Jouni Malinen [Tue, 21 Jan 2020 10:57:07 +0000 (12:57 +0200)]
tests: Remove mesh SAE Password Identifier test cases for now
IEEE P802.11-REVmd was modified to require H2E to be used whenever
Password Identifier is used with SAE. Since wpa_supplicant and mac80211
do not yet support SAE H2E in mesh, Password Identifier cannot be used
in mesh cases. Remove the test cases that verified this behavior for now
to allow H2E to be required per updated REVmd definition. These test
cases will be restored once H2E is fully functionality in mesh cases.
Jouni Malinen [Mon, 20 Jan 2020 18:23:48 +0000 (20:23 +0200)]
SAE H2E: Check H2E-only BSS membership selector only if SAE is enabled
This BSS membership selector has impact only for SAE functionality, so
ignore it when configured not to use SAE. This allows WPA-PSK connection
to and AP that advertises WPA-PSK and SAE while requiring H2E for SAE.
Johannes Berg [Wed, 15 Jan 2020 11:48:43 +0000 (12:48 +0100)]
tests: parallel-vm: allow running without curses
Allow running without curses, in which case the log is simply written to
stdout instead of a file. This is useful for automated (but parallel)
testing. Note that in most cases, you'd want to specify --debug, and so
I added a .rstrip() there on the lines to clean that up a bit.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Hai Shalom [Fri, 27 Dec 2019 17:44:49 +0000 (09:44 -0800)]
EAP-SIM/AKA peer: Add support for EAP Method prefix
Add support for EAP method prefix in the anonymous identity
used during EAP-SIM/AKA/AKA' authentication when encrypted IMSI
is used. The prefix is a single character that indicates which
EAP method is required by the client.
Jouni Malinen [Thu, 9 Jan 2020 22:19:27 +0000 (00:19 +0200)]
tests: Call stop_sigma_dut() in more failure cases
Some of the sigma_dut test cases were not yet using try/finally to
ensure stop_sigma_dut() gets called. That could result in not logging
all failure reasons in the log and getting stuck with being unable to
start new sigma_dut processes after failed test cases.
Vamsi Krishna [Tue, 24 Dec 2019 12:01:39 +0000 (17:31 +0530)]
ACS: Populate channel config from external ACS per documented behavior
Based on the now documented seg0/seg1 values from offloaded ACS, there
is a mismatch between the driver interface and internal hostapd use.
The value of segment0 field in ACS results is the index of the channel
center frequency for 20 MHz, 40 MHz, and 80M Hz channels. The value is
the center frequency index of the primary 80 MHz segment for 160 MHz and
80+80 MHz channels.
The value of segment1 field in ACS results is zero for 20 MHz, 40 MHz,
and 80 MHz channels. The value is the index of the channel center
frequency for 160 MHz channels and the center frequency index of the
secondary 80 MHz segment for 80+80 MHz channels.
However, in struct hostapd_config, for 160 MHz channels, the value of
the segment0 field is the index of the channel center frequency of 160
MHz channel and the value of the segment1 field is zero. Map the values
from ACS event into hostapd_config fields accordingly.
Vamsi Krishna [Tue, 24 Dec 2019 12:46:37 +0000 (18:16 +0530)]
ACS: Update documentation of external ACS results event parameters
Update the documentation with values to be sent for seg0 and seg1 fields
in external ACS result event for 20 MHz, 40 MHz, 80 MHz, 160 MHz, and
80+80 MHz channels. These values match the changes done to definitions
of seg0 and seg1 fields in the IEEE 802.11 standard.
This vendor command had not previously been documented in this level of
detail and had not actually been used for the only case that could have
two different interpretation (160 MHz) based on which version of IEEE
802.11 standard is used.
Vamsi Krishna [Tue, 24 Dec 2019 12:02:29 +0000 (17:32 +0530)]
6 GHz: Fix Channel Width value for 80+80 in 6 GHZ Operation Info field
The Channel Width field value is 0 for 20 MHz, 1 for 40 MHz, 2 for 80
MHz, and 3 for both 160 MHz and 80+80 MHz channels. The 80+80 MHz case
was not addressed previously correctly since it cannot be derived from
seg0 only.
The Channel Center Frequency Segment 0 field value is the index of
channel center frequency for 20 MHz, 40 MHz, and 80 MHz channels. The
value is the center frequency index of the primary 80 MHz segment for
160 MHz and 80+80 MHz channels.
The Channel Center Frequency Segment 1 field value is zero for 20 MHz,
40 MHz, and 80 MHz channels. The value is the index of the channel
center frequency for 160 MHz channel and the center frequency index of
the secondary 80 MHz segment for 80+80 MHz channels.
Add the new set_key() parameter "key_flag" to provide more specific
description of what type of a key is being configured. This is needed to
be able to add support for "Extended Key ID for Individually Addressed
Frames" from IEEE Std 802.11-2016. In addition, this may be used to
replace the set_tx boolean eventually once all the driver wrappers have
moved to using the new key_flag.
The following flag are defined:
KEY_FLAG_MODIFY
Set when an already installed key must be updated.
So far the only use-case is changing RX/TX status of installed
keys. Must not be set when deleting a key.
KEY_FLAG_DEFAULT
Set when the key is also a default key. Must not be set when
deleting a key. (This is the replacement for set_tx.)
KEY_FLAG_RX
The key is valid for RX. Must not be set when deleting a key.
KEY_FLAG_TX
The key is valid for TX. Must not be set when deleting a key.
KEY_FLAG_GROUP
The key is a broadcast or group key.
KEY_FLAG_PAIRWISE
The key is a pairwise key.
KEY_FLAG_PMK
The key is a Pairwise Master Key (PMK).
Predefined and needed flag combinations so far are:
KEY_FLAG_GROUP_RX_TX
WEP key not used as default key (yet).
KEY_FLAG_GROUP_RX_TX_DEFAULT
Default WEP or WPA-NONE key.
KEY_FLