}
ap: support PTK rekeys This adds support for rekeys to AP mode. A single timer is used and reset to the next station needing a rekey. A default rekey timer of 600 seconds is used unless the profile sets a timeout.
2023-01-18 22:18:00 +01:00
if (l_settings_has_key(config, "General", "RekeyTimeout")) {
unsigned int uintval;
if (!l_settings_get_uint(config, "General",
"RekeyTimeout", &uintval)) {
l_error("AP [General].RekeyTimeout is not valid");
return -EINVAL;
}
ap->rekey_time = uintval * L_USEC_PER_SEC;
} else
ap->rekey_time = 0;
ap: add support for 5GHz frequencies in AP mode This enables IWD to use 5GHz frequencies in AP mode. Currently 6GHz is not supported so we can assume a [General].Channel value 36 or above indicates the 5GHz band. It should be noted that the system will probably need a regulatory domain set in order for 5GHz to be allowed in AP mode. This is due to world roaming (00) restricting any/all 5GHz frequencies. This can be accomplished by setting main.conf [General].Country=CC to the country this AP will operate in.
2022-12-09 20:39:28 +01:00
/*
* Since 5GHz won't ever support only CCK rates we can ignore this
* setting on that band.
*/
if (ap->band & BAND_FREQ_5_GHZ)
*out_cck_rates = false;
else if (l_settings_get_value(config, "General", "NoCCKRates")) {
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
bool boolval;
if (!l_settings_get_bool(config, "General", "NoCCKRates",
&boolval)) {
l_error("AP [General].NoCCKRates not a valid "
"boolean");
return -EINVAL;
}
*out_cck_rates = !boolval;
} else
*out_cck_rates = true;
ap: add profile settings PairwiseCiphers/GroupCipher These can now be optionally provided in an AP profile and provide a way to limit what ciphers can be chosen. This still is dependent on what the hardware supports.
2022-11-02 23:51:08 +01:00
cipher_mask = wiphy_get_supported_ciphers(wiphy, IE_GROUP_CIPHERS);
/* If the config sets a group cipher use that directly */
strval = l_settings_get_string(config, "Security", "GroupCipher");
if (strval) {
enum ie_rsn_cipher_suite cipher = ap_string_to_cipher(strval);
if (!cipher || !(cipher & cipher_mask)) {
l_error("Unsupported or unknown group cipher %s",
strval);
return -ENOTSUP;
}
ap->group_cipher = cipher;
l_free(l_steal_ptr(strval));
} else {
/* No config override, use CCMP (or TKIP if not supported) */
if (cipher_mask & IE_RSN_CIPHER_SUITE_CCMP)
ap->group_cipher = IE_RSN_CIPHER_SUITE_CCMP;
else
ap->group_cipher = IE_RSN_CIPHER_SUITE_TKIP;
}
cipher_mask = wiphy_get_supported_ciphers(wiphy, IE_PAIRWISE_CIPHERS);
ciphers_str = l_settings_get_string_list(config, "Security",
"PairwiseCiphers", ',');
for (i = 0; ciphers_str && ciphers_str[i]; i++) {
enum ie_rsn_cipher_suite cipher =
ap_string_to_cipher(ciphers_str[i]);
/*
* Constrain list to only values in both supported ciphers and
* the cipher list provided.
*/
if (!cipher || !(cipher & cipher_mask)) {
l_error("Unsupported or unknown pairwise cipher %s",
ciphers_str[i]);
return -ENOTSUP;
}
ap->ciphers |= cipher;
}
if (!ap->ciphers) {
/*
* Default behavior if no ciphers are specified, disable TKIP
* for security if CCMP is available
*/
if (cipher_mask & IE_RSN_CIPHER_SUITE_CCMP)
cipher_mask &= ~IE_RSN_CIPHER_SUITE_TKIP;
ap->ciphers = cipher_mask;
}
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
return 0;
}
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
/*
* Start a simple independent WPA2 AP on given netdev.
*
ap: Pass "ops" struct to ap_start() Pass the event callback function pointer in a "struct ap_ops" instead of as individual ap_start() argument to make adding new callbacks easier.
2020-09-16 11:17:49 +02:00
* @ops.handle_event is required and must react to AP_EVENT_START_FAILED
* and AP_EVENT_STOPPING by forgetting the ap_state struct, which is
* going to be freed automatically.
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
* In the @config struct the [General].SSID key is required and one of
* [Security].Passphrase and [Security].PreSharedKey must be filled in.
* All other fields are optional.
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
*/
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
struct ap_state *ap_start(struct netdev *netdev, struct l_settings *config,
ap: add error out param to ap_start This allows the caller to extract a bit more information about what exactly went wrong.
2020-10-26 19:05:41 +01:00
const struct ap_ops *ops, int *err_out,
void *user_data)
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
{
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
struct ap_state *ap;
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
struct wiphy *wiphy = netdev_get_wiphy(netdev);
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
uint64_t wdev_id = netdev_get_wdev_id(netdev);
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
int err;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
bool cck_rates = true;
ap: select rates from hardware capabilities Rather than a list of static rates, choose what the hardware advertises. For the case of CCK rates only include rates 11mbps or lower.
2022-12-09 20:39:29 +01:00
const uint8_t *rates;
unsigned int num_rates;
unsigned int i;
ap: add error out param to ap_start This allows the caller to extract a bit more information about what exactly went wrong.
2020-10-26 19:05:41 +01:00
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (L_WARN_ON(!config)) {
if (err_out)
*err_out = -EINVAL;
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
ap: Support working without passphrase Add a "psk" setting to allow the user to pass the binary PSK directly instead of generating it from the passphrase and the SSID. In that case we'll only send the PSK to WSC enrollees.
2020-09-16 11:17:50 +02:00
return NULL;
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
}
ap: Support working without passphrase Add a "psk" setting to allow the user to pass the binary PSK directly instead of generating it from the passphrase and the SSID. In that case we'll only send the PSK to WSC enrollees.
2020-09-16 11:17:50 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap = l_new(struct ap_state, 1);
ap->nl80211 = l_genl_family_new(iwd_get_genl(), NL80211_GENL_NAME);
ap->netdev = netdev;
ap: Pass "ops" struct to ap_start() Pass the event callback function pointer in a "struct ap_ops" instead of as individual ap_start() argument to make adding new callbacks easier.
2020-09-16 11:17:49 +02:00
ap->ops = ops;
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap->user_data = user_data;
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
err = ap_load_config(ap, config, &cck_rates);
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
if (err)
ap: add StartProfile DBus method Users can now start an AP from settings based on a profile on disk. The only argument is the SSID which will be used to lookup the profile. If no profile is found a NotFound error will be returned. Any invalid profiles will result in an Invalid return.
2020-11-02 19:28:49 +01:00
goto error;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
err = -EINVAL;
ap: Push Button mode API and beacon changes This adds the API for putting the AP in Push Button mode, which we'll need to P2P GO side but may be useful on its own too. A WSC IE is added to our beacons and probe responses indicating whether the PBC mode is active.
2020-08-28 14:46:46 +02:00
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
ap->beacon_interval = 100;
ap: add support for scanning while in AP mode Scanning while in AP mode is somewhat of an edge case, but it does have some usefulness specifically with onboarding new devices, i.e. a new device starts an AP, a station connects and provides the new device with network credentials, the new device switches to station mode and connects to the desired network. In addition this could be used later for ACS (though this is a bit overkill for IWD's access point needs). Since AP performance is basically non-existant while scanning this feature is meant to be used in a limited scope. Two DBus API's were added which mirror the station interface: Scan and GetOrderedNetworks. Scan is no different than the station variant, and will perform an active scan on all channels. GetOrderedNetworks diverges from station and simply returns an array of dictionaries containing basic information about networks: { Name: <ssid> SignalStrength: <mBm> Security: <psk, open, or 8021x> } Limitations: - Hidden networks are not supported. This isn't really possible since the SSID's are unknown from the AP perspective. - Sharing scan results with station is not supported. This would be a convenient improvement in the future with respect to onboarding new devices. The scan could be performed in AP mode, then switch to station and connect immediately without needing to rescan. A quick hack could better this situation by not flushing scan results in station (if the kernel retains these on an iftype change).
2022-02-28 19:49:37 +01:00
ap->networks = l_queue_new();
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
wsc_uuid_from_addr(netdev_get_address(netdev), ap->wsc_uuid_r);
ap: select rates from hardware capabilities Rather than a list of static rates, choose what the hardware advertises. For the case of CCK rates only include rates 11mbps or lower.
2022-12-09 20:39:29 +01:00
rates = wiphy_get_supported_rates(wiphy, ap->band, &num_rates);
if (!rates)
goto error;
ap: Fix NULL ap->rates Make sure ap->rates is non-NULL both with and without no_cck_rates.
2020-09-11 21:50:56 +02:00
ap->rates = l_uintset_new(200);
ap: Add a no_cck_rates flag
2020-08-04 16:53:02 +02:00
ap: select rates from hardware capabilities Rather than a list of static rates, choose what the hardware advertises. For the case of CCK rates only include rates 11mbps or lower.
2022-12-09 20:39:29 +01:00
for (i = 0; i < num_rates; i++) {
if (cck_rates && !L_IN_SET(rates[i], 2, 4, 11, 22))
continue;
l_uintset_put(ap->rates, rates[i]);
ap: Add a no_cck_rates flag
2020-08-04 16:53:02 +02:00
}
ap: Handle PTK handshake frame 2/4, send frame 3/4
2017-09-22 05:06:42 +02:00
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
if (!frame_watch_add(wdev_id, 0, 0x0000 |
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
(MPDU_MANAGEMENT_SUBTYPE_ASSOCIATION_REQUEST << 4),
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
NULL, 0, ap_assoc_req_cb, ap, NULL))
goto error;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
if (!frame_watch_add(wdev_id, 0, 0x0000 |
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
(MPDU_MANAGEMENT_SUBTYPE_REASSOCIATION_REQUEST << 4),
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
NULL, 0, ap_reassoc_req_cb, ap, NULL))
goto error;
ap: Implement Reassociation Add some form of the Reassociation procedure with comment on the interpretation choice.
2017-09-30 04:28:12 +02:00
ap: include PROBE_RESP attribute if required by driver Certain drivers support/require probe response offloading which IWD did not check for or properly handle. If probe response offloading is required the probe response frame watch will not be added and instead the ATTR_PROBE_RESP will be included with START_AP. The head/tail builders were reused but slightly modified to check if the probe request frame is NULL, since it will be for use with START_AP.
2022-09-28 21:48:52 +02:00
if (!wiphy_supports_probe_resp_offload(wiphy)) {
if (!frame_watch_add(wdev_id, 0, 0x0000 |
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
(MPDU_MANAGEMENT_SUBTYPE_PROBE_REQUEST << 4),
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
NULL, 0, ap_probe_req_cb, ap, NULL))
ap: include PROBE_RESP attribute if required by driver Certain drivers support/require probe response offloading which IWD did not check for or properly handle. If probe response offloading is required the probe response frame watch will not be added and instead the ATTR_PROBE_RESP will be included with START_AP. The head/tail builders were reused but slightly modified to check if the probe request frame is NULL, since it will be for use with START_AP.
2022-09-28 21:48:52 +02:00
goto error;
}
ap: Handle Probe Request frames Parse probe requests and send probe responses to show up in active scans.
2017-09-22 05:06:37 +02:00
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
if (!frame_watch_add(wdev_id, 0, 0x0000 |
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
(MPDU_MANAGEMENT_SUBTYPE_DISASSOCIATION << 4),
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
NULL, 0, ap_disassoc_cb, ap, NULL))
goto error;
ap: Handle Association / Disassociation Parse Association Request frames and send Association Responses, handle Disassociation. With this we should be able to receive uncontrolled port data frames since we register the STAs with the kernel. In this version I don't register for Reassociation frames.
2017-09-22 05:06:39 +02:00
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
if (!frame_watch_add(wdev_id, 0, 0x0000 |
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
(MPDU_MANAGEMENT_SUBTYPE_AUTHENTICATION << 4),
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
NULL, 0, ap_auth_cb, ap, NULL))
goto error;
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
if (!frame_watch_add(wdev_id, 0, 0x0000 |
ap: Handle Authentication and Deauthentication frames Parse authentication frames and send responses, manage a list of STAs.
2017-09-22 05:06:38 +02:00
(MPDU_MANAGEMENT_SUBTYPE_DEAUTHENTICATION << 4),
ap: Switch to new frame watch API
2020-01-06 13:39:39 +01:00
NULL, 0, ap_deauth_cb, ap, NULL))
goto error;
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
ap: React to NL80211_CMD_STOP_AP events These events will tell use when our AP gets stopped without our request, for example due to suspend/resume.
2020-02-15 01:17:46 +01:00
ap->mlme_watch = l_genl_family_register(ap->nl80211, "mlme",
ap_mlme_notify, ap, NULL);
if (!ap->mlme_watch)
l_error("Registering for MLME notification failed");
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
if (ap->netconfig_set_addr4) {
ap: Save AP address as l_rtnl_address Change the char *addr_str and uint8_t prefix_len pair to an l_rtnl_address object and use ell/rtnl.h utilities that use that directly. Extend broadcast_from_ip to handle prefix_len.
2021-05-28 02:47:35 +02:00
ap->rtnl_add_cmd = l_rtnl_ifaddr_add(rtnl,
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
netdev_get_ifindex(netdev),
ap: Save AP address as l_rtnl_address Change the char *addr_str and uint8_t prefix_len pair to an l_rtnl_address object and use ell/rtnl.h utilities that use that directly. Extend broadcast_from_ip to handle prefix_len.
2021-05-28 02:47:35 +02:00
ap->netconfig_addr4,
ap: Refactor DHCP settings loading Extend the [IPv4].Address setting's syntax to allow a new format: a list of <IP>/<prefix_len> -form strings that define the address space from which a subnet is selected. Rewrite the DHCP settings loading with other notable changes: * validate some of the settings more thoroughly, * name all netconfig-related ap_state members with the netconfig_ prefix, * make sure we always call l_dhcp_server_set_netmask(), * allow netmasks other than 24-bit and change the default to 28 bits, * as requested avoid using the l_net_ ioctl-based functions although l_dhcp still uses them internally, * as requested avoid touching the ap_state members until the end of some functions so that on error they're basically a no-op (for readability).
2021-05-28 02:47:32 +02:00
ap_ifaddr4_added_cb, ap, NULL);
if (!ap->rtnl_add_cmd) {
l_error("Failed to add the IPv4 address");
goto error;
}
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
return ap;
}
ap: Move sending CMD_START_AP to common function
2021-05-10 10:57:56 +02:00
if (ap_start_send(ap)) {
if (err_out)
*err_out = 0;
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
ap: Move sending CMD_START_AP to common function
2021-05-10 10:57:56 +02:00
return ap;
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
}
error:
ap: add error out param to ap_start This allows the caller to extract a bit more information about what exactly went wrong.
2020-10-26 19:05:41 +01:00
if (err_out)
*err_out = err;
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap_reset(ap);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
l_genl_family_free(ap->nl80211);
l_free(ap);
return NULL;
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
}
static void ap_stop_cb(struct l_genl_msg *msg, void *user_data)
{
struct ap_state *ap = user_data;
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
int error = l_genl_msg_get_error(msg);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
ap->start_stop_cmd_id = 0;
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
if (error < 0)
ap: Make sure strerror argument is positive
2021-02-08 22:05:54 +01:00
l_error("STOP_AP failed: %s (%i)", strerror(-error), -error);
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
if (ap->stopped_func)
ap->stopped_func(ap->user_data);
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
l_genl_family_free(ap->nl80211);
l_free(ap);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
}
static struct l_genl_msg *ap_build_cmd_stop_ap(struct ap_state *ap)
{
struct l_genl_msg *cmd;
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
uint32_t ifindex = netdev_get_ifindex(ap->netdev);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
cmd = l_genl_msg_new_sized(NL80211_CMD_STOP_AP, 16);
l_genl_msg_append_attr(cmd, NL80211_ATTR_IFINDEX, 4, &ifindex);
return cmd;
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
}
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
/*
* Schedule the running @ap to be stopped and freed. The original
ap: Pass "ops" struct to ap_start() Pass the event callback function pointer in a "struct ap_ops" instead of as individual ap_start() argument to make adding new callbacks easier.
2020-09-16 11:17:49 +02:00
* ops and user_data are forgotten and a new callback can be
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
* provided if the caller needs to know when the interface becomes
* free, for example for a new ap_start call.
*
* The user must forget @ap when @stopped_func is called. If the
* @user_data ends up being destroyed before that, ap_free(ap) should
* be used to prevent @stopped_func from being called.
* If @stopped_func is not provided, the caller must forget @ap
* immediately.
*/
void ap_shutdown(struct ap_state *ap, ap_stopped_func_t stopped_func,
void *user_data)
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
{
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
struct l_genl_msg *cmd;
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
if (ap->started) {
ap->started = false;
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
if (ap_event(ap, AP_EVENT_STOPPING, NULL))
return;
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
}
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap_reset(ap);
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
if (ap->gtk_set) {
ap->gtk_set = false;
cmd = ap_build_cmd_del_key(ap);
if (!cmd) {
l_error("ap_build_cmd_del_key failed");
goto free_ap;
}
if (!l_genl_family_send(ap->nl80211, cmd, ap_gtk_op_cb, NULL,
NULL)) {
l_genl_msg_unref(cmd);
l_error("Issuing DEL_KEY failed");
goto free_ap;
}
}
cmd = ap_build_cmd_stop_ap(ap);
if (!cmd) {
l_error("ap_build_cmd_stop_ap failed");
goto free_ap;
}
ap: React to NL80211_CMD_STOP_AP events These events will tell use when our AP gets stopped without our request, for example due to suspend/resume.
2020-02-15 01:17:46 +01:00
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
ap->start_stop_cmd_id = l_genl_family_send(ap->nl80211, cmd, ap_stop_cb,
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
ap, NULL);
if (!ap->start_stop_cmd_id) {
l_genl_msg_unref(cmd);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
l_error("Sending STOP_AP failed");
goto free_ap;
ap: Build and send NL80211_CMD_START_AP and STOP_AP This should be enough for the AP to start sending beacons and appear in a passive scan.
2017-09-12 02:45:44 +02:00
}
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap->stopped_func = stopped_func;
ap->user_data = user_data;
return;
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
free_ap:
if (stopped_func)
stopped_func(user_data);
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
l_genl_family_free(ap->nl80211);
l_free(ap);
}
ap: Set up the GTK and pass to handshake Set a default GTK cipher type same as our current PTK type, generate a random GTK when the first STA connects and set it up in the kernel, then pass the values that EAPoL is going to need to the handshake_state.
2018-09-25 17:43:48 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
/* Free @ap without a graceful shutdown */
void ap_free(struct ap_state *ap)
{
ap: Don't defer ap_reset when ap_free called in event Previously we added logic to defer doing anything in ap_free() to after the AP event handler has returned so that ap_event() has a chance to inform whoever called it that the ap_state has been freed. But there's also a chance that the event handler is destroying both the AP and the netdev it runs on, so after the handler has returned we can't even use netdev_get_wdev_id or netdev_get_ifindex. The easiest solution seems to be to call ap_reset() in ap_free() even if we're within an event handler to ensure we no longer need any external objects. Also make sure ap_reset() can be called multiple times. Another option would be to watch for NETDEV_WATCH_EVENT_DEL and remove our reference to the netdev (because there's no need actually call l_rtnl_ifaddr_delete or frame_watch_wdev_remove if the netdev was destroyed -- frame_watch already tracks netdev removals), or to save just the ifindex and the wdev id...
2022-01-21 11:24:37 +01:00
ap_reset(ap);
ap: Delay ap_free if called inside event handler ap.c has been mostly careful to call the event handler at the end of any externally called function to allow methods like ap_free() to be called within the handler, but that isn't enough. For example in ap_del_station we may end up emitting two events: STATION_REMOVED and DHCP_LEASE_EXPIRED. Use a slightly more complicated mechanism to explicitly guard ap_free calls inside the event handler. To make it easier, simplify cleanup in ap_assoc_reassoc with the use of _auto_. In ap_del_station reorder the actions to send the STATION_REMOVED event first as the DHCP_LEASE_EXPIRED is a consequence of the former and it makes sense for the handler to react to it first.
2021-11-08 12:28:36 +01:00
if (ap->in_event) {
ap->free_pending = true;
return;
}
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
l_genl_family_free(ap->nl80211);
l_free(ap);
}
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
bool ap_station_disconnect(struct ap_state *ap, const uint8_t *mac,
enum mmpdu_reason_code reason)
{
struct sta_state *sta;
if (!ap->started)
return false;
sta = l_queue_remove_if(ap->sta_states, ap_sta_match_addr, mac);
if (!sta)
return false;
ap: build HT Capabilities/Operation elements If supported this will include the HT capabilities and HT operations elements in beacons/probes. Some shortcuts were taken here since not all the information is currently parsed from the hardware. Namely the HT operation element does not include the basic MCS set. Still, this will at least show stations that the AP is capable of more than just basic rates. The builders themselves are structured similar to the basic rates builder where they build only the contents and return the length. The caller must set the type/length manually. This is to support the two use cases of using with an IE builder vs direct pointer.
2022-12-28 23:32:31 +01:00
if (ap->supports_ht)
ap_update_beacon(ap);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap_del_station(sta, reason, false);
ap_sta_free(sta);
return true;
}
ap: Push Button mode API and beacon changes This adds the API for putting the AP in Push Button mode, which we'll need to P2P GO side but may be useful on its own too. A WSC IE is added to our beacons and probe responses indicating whether the PBC mode is active.
2020-08-28 14:46:46 +02:00
static void ap_wsc_pbc_timeout_cb(struct l_timeout *timeout, void *user_data)
{
struct ap_state *ap = user_data;
l_debug("PBC mode timeout");
ap_wsc_exit_pbc(ap);
}
static void ap_wsc_pbc_timeout_destroy(void *user_data)
{
struct ap_state *ap = user_data;
ap->wsc_pbc_timeout = NULL;
}
bool ap_push_button(struct ap_state *ap)
{
if (!ap->started)
return false;
ap: WSC Probe Request processing logic Implement the caching of WSC probe requests -- when an Enrollee later associates to start registration we need to have its Probe Request on file. Also use this cache for PBC "Session Overlap" detection.
2020-08-28 14:46:47 +02:00
if (l_queue_length(ap->wsc_pbc_probes) > 1) {
l_debug("Can't start PBC mode due to Session Overlap");
return false;
}
ap: Push Button mode API and beacon changes This adds the API for putting the AP in Push Button mode, which we'll need to P2P GO side but may be useful on its own too. A WSC IE is added to our beacons and probe responses indicating whether the PBC mode is active.
2020-08-28 14:46:46 +02:00
/*
* WSC v2.0.5 Section 11.3: "Multiple presses of the button are
* permitted. If a PBC button on an Enrollee or Registrar is
* pressed again during Walk Time, the timers for that device are
* restarted at that time [...]"
*/
if (ap->wsc_pbc_timeout) {
l_timeout_modify(ap->wsc_pbc_timeout, AP_WSC_PBC_WALK_TIME);
return true;
}
ap->wsc_pbc_timeout = l_timeout_create(AP_WSC_PBC_WALK_TIME,
ap_wsc_pbc_timeout_cb, ap,
ap_wsc_pbc_timeout_destroy);
ap->wsc_dpid = WSC_DEVICE_PASSWORD_ID_PUSH_BUTTON;
ap_update_beacon(ap);
return true;
}
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
struct ap_if_data {
struct netdev *netdev;
struct ap_state *ap;
struct l_dbus_message *pending;
};
ap: update Scanning property when AP starts This wasn't being updated meaning the property is missing until a scan is issued over DBus. Rather than duplicate all the property changed calls they were all factored out into a helper function.
2022-11-04 17:37:55 +01:00
static void ap_properties_changed(struct ap_if_data *ap_if)
{
l_dbus_property_changed(dbus_get_bus(),
netdev_get_path(ap_if->netdev),
IWD_AP_INTERFACE, "Started");
l_dbus_property_changed(dbus_get_bus(),
netdev_get_path(ap_if->netdev),
IWD_AP_INTERFACE, "Name");
l_dbus_property_changed(dbus_get_bus(),
netdev_get_path(ap_if->netdev),
IWD_AP_INTERFACE, "Frequency");
l_dbus_property_changed(dbus_get_bus(),
netdev_get_path(ap_if->netdev),
IWD_AP_INTERFACE, "PairwiseCiphers");
l_dbus_property_changed(dbus_get_bus(),
netdev_get_path(ap_if->netdev),
IWD_AP_INTERFACE, "GroupCipher");
l_dbus_property_changed(dbus_get_bus(),
netdev_get_path(ap_if->netdev),
IWD_AP_INTERFACE, "Scanning");
}
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
static void ap_if_event_func(enum ap_event_type type, const void *event_data,
void *user_data)
{
struct ap_if_data *ap_if = user_data;
struct l_dbus_message *reply;
switch (type) {
case AP_EVENT_START_FAILED:
ap: Send a specific error message on async AP start failure We generate the DBus error reply type from the errno only when ap_start() was failing synchronously, now also send the errno through the callbacks so that we can also return a specific DBus reply when failing asynchronously. Thea AP autotest relies on receiving the AlreadyExists DBus error.
2021-05-28 02:47:34 +02:00
{
const struct ap_event_start_failed_data *data = event_data;
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
if (L_WARN_ON(!ap_if->pending))
break;
ap: Send a specific error message on async AP start failure We generate the DBus error reply type from the errno only when ap_start() was failing synchronously, now also send the errno through the callbacks so that we can also return a specific DBus reply when failing asynchronously. Thea AP autotest relies on receiving the AlreadyExists DBus error.
2021-05-28 02:47:34 +02:00
reply = dbus_error_from_errno(data->error, ap_if->pending);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
dbus_pending_reply(&ap_if->pending, reply);
ap_if->ap = NULL;
break;
ap: Send a specific error message on async AP start failure We generate the DBus error reply type from the errno only when ap_start() was failing synchronously, now also send the errno through the callbacks so that we can also return a specific DBus reply when failing asynchronously. Thea AP autotest relies on receiving the AlreadyExists DBus error.
2021-05-28 02:47:34 +02:00
}
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
case AP_EVENT_STARTED:
if (L_WARN_ON(!ap_if->pending))
break;
ap: tie diagnostic interface to AP Start/Stop The diagnostic interface serves no purpose until the AP has been started. Any calls on it will return an error so instead it makes more sense to bring it up when the AP is started, and down when the AP is stopped.
2021-02-02 00:52:17 +01:00
l_dbus_object_add_interface(dbus_get_bus(),
netdev_get_path(ap_if->netdev),
IWD_AP_DIAGNOSTIC_INTERFACE,
ap_if);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
reply = l_dbus_message_new_method_return(ap_if->pending);
dbus_pending_reply(&ap_if->pending, reply);
ap: update Scanning property when AP starts This wasn't being updated meaning the property is missing until a scan is issued over DBus. Rather than duplicate all the property changed calls they were all factored out into a helper function.
2022-11-04 17:37:55 +01:00
ap_properties_changed(ap_if);
ap: set link/operstate on AP start/stop Prior to this netdev_connect_ok set setting this which really only applies to station mode. In addition this happens for each new station that connects to the AP. Instead set the operstate / link mode when AP starts and stops.
2021-04-28 01:34:52 +02:00
l_rtnl_set_linkmode_and_operstate(rtnl,
netdev_get_ifindex(ap_if->netdev),
IF_LINK_MODE_DEFAULT, IF_OPER_UP,
NULL, NULL, NULL);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
break;
case AP_EVENT_STOPPING:
ap: tie diagnostic interface to AP Start/Stop The diagnostic interface serves no purpose until the AP has been started. Any calls on it will return an error so instead it makes more sense to bring it up when the AP is started, and down when the AP is stopped.
2021-02-02 00:52:17 +01:00
l_dbus_object_remove_interface(dbus_get_bus(),
netdev_get_path(ap_if->netdev),
IWD_AP_DIAGNOSTIC_INTERFACE);
ap: update Scanning property when AP starts This wasn't being updated meaning the property is missing until a scan is issued over DBus. Rather than duplicate all the property changed calls they were all factored out into a helper function.
2022-11-04 17:37:55 +01:00
ap_properties_changed(ap_if);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap: set link/operstate on AP start/stop Prior to this netdev_connect_ok set setting this which really only applies to station mode. In addition this happens for each new station that connects to the AP. Instead set the operstate / link mode when AP starts and stops.
2021-04-28 01:34:52 +02:00
l_rtnl_set_linkmode_and_operstate(rtnl,
netdev_get_ifindex(ap_if->netdev),
IF_LINK_MODE_DORMANT, IF_OPER_DOWN,
NULL, NULL, NULL);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
if (!ap_if->pending)
ap_if->ap = NULL;
break;
case AP_EVENT_STATION_ADDED:
case AP_EVENT_STATION_REMOVED:
ap: Push Button mode API and beacon changes This adds the API for putting the AP in Push Button mode, which we'll need to P2P GO side but may be useful on its own too. A WSC IE is added to our beacons and probe responses indicating whether the PBC mode is active.
2020-08-28 14:46:46 +02:00
case AP_EVENT_REGISTRATION_START:
case AP_EVENT_REGISTRATION_SUCCESS:
case AP_EVENT_PBC_MODE_EXIT:
ap: Forward DHCP events to AP event handler
2021-06-04 03:50:45 +02:00
case AP_EVENT_DHCP_NEW_LEASE:
case AP_EVENT_DHCP_LEASE_EXPIRED:
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
/* Ignored */
break;
}
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
}
ap: Pass "ops" struct to ap_start() Pass the event callback function pointer in a "struct ap_ops" instead of as individual ap_start() argument to make adding new callbacks easier.
2020-09-16 11:17:49 +02:00
static const struct ap_ops ap_dbus_ops = {
.handle_event = ap_if_event_func,
};
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
static struct l_dbus_message *ap_dbus_start(struct l_dbus *dbus,
struct l_dbus_message *message, void *user_data)
{
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
struct ap_if_data *ap_if = user_data;
ap: Rename wpa2_psk to wpa2_passphrase on DBus Use the passphrase naming instead of PSK.
2020-09-16 11:17:51 +02:00
const char *ssid, *wpa2_passphrase;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
struct l_settings *config;
ap: add error out param to ap_start This allows the caller to extract a bit more information about what exactly went wrong.
2020-10-26 19:05:41 +01:00
int err;
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
if (ap_if->ap && ap_if->ap->started)
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
return dbus_error_already_exists(message);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
if (ap_if->ap || ap_if->pending)
return dbus_error_busy(message);
ap: Rename wpa2_psk to wpa2_passphrase on DBus Use the passphrase naming instead of PSK.
2020-09-16 11:17:51 +02:00
if (!l_dbus_message_get_arguments(message, "ss",
&ssid, &wpa2_passphrase))
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
return dbus_error_invalid_args(message);
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
config = l_settings_new();
l_settings_set_string(config, "General", "SSID", ssid);
l_settings_set_string(config, "Security", "Passphrase",
wpa2_passphrase);
l_settings_add_group(config, "IPv4");
ap: Move AP parameters to a struct To limit the number of ap_start parameters, group basic AP config parameters in the ap_config struct that is passed as a pointer and owned by the ap_state.
2020-08-27 20:14:46 +02:00
ap: add error out param to ap_start This allows the caller to extract a bit more information about what exactly went wrong.
2020-10-26 19:05:41 +01:00
ap_if->ap = ap_start(ap_if->netdev, config, &ap_dbus_ops, &err, ap_if);
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
l_settings_free(config);
if (!ap_if->ap)
ap: add error out param to ap_start This allows the caller to extract a bit more information about what exactly went wrong.
2020-10-26 19:05:41 +01:00
return dbus_error_from_errno(err, message);
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap_if->pending = l_dbus_message_ref(message);
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
return NULL;
}
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
static void ap_dbus_stop_cb(void *user_data)
{
struct ap_if_data *ap_if = user_data;
struct l_dbus_message *reply;
if (L_WARN_ON(!ap_if->pending))
return;
reply = l_dbus_message_new_method_return(ap_if->pending);
dbus_pending_reply(&ap_if->pending, reply);
ap_if->ap = NULL;
}
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
static struct l_dbus_message *ap_dbus_stop(struct l_dbus *dbus,
struct l_dbus_message *message, void *user_data)
{
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
struct ap_if_data *ap_if = user_data;
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
if (!ap_if->ap) {
if (ap_if->pending)
return dbus_error_busy(message);
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
/* already stopped, no-op */
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
return l_dbus_message_new_method_return(message);
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
}
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
if (ap_if->pending) {
struct l_dbus_message *reply;
reply = dbus_error_aborted(ap_if->pending);
dbus_pending_reply(&ap_if->pending, reply);
}
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap_if->pending = l_dbus_message_ref(message);
ap_shutdown(ap_if->ap, ap_dbus_stop_cb, ap_if);
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
return NULL;
}
ap: add StartProfile DBus method Users can now start an AP from settings based on a profile on disk. The only argument is the SSID which will be used to lookup the profile. If no profile is found a NotFound error will be returned. Any invalid profiles will result in an Invalid return.
2020-11-02 19:28:49 +01:00
static struct l_dbus_message *ap_dbus_start_profile(struct l_dbus *dbus,
struct l_dbus_message *message,
void *user_data)
{
struct ap_if_data *ap_if = user_data;
const char *ssid;
ap: Fix l_settings leak l_settings is leaked if l_settings_load_from_file() fails.
2021-04-30 18:36:03 +02:00
_auto_(l_settings_free) struct l_settings *config = NULL;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
char *config_path;
ap: add StartProfile DBus method Users can now start an AP from settings based on a profile on disk. The only argument is the SSID which will be used to lookup the profile. If no profile is found a NotFound error will be returned. Any invalid profiles will result in an Invalid return.
2020-11-02 19:28:49 +01:00
int err;
if (ap_if->ap && ap_if->ap->started)
return dbus_error_already_exists(message);
if (ap_if->ap || ap_if->pending)
return dbus_error_busy(message);
if (!l_dbus_message_get_arguments(message, "s", &ssid))
return dbus_error_invalid_args(message);
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
config = l_settings_new();
config_path = storage_get_path("ap/%s.ap", ssid);
err = l_settings_load_from_file(config, config_path) ? 0 : -EIO;
l_free(config_path);
if (err)
goto error;
/*
* Since [General].SSID is not an allowed setting for a profile on
* disk, we're free to potentially overwrite it with the SSID that
* the DBus user asked for.
*/
l_settings_set_string(config, "General", "SSID", ssid);
ap: add StartProfile DBus method Users can now start an AP from settings based on a profile on disk. The only argument is the SSID which will be used to lookup the profile. If no profile is found a NotFound error will be returned. Any invalid profiles will result in an Invalid return.
2020-11-02 19:28:49 +01:00
ap_if->ap = ap_start(ap_if->netdev, config, &ap_dbus_ops, &err, ap_if);
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
if (!ap_if->ap)
goto error;
ap: add StartProfile DBus method Users can now start an AP from settings based on a profile on disk. The only argument is the SSID which will be used to lookup the profile. If no profile is found a NotFound error will be returned. Any invalid profiles will result in an Invalid return.
2020-11-02 19:28:49 +01:00
ap_if->pending = l_dbus_message_ref(message);
return NULL;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
error:
return dbus_error_from_errno(err, message);
ap: add StartProfile DBus method Users can now start an AP from settings based on a profile on disk. The only argument is the SSID which will be used to lookup the profile. If no profile is found a NotFound error will be returned. Any invalid profiles will result in an Invalid return.
2020-11-02 19:28:49 +01:00
}
ap: add support for scanning while in AP mode Scanning while in AP mode is somewhat of an edge case, but it does have some usefulness specifically with onboarding new devices, i.e. a new device starts an AP, a station connects and provides the new device with network credentials, the new device switches to station mode and connects to the desired network. In addition this could be used later for ACS (though this is a bit overkill for IWD's access point needs). Since AP performance is basically non-existant while scanning this feature is meant to be used in a limited scope. Two DBus API's were added which mirror the station interface: Scan and GetOrderedNetworks. Scan is no different than the station variant, and will perform an active scan on all channels. GetOrderedNetworks diverges from station and simply returns an array of dictionaries containing basic information about networks: { Name: <ssid> SignalStrength: <mBm> Security: <psk, open, or 8021x> } Limitations: - Hidden networks are not supported. This isn't really possible since the SSID's are unknown from the AP perspective. - Sharing scan results with station is not supported. This would be a convenient improvement in the future with respect to onboarding new devices. The scan could be performed in AP mode, then switch to station and connect immediately without needing to rescan. A quick hack could better this situation by not flushing scan results in station (if the kernel retains these on an iftype change).
2022-02-28 19:49:37 +01:00
static void ap_set_scanning(struct ap_state *ap, bool scanning)
{
if (ap->scanning == scanning)
return;
ap->scanning = scanning;
l_dbus_property_changed(dbus_get_bus(), netdev_get_path(ap->netdev),
IWD_AP_INTERFACE, "Scanning");
}
static void ap_scan_triggered(int err, void *user_data)
{
struct ap_state *ap = user_data;
struct l_dbus_message *reply;
if (err < 0) {
reply = dbus_error_from_errno(err, ap->scan_pending);
dbus_pending_reply(&ap->scan_pending, reply);
return;
}
l_debug("AP scan triggered for %s", netdev_get_name(ap->netdev));
reply = l_dbus_message_new_method_return(ap->scan_pending);
l_dbus_message_set_arguments(reply, "");
dbus_pending_reply(&ap->scan_pending, reply);
ap_set_scanning(ap, true);
}
static bool ap_scan_notify(int err, struct l_queue *bss_list,
const struct scan_freq_set *freqs,
void *user_data)
{
struct ap_state *ap = user_data;
const struct l_queue_entry *bss_entry;
ap_set_scanning(ap, false);
/* Remove all networks, then re-populate with fresh BSS list */
l_queue_clear(ap->networks, l_free);
for (bss_entry = l_queue_get_entries(bss_list); bss_entry;
bss_entry = bss_entry->next) {
struct scan_bss *bss = bss_entry->data;
ap_network_append(ap, bss);
}
l_debug("");
return false;
}
static void ap_scan_destroy(void *user_data)
{
struct ap_state *ap = user_data;
ap->scan_id = 0;
}
static struct l_dbus_message *ap_dbus_scan(struct l_dbus *dbus,
struct l_dbus_message *message,
void *user_data)
{
struct ap_if_data *ap_if = user_data;
uint64_t wdev_id = netdev_get_wdev_id(ap_if->netdev);
struct scan_parameters params = { 0};
if (wiphy_has_feature(wiphy_find_by_wdev(wdev_id),
NL80211_FEATURE_AP_SCAN))
params.ap_scan = true;
/*
* TODO: There is really nothing preventing scanning while stopped.
* The only consideration would be if a scan is ongoing and the
* AP is started. Queuing Start() as wiphy work may be required to
* handle this case if needed. For now just limit to started APs.
*/
if (!ap_if->ap || !ap_if->ap->started)
return dbus_error_not_available(message);
if (ap_if->ap->scan_id)
return dbus_error_busy(message);
ap_if->ap->scan_id = scan_active_full(wdev_id, &params,
ap_scan_triggered,
ap_scan_notify,
ap_if->ap, ap_scan_destroy);
if (!ap_if->ap->scan_id)
return dbus_error_failed(message);
ap_if->ap->scan_pending = l_dbus_message_ref(message);
return NULL;
}
static void dbus_append_network(struct l_dbus_message_builder *builder,
struct ap_network *network)
{
l_dbus_message_builder_enter_array(builder, "{sv}");
dbus_append_dict_basic(builder, "Name", 's', network->ssid);
dbus_append_dict_basic(builder, "SignalStrength", 'n',
&network->signal);
dbus_append_dict_basic(builder, "Type", 's',
security_to_str(network->security));
l_dbus_message_builder_leave_array(builder);
}
static struct l_dbus_message *ap_dbus_get_networks(struct l_dbus *dbus,
struct l_dbus_message *message,
void *user_data)
{
struct ap_if_data *ap_if = user_data;
struct l_dbus_message *reply;
struct l_dbus_message_builder *builder;
const struct l_queue_entry *entry;
if (!ap_if->ap || !ap_if->ap->started)
return dbus_error_not_available(message);
reply = l_dbus_message_new_method_return(message);
builder = l_dbus_message_builder_new(reply);
l_dbus_message_builder_enter_array(builder, "a{sv}");
for (entry = l_queue_get_entries(ap_if->ap->networks); entry;
entry = entry->next) {
struct ap_network *network = entry->data;
dbus_append_network(builder, network);
}
l_dbus_message_builder_leave_array(builder);
l_dbus_message_builder_finalize(builder);
l_dbus_message_builder_destroy(builder);
return reply;
}
ap: expose property 'Started' property on the Dbus
2018-07-30 00:32:55 +02:00
static bool ap_dbus_property_get_started(struct l_dbus *dbus,
struct l_dbus_message *message,
struct l_dbus_message_builder *builder,
void *user_data)
{
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
struct ap_if_data *ap_if = user_data;
bool started = ap_if->ap && ap_if->ap->started;
ap: expose property 'Started' property on the Dbus
2018-07-30 00:32:55 +02:00
l_dbus_message_builder_append_basic(builder, 'b', &started);
return true;
}
ap: add Name property Its useful being able to refer to the network Name/SSID once an AP is started. For example opening an iwctl session with an already started AP provides no way of obtaining the SSID.
2021-02-02 00:52:14 +01:00
static bool ap_dbus_property_get_name(struct l_dbus *dbus,
struct l_dbus_message *message,
struct l_dbus_message_builder *builder,
void *user_data)
{
struct ap_if_data *ap_if = user_data;
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
if (!ap_if->ap || !ap_if->ap->started)
ap: add Name property Its useful being able to refer to the network Name/SSID once an AP is started. For example opening an iwctl session with an already started AP provides no way of obtaining the SSID.
2021-02-02 00:52:14 +01:00
return false;
l_dbus_message_builder_append_basic(builder, 's',
ap: Drop struct ap_config in favor of l_settings Change ap_start to load all of the AP configuration from a struct l_settings, moving the 6 or so parameters from struct ap_config members to the l_settings groups and keys. This extends the ap profile concept used for the DHCP settings. ap_start callers create the l_settings object and fill the values in it or read the settings in from a file. Since ap_setup_dhcp and ap_load_profile_and_dhcp no longer do the settings file loading, they needed to be refactored and some issues were fixed in their logic, e.g. l_dhcp_server_set_ip_address() was never called when the "IP pool" was used. Also the IP pool was previously only used if the ap->config->profile was NULL and this didn't match what the docs said: "If [IPv4].Address is not provided and no IP address is set on the interface prior to calling StartProfile the IP pool will be used."
2021-04-27 01:34:06 +02:00
ap_if->ap->ssid);
ap: add Name property Its useful being able to refer to the network Name/SSID once an AP is started. For example opening an iwctl session with an already started AP provides no way of obtaining the SSID.
2021-02-02 00:52:14 +01:00
return true;
}
ap: add support for scanning while in AP mode Scanning while in AP mode is somewhat of an edge case, but it does have some usefulness specifically with onboarding new devices, i.e. a new device starts an AP, a station connects and provides the new device with network credentials, the new device switches to station mode and connects to the desired network. In addition this could be used later for ACS (though this is a bit overkill for IWD's access point needs). Since AP performance is basically non-existant while scanning this feature is meant to be used in a limited scope. Two DBus API's were added which mirror the station interface: Scan and GetOrderedNetworks. Scan is no different than the station variant, and will perform an active scan on all channels. GetOrderedNetworks diverges from station and simply returns an array of dictionaries containing basic information about networks: { Name: <ssid> SignalStrength: <mBm> Security: <psk, open, or 8021x> } Limitations: - Hidden networks are not supported. This isn't really possible since the SSID's are unknown from the AP perspective. - Sharing scan results with station is not supported. This would be a convenient improvement in the future with respect to onboarding new devices. The scan could be performed in AP mode, then switch to station and connect immediately without needing to rescan. A quick hack could better this situation by not flushing scan results in station (if the kernel retains these on an iftype change).
2022-02-28 19:49:37 +01:00
static bool ap_dbus_property_get_scanning(struct l_dbus *dbus,
struct l_dbus_message *message,
struct l_dbus_message_builder *builder,
void *user_data)
{
struct ap_if_data *ap_if = user_data;
bool bval;
if (!ap_if->ap || !ap_if->ap->started)
return false;
bval = ap_if->ap->scanning;
l_dbus_message_builder_append_basic(builder, 'b', &bval);
return true;
}
ap: add frequency to AP interface
2022-11-01 21:17:36 +01:00
static bool ap_dbus_property_get_freq(struct l_dbus *dbus,
struct l_dbus_message *message,
struct l_dbus_message_builder *builder,
void *user_data)
{
struct ap_if_data *ap_if = user_data;
uint32_t freq;
if (!ap_if->ap || !ap_if->ap->started)
return false;
ap: add support for 5GHz frequencies in AP mode This enables IWD to use 5GHz frequencies in AP mode. Currently 6GHz is not supported so we can assume a [General].Channel value 36 or above indicates the 5GHz band. It should be noted that the system will probably need a regulatory domain set in order for 5GHz to be allowed in AP mode. This is due to world roaming (00) restricting any/all 5GHz frequencies. This can be accomplished by setting main.conf [General].Country=CC to the country this AP will operate in.
2022-12-09 20:39:28 +01:00
freq = band_channel_to_freq(ap_if->ap->channel, ap_if->ap->band);
ap: add frequency to AP interface
2022-11-01 21:17:36 +01:00
l_dbus_message_builder_append_basic(builder, 'u', &freq);
return true;
}
ap: add PairwiseCiphers/GroupCipher to dbus interface Expose these values on the DBus interface so clients can view them.
2022-11-02 23:51:11 +01:00
static bool ap_dbus_property_get_pairwise(struct l_dbus *dbus,
struct l_dbus_message *message,
struct l_dbus_message_builder *builder,
void *user_data)
{
struct ap_if_data *ap_if = user_data;
ap: change PairwiseCiphers to string array A comma separated list as a string was ok for pure display purposes but if any processing needed to be done on these values by external consumers it really makes more sense to use a DBus array.
2022-11-04 17:37:58 +01:00
char **ciphers;
size_t i;
ap: add PairwiseCiphers/GroupCipher to dbus interface Expose these values on the DBus interface so clients can view them.
2022-11-02 23:51:11 +01:00
if (!ap_if->ap || !ap_if->ap->started)
return false;
ap: change PairwiseCiphers to string array A comma separated list as a string was ok for pure display purposes but if any processing needed to be done on these values by external consumers it really makes more sense to use a DBus array.
2022-11-04 17:37:58 +01:00
ciphers = ap_ciphers_to_strv(ap_if->ap->ciphers);
l_dbus_message_builder_enter_array(builder, "s");
for (i = 0; ciphers[i]; i++)
l_dbus_message_builder_append_basic(builder, 's', ciphers[i]);
l_dbus_message_builder_leave_array(builder);
ap: add PairwiseCiphers/GroupCipher to dbus interface Expose these values on the DBus interface so clients can view them.
2022-11-02 23:51:11 +01:00
ap: change PairwiseCiphers to string array A comma separated list as a string was ok for pure display purposes but if any processing needed to be done on these values by external consumers it really makes more sense to use a DBus array.
2022-11-04 17:37:58 +01:00
l_strv_free(ciphers);
ap: add PairwiseCiphers/GroupCipher to dbus interface Expose these values on the DBus interface so clients can view them.
2022-11-02 23:51:11 +01:00
return true;
}
static bool ap_dbus_property_get_group(struct l_dbus *dbus,
struct l_dbus_message *message,
struct l_dbus_message_builder *builder,
void *user_data)
{
struct ap_if_data *ap_if = user_data;
ap: change PairwiseCiphers to string array A comma separated list as a string was ok for pure display purposes but if any processing needed to be done on these values by external consumers it really makes more sense to use a DBus array.
2022-11-04 17:37:58 +01:00
char **cipher;
ap: add PairwiseCiphers/GroupCipher to dbus interface Expose these values on the DBus interface so clients can view them.
2022-11-02 23:51:11 +01:00
if (!ap_if->ap || !ap_if->ap->started)
return false;
ap: change PairwiseCiphers to string array A comma separated list as a string was ok for pure display purposes but if any processing needed to be done on these values by external consumers it really makes more sense to use a DBus array.
2022-11-04 17:37:58 +01:00
cipher = ap_ciphers_to_strv(ap_if->ap->group_cipher);
ap: add PairwiseCiphers/GroupCipher to dbus interface Expose these values on the DBus interface so clients can view them.
2022-11-02 23:51:11 +01:00
ap: change PairwiseCiphers to string array A comma separated list as a string was ok for pure display purposes but if any processing needed to be done on these values by external consumers it really makes more sense to use a DBus array.
2022-11-04 17:37:58 +01:00
/* Group cipher will only ever be a single value */
l_dbus_message_builder_append_basic(builder, 's', cipher[0]);
l_strv_free(cipher);
ap: add PairwiseCiphers/GroupCipher to dbus interface Expose these values on the DBus interface so clients can view them.
2022-11-02 23:51:11 +01:00
return true;
}
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
static void ap_setup_interface(struct l_dbus_interface *interface)
{
l_dbus_interface_method(interface, "Start", 0, ap_dbus_start, "",
ap: Rename wpa2_psk to wpa2_passphrase on DBus Use the passphrase naming instead of PSK.
2020-09-16 11:17:51 +02:00
"ss", "ssid", "wpa2_passphrase");
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
l_dbus_interface_method(interface, "Stop", 0, ap_dbus_stop, "", "");
ap: add StartProfile DBus method Users can now start an AP from settings based on a profile on disk. The only argument is the SSID which will be used to lookup the profile. If no profile is found a NotFound error will be returned. Any invalid profiles will result in an Invalid return.
2020-11-02 19:28:49 +01:00
l_dbus_interface_method(interface, "StartProfile", 0,
ap_dbus_start_profile, "", "s",
"ssid");
ap: add support for scanning while in AP mode Scanning while in AP mode is somewhat of an edge case, but it does have some usefulness specifically with onboarding new devices, i.e. a new device starts an AP, a station connects and provides the new device with network credentials, the new device switches to station mode and connects to the desired network. In addition this could be used later for ACS (though this is a bit overkill for IWD's access point needs). Since AP performance is basically non-existant while scanning this feature is meant to be used in a limited scope. Two DBus API's were added which mirror the station interface: Scan and GetOrderedNetworks. Scan is no different than the station variant, and will perform an active scan on all channels. GetOrderedNetworks diverges from station and simply returns an array of dictionaries containing basic information about networks: { Name: <ssid> SignalStrength: <mBm> Security: <psk, open, or 8021x> } Limitations: - Hidden networks are not supported. This isn't really possible since the SSID's are unknown from the AP perspective. - Sharing scan results with station is not supported. This would be a convenient improvement in the future with respect to onboarding new devices. The scan could be performed in AP mode, then switch to station and connect immediately without needing to rescan. A quick hack could better this situation by not flushing scan results in station (if the kernel retains these on an iftype change).
2022-02-28 19:49:37 +01:00
l_dbus_interface_method(interface, "Scan", 0, ap_dbus_scan, "", "");
l_dbus_interface_method(interface, "GetOrderedNetworks", 0,
ap_dbus_get_networks, "aa{sv}",
"", "networks");
ap: expose property 'Started' property on the Dbus
2018-07-30 00:32:55 +02:00
l_dbus_interface_property(interface, "Started", 0, "b",
ap_dbus_property_get_started, NULL);
ap: add Name property Its useful being able to refer to the network Name/SSID once an AP is started. For example opening an iwctl session with an already started AP provides no way of obtaining the SSID.
2021-02-02 00:52:14 +01:00
l_dbus_interface_property(interface, "Name", 0, "s",
ap_dbus_property_get_name, NULL);
ap: add support for scanning while in AP mode Scanning while in AP mode is somewhat of an edge case, but it does have some usefulness specifically with onboarding new devices, i.e. a new device starts an AP, a station connects and provides the new device with network credentials, the new device switches to station mode and connects to the desired network. In addition this could be used later for ACS (though this is a bit overkill for IWD's access point needs). Since AP performance is basically non-existant while scanning this feature is meant to be used in a limited scope. Two DBus API's were added which mirror the station interface: Scan and GetOrderedNetworks. Scan is no different than the station variant, and will perform an active scan on all channels. GetOrderedNetworks diverges from station and simply returns an array of dictionaries containing basic information about networks: { Name: <ssid> SignalStrength: <mBm> Security: <psk, open, or 8021x> } Limitations: - Hidden networks are not supported. This isn't really possible since the SSID's are unknown from the AP perspective. - Sharing scan results with station is not supported. This would be a convenient improvement in the future with respect to onboarding new devices. The scan could be performed in AP mode, then switch to station and connect immediately without needing to rescan. A quick hack could better this situation by not flushing scan results in station (if the kernel retains these on an iftype change).
2022-02-28 19:49:37 +01:00
l_dbus_interface_property(interface, "Scanning", 0, "b",
ap_dbus_property_get_scanning, NULL);
ap: add frequency to AP interface
2022-11-01 21:17:36 +01:00
l_dbus_interface_property(interface, "Frequency", 0, "u",
ap_dbus_property_get_freq, NULL);
ap: change PairwiseCiphers to string array A comma separated list as a string was ok for pure display purposes but if any processing needed to be done on these values by external consumers it really makes more sense to use a DBus array.
2022-11-04 17:37:58 +01:00
l_dbus_interface_property(interface, "PairwiseCiphers", 0, "as",
ap: add PairwiseCiphers/GroupCipher to dbus interface Expose these values on the DBus interface so clients can view them.
2022-11-02 23:51:11 +01:00
ap_dbus_property_get_pairwise, NULL);
l_dbus_interface_property(interface, "GroupCipher", 0, "s",
ap_dbus_property_get_group, NULL);
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
}
static void ap_destroy_interface(void *user_data)
{
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
struct ap_if_data *ap_if = user_data;
if (ap_if->pending) {
struct l_dbus_message *reply;
reply = dbus_error_aborted(ap_if->pending);
dbus_pending_reply(&ap_if->pending, reply);
}
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
if (ap_if->ap)
ap_free(ap_if->ap);
l_free(ap_if);
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
}
ap: add AP diagnostic interface This adds a new AccessPointDiagnostic interface. This interface provides similar low level functionality as StationDiagnostic, but for when IWD is in AP mode. This uses netdev_get_all_stations which will dump all stations, parse, and return each station in an individual callback. Once the dump is complete the destroy is called and all data is packaged as an array of dictionaries.
2021-01-22 21:38:45 +01:00
struct diagnostic_data {
struct l_dbus_message *pending;
struct l_dbus_message_builder *builder;
};
static void ap_get_station_cb(const struct diagnostic_station_info *info,
void *user_data)
{
struct diagnostic_data *data = user_data;
/* First station info */
if (!data->builder) {
struct l_dbus_message *reply =
l_dbus_message_new_method_return(data->pending);
data->builder = l_dbus_message_builder_new(reply);
l_dbus_message_builder_enter_array(data->builder, "a{sv}");
}
l_dbus_message_builder_enter_array(data->builder, "{sv}");
dbus_append_dict_basic(data->builder, "Address", 's',
util_address_to_string(info->addr));
diagnostic_info_to_dict(info, data->builder);
l_dbus_message_builder_leave_array(data->builder);
}
static void ap_get_station_destroy(void *user_data)
{
struct diagnostic_data *data = user_data;
struct l_dbus_message *reply;
if (!data->builder) {
reply = l_dbus_message_new_method_return(data->pending);
data->builder = l_dbus_message_builder_new(reply);
l_dbus_message_builder_enter_array(data->builder, "a{sv}");
}
l_dbus_message_builder_leave_array(data->builder);
reply = l_dbus_message_builder_finalize(data->builder);
l_dbus_message_builder_destroy(data->builder);
dbus_pending_reply(&data->pending, reply);
l_free(data);
}
static struct l_dbus_message *ap_dbus_get_diagnostics(struct l_dbus *dbus,
struct l_dbus_message *message, void *user_data)
{
struct ap_if_data *ap_if = user_data;
struct diagnostic_data *data;
int ret;
data = l_new(struct diagnostic_data, 1);
data->pending = l_dbus_message_ref(message);
ret = netdev_get_all_stations(ap_if->ap->netdev, ap_get_station_cb,
data, ap_get_station_destroy);
if (ret < 0) {
l_dbus_message_unref(data->pending);
l_free(data);
return dbus_error_from_errno(ret, message);
}
return NULL;
}
static void ap_setup_diagnostic_interface(struct l_dbus_interface *interface)
{
l_dbus_interface_method(interface, "GetDiagnostics", 0,
ap_dbus_get_diagnostics,
"aa{sv}", "", "diagnostic");
}
static void ap_diagnostic_interface_destroy(void *user_data)
{
}
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
static void ap_add_interface(struct netdev *netdev)
device: added DEVICE_EVENT_MODE_CHANGED Rather than have device.c manage the creation/removal of AP/AdHoc interfaces this new event was introduced. Now anyone can listen for device events and if the mode changes handle accordingly. This fixes potential memory leaks in WSC when switching modes as well.
2018-07-18 01:39:54 +02:00
{
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
struct ap_if_data *ap_if;
device: added DEVICE_EVENT_MODE_CHANGED Rather than have device.c manage the creation/removal of AP/AdHoc interfaces this new event was introduced. Now anyone can listen for device events and if the mode changes handle accordingly. This fixes potential memory leaks in WSC when switching modes as well.
2018-07-18 01:39:54 +02:00
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
/*
* TODO: Check wiphy supported channels and NL80211_ATTR_TX_FRAME_TYPES
*/
device: added DEVICE_EVENT_MODE_CHANGED Rather than have device.c manage the creation/removal of AP/AdHoc interfaces this new event was introduced. Now anyone can listen for device events and if the mode changes handle accordingly. This fixes potential memory leaks in WSC when switching modes as well.
2018-07-18 01:39:54 +02:00
/* just allocate/set device, Start method will complete setup */
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
ap_if = l_new(struct ap_if_data, 1);
ap_if->netdev = netdev;
device: added DEVICE_EVENT_MODE_CHANGED Rather than have device.c manage the creation/removal of AP/AdHoc interfaces this new event was introduced. Now anyone can listen for device events and if the mode changes handle accordingly. This fixes potential memory leaks in WSC when switching modes as well.
2018-07-18 01:39:54 +02:00
/* setup ap dbus interface */
l_dbus_object_add_interface(dbus_get_bus(),
ap: Put a public api between AP logic and DBus code Separate AP logic from DBus code, add a public API to make the AP logic reusable from other files.
2020-08-04 16:53:01 +02:00
netdev_get_path(netdev), IWD_AP_INTERFACE, ap_if);
device: added DEVICE_EVENT_MODE_CHANGED Rather than have device.c manage the creation/removal of AP/AdHoc interfaces this new event was introduced. Now anyone can listen for device events and if the mode changes handle accordingly. This fixes potential memory leaks in WSC when switching modes as well.
2018-07-18 01:39:54 +02:00
}
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
static void ap_remove_interface(struct netdev *netdev)
device: added DEVICE_EVENT_MODE_CHANGED Rather than have device.c manage the creation/removal of AP/AdHoc interfaces this new event was introduced. Now anyone can listen for device events and if the mode changes handle accordingly. This fixes potential memory leaks in WSC when switching modes as well.
2018-07-18 01:39:54 +02:00
{
l_dbus_object_remove_interface(dbus_get_bus(),
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
netdev_get_path(netdev), IWD_AP_INTERFACE);
ap: remove diagnostic interface when netdev goes down This fixes a crash associated with toggling the iftype to AP mode then calling GetDiagnostics. The diagnostic interface is never cleaned up when netdev goes down so DBus calls can still be made which ends up crashing since the AP interface objects are no longer valid. Running the following iwctl commands in a script (once or twice) triggers this crash reliably: iwctl device wlp2s0 set-property Mode ap iwctl device wlp2s0 set-property Mode station iwctl device wlp2s0 set-property Mode ap iwctl ap wlp2s0 start myssid secret123 iwctl ap wlp2s0 show ++++++++ backtrace ++++++++ 0 0x7f8f1a8fe320 in /lib64/libc.so.6 1 0x451f35 in ap_dbus_get_diagnostics() at src/ap.c:4043 2 0x4cdf5a in _dbus_object_tree_dispatch() at ell/dbus-service.c:1815 3 0x4bffc7 in message_read_handler() at ell/dbus.c:285 4 0x4b5d7b in io_callback() at ell/io.c:120 5 0x4b489b in l_main_iterate() at ell/main.c:476 6 0x4b49a6 in l_main_run() at ell/main.c:519 7 0x4b4cd9 in l_main_run_with_signal() at ell/main.c:645 8 0x404f5b in main() at src/main.c:600 9 0x7f8f1a8e8b75 in /lib64/libc.so.6 +++++++++++++++++++++++++++
2022-07-01 18:31:08 +02:00
l_dbus_object_remove_interface(dbus_get_bus(),
netdev_get_path(netdev), IWD_AP_DIAGNOSTIC_INTERFACE);
device: added DEVICE_EVENT_MODE_CHANGED Rather than have device.c manage the creation/removal of AP/AdHoc interfaces this new event was introduced. Now anyone can listen for device events and if the mode changes handle accordingly. This fixes potential memory leaks in WSC when switching modes as well.
2018-07-18 01:39:54 +02:00
}
netdev/device: Combine watches There was somewhat overlapping functionality in the device_watch infrastructure as well as the netdev_event_watch. This commit combines the two into a single watch based on the netdev object and cleans up the various interface additions / removals. With this commit the interfaces are created when the netdev/device is switched to Powered=True state AND when the netdev iftype is also in the correct state for that interface. If the device is brought down, then all interfaces except the .Device interface are removed. This will make it easy to implement Device.Mode property properly since most nl80211 devices need to be brought into Powered=False state prior to switching the iftype.
2018-08-18 06:40:49 +02:00
static void ap_netdev_watch(struct netdev *netdev,
enum netdev_watch_event event, void *userdata)
device: added DEVICE_EVENT_MODE_CHANGED Rather than have device.c manage the creation/removal of AP/AdHoc interfaces this new event was introduced. Now anyone can listen for device events and if the mode changes handle accordingly. This fixes potential memory leaks in WSC when switching modes as well.
2018-07-18 01:39:54 +02:00
{
switch (event) {
netdev/device: Combine watches There was somewhat overlapping functionality in the device_watch infrastructure as well as the netdev_event_watch. This commit combines the two into a single watch based on the netdev object and cleans up the various interface additions / removals. With this commit the interfaces are created when the netdev/device is switched to Powered=True state AND when the netdev iftype is also in the correct state for that interface. If the device is brought down, then all interfaces except the .Device interface are removed. This will make it easy to implement Device.Mode property properly since most nl80211 devices need to be brought into Powered=False state prior to switching the iftype.
2018-08-18 06:40:49 +02:00
case NETDEV_WATCH_EVENT_UP:
case NETDEV_WATCH_EVENT_NEW:
ap, adhoc: On NETDEV_WATCH_EVENT_NEW check device is UP Since the interfaces are not supposed to exist when the device is DOWN (we destroy the interfaces on NETDEV_WATCH_EVENT_DOWN too), don't create the interfaces if the device hasn't been brought up yet.
2018-09-22 18:48:22 +02:00
if (netdev_get_iftype(netdev) == NETDEV_IFTYPE_AP &&
netdev_get_is_up(netdev))
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
ap_add_interface(netdev);
netdev/device: Combine watches There was somewhat overlapping functionality in the device_watch infrastructure as well as the netdev_event_watch. This commit combines the two into a single watch based on the netdev object and cleans up the various interface additions / removals. With this commit the interfaces are created when the netdev/device is switched to Powered=True state AND when the netdev iftype is also in the correct state for that interface. If the device is brought down, then all interfaces except the .Device interface are removed. This will make it easy to implement Device.Mode property properly since most nl80211 devices need to be brought into Powered=False state prior to switching the iftype.
2018-08-18 06:40:49 +02:00
break;
case NETDEV_WATCH_EVENT_DOWN:
case NETDEV_WATCH_EVENT_DEL:
ap: Make AP use netdev instead of device The information is always on the netdev object anyway, so this just makes it explicit.
2018-08-20 06:12:29 +02:00
ap_remove_interface(netdev);
netdev/device: Combine watches There was somewhat overlapping functionality in the device_watch infrastructure as well as the netdev_event_watch. This commit combines the two into a single watch based on the netdev object and cleans up the various interface additions / removals. With this commit the interfaces are created when the netdev/device is switched to Powered=True state AND when the netdev iftype is also in the correct state for that interface. If the device is brought down, then all interfaces except the .Device interface are removed. This will make it easy to implement Device.Mode property properly since most nl80211 devices need to be brought into Powered=False state prior to switching the iftype.
2018-08-18 06:40:49 +02:00
break;
device: added DEVICE_EVENT_MODE_CHANGED Rather than have device.c manage the creation/removal of AP/AdHoc interfaces this new event was introduced. Now anyone can listen for device events and if the mode changes handle accordingly. This fixes potential memory leaks in WSC when switching modes as well.
2018-07-18 01:39:54 +02:00
default:
break;
}
}
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
static int ap_init(void)
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
{
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
const struct l_settings *settings = iwd_get_config();
netdev/device: Combine watches There was somewhat overlapping functionality in the device_watch infrastructure as well as the netdev_event_watch. This commit combines the two into a single watch based on the netdev object and cleans up the various interface additions / removals. With this commit the interfaces are created when the netdev/device is switched to Powered=True state AND when the netdev iftype is also in the correct state for that interface. If the device is brought down, then all interfaces except the .Device interface are removed. This will make it easy to implement Device.Mode property properly since most nl80211 devices need to be brought into Powered=False state prior to switching the iftype.
2018-08-18 06:40:49 +02:00
netdev_watch = netdev_watch_add(ap_netdev_watch, NULL, NULL);
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
l_dbus_register_interface(dbus_get_bus(), IWD_AP_INTERFACE,
device/ap: Moved AP dbus interface into ap.c Now, a user can setup an AP as follows: - Set device "Mode" to ap (ap interface will appear on bus) - call "Start()" on AP interface Issuing "Stop()" on the AP interface will stop and cleanup the internal AP structures, but the AP interface will remain up. To shutdown completely the device Mode must be switched back to station. If the AP interface is running, the Mode can directly be switched to station without calling Stop; this has the same effect and will take down the AP interface.
2018-06-30 01:29:17 +02:00
ap_setup_interface, ap_destroy_interface, false);
ap: add AP diagnostic interface This adds a new AccessPointDiagnostic interface. This interface provides similar low level functionality as StationDiagnostic, but for when IWD is in AP mode. This uses netdev_get_all_stations which will dump all stations, parse, and return each station in an individual callback. Once the dump is complete the destroy is called and all data is packaged as an array of dictionaries.
2021-01-22 21:38:45 +01:00
l_dbus_register_interface(dbus_get_bus(), IWD_AP_DIAGNOSTIC_INTERFACE,
ap_setup_diagnostic_interface,
ap_diagnostic_interface_destroy, false);
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
/*
ap: Refactor global address pool loading Deprecate the global [General].APRanges setting in favour of [IPv4].APAddressPool with an extended (but backwards-compatible) syntax. Drop the existing address pool creation code. The new APAddressPool setting has the same syntax as the profile-local [IPv4].Address setting and the subnet selection code will fall back to the global setting if it's missing, this way we use common code to handle both settings.
2021-05-28 02:47:33 +02:00
* Enable network configuration and DHCP only if
* [General].EnableNetworkConfiguration is true.
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
*/
treewide: Parse EnableNetworkConfiguration in one place Add netconfig_enabled() and use that in all places that want to know whether network configuration is enabled. Drop the enable_network_config deprecated setting, which was only being handled in one of these 5 or so places.
2021-10-21 10:50:30 +02:00
if (netconfig_enabled()) {
ap: Remove support for deprecated APRanges setting In preparation for 2.0 release, remove settings that were marked deprecated during 1.x cycle.
2022-09-29 19:59:16 +02:00
global_addr4_strs =
l_settings_get_string_list(settings, "IPv4",
"APAddressPool", ',');
ap: Fix spurious warning message The warning message would be printed even if no setting was present Fixes: ac976c6f7147 ("ap: Remove support for deprecated APRanges setting")
2022-10-24 18:03:08 +02:00
if (global_addr4_strs && !global_addr4_strs[0]) {
ap: Remove support for deprecated APRanges setting In preparation for 2.0 release, remove settings that were marked deprecated during 1.x cycle.
2022-09-29 19:59:16 +02:00
l_error("Can't parse the [IPv4].APAddressPool "
ap: Refactor global address pool loading Deprecate the global [General].APRanges setting in favour of [IPv4].APAddressPool with an extended (but backwards-compatible) syntax. Drop the existing address pool creation code. The new APAddressPool setting has the same syntax as the profile-local [IPv4].Address setting and the subnet selection code will fall back to the global setting if it's missing, this way we use common code to handle both settings.
2021-05-28 02:47:33 +02:00
"setting as a string list");
ap: Remove support for deprecated APRanges setting In preparation for 2.0 release, remove settings that were marked deprecated during 1.x cycle.
2022-09-29 19:59:16 +02:00
l_strv_free(global_addr4_strs);
global_addr4_strs = NULL;
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
}
ap: Refactor global address pool loading Deprecate the global [General].APRanges setting in favour of [IPv4].APAddressPool with an extended (but backwards-compatible) syntax. Drop the existing address pool creation code. The new APAddressPool setting has the same syntax as the profile-local [IPv4].Address setting and the subnet selection code will fall back to the global setting if it's missing, this way we use common code to handle both settings.
2021-05-28 02:47:33 +02:00
/* Fall back to 192.168.0.0/16 */
if (!global_addr4_strs)
global_addr4_strs =
l_strv_append(NULL, "192.168.0.0/16");
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
}
ap: allow DHCP settings in provisioning files Users can now supply an AP provisioning file containing an [IPv4] section and define various DHCP settings: [IPv4] Address=<address> Netmask=<netmask> Gateway=<gateway> IPRange=<start_address>,<end_address> DNSList=<dns1>,<dns2>,...<dnsN> LeaseTime=<lease_time> There are a few notes/requirements to keep in mind when using a provisioning file: - All settings are optional but [IPv4].Address is required if the interface does not already have an address set. - If no [IPv4].Address is defined in the provisioning file and the AP interface does not already have an address set, StartWithConfig() will fail with -EINVAL. - If a provisioning file is provided it will take precedence, and the AP will not pull from the IP pool. - A provisioning file containing an IPv4 section assumes DHCP is being enabled and will override [General].EnableNetworkConfiguration. - Any address that AP sets on the interface will be deleted when the AP is stopped.
2020-11-02 19:28:50 +01:00
rtnl = iwd_get_rtnl();
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
return 0;
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
}
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
static void ap_exit(void)
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
{
netdev/device: Combine watches There was somewhat overlapping functionality in the device_watch infrastructure as well as the netdev_event_watch. This commit combines the two into a single watch based on the netdev object and cleans up the various interface additions / removals. With this commit the interfaces are created when the netdev/device is switched to Powered=True state AND when the netdev iftype is also in the correct state for that interface. If the device is brought down, then all interfaces except the .Device interface are removed. This will make it easy to implement Device.Mode property properly since most nl80211 devices need to be brought into Powered=False state prior to switching the iftype.
2018-08-18 06:40:49 +02:00
netdev_watch_remove(netdev_watch);
device: added DEVICE_EVENT_MODE_CHANGED Rather than have device.c manage the creation/removal of AP/AdHoc interfaces this new event was introduced. Now anyone can listen for device events and if the mode changes handle accordingly. This fixes potential memory leaks in WSC when switching modes as well.
2018-07-18 01:39:54 +02:00
l_dbus_unregister_interface(dbus_get_bus(), IWD_AP_INTERFACE);
ap: add support for DHCPv4 server The DHCP server can be enabled by enabling network configuration with [General].EnableNetworkConfiguration. If an IP is not set on the interface before the AP is started a valid IP range must also be provided under [General].APRanges in IP prefix format e.g. [General] EnableNetworkConfiguration=true APRanges=192.168.1.1/24 Each AP started will get assigned a new subnet within the range specified by APRanges as to not conflict with other AP interfaces. If there are no subnets left in the pool when an AP is started it will fail with -EEXIST. Any AP's that are stopped will release their subnet back into the pool to be used with other APs. The DHCP IP pool will be automatically chosen by the ELL DHCP implementation (+1 the AP's IP to *.254). The remaining DHCP settings will be defaults chosen by ELL (DNS, lease time, etc).
2020-10-27 20:57:47 +01:00
ap: Refactor global address pool loading Deprecate the global [General].APRanges setting in favour of [IPv4].APAddressPool with an extended (but backwards-compatible) syntax. Drop the existing address pool creation code. The new APAddressPool setting has the same syntax as the profile-local [IPv4].Address setting and the subnet selection code will fall back to the global setting if it's missing, this way we use common code to handle both settings.
2021-05-28 02:47:33 +02:00
l_strv_free(global_addr4_strs);
ap: Add AP mode api Very basic WPA2-PSK only access point mode API with stubs of the start and stop functions.
2017-09-12 02:45:43 +02:00
}
ap: Convert to IWD_MODULE
2019-05-28 21:43:42 +02:00
IWD_MODULE(ap, ap_init, ap_exit)
treewide: Add missing netdev module dependencies
2020-08-20 18:07:51 +02:00
IWD_MODULE_DEPENDS(ap, netdev);
Copy Permalink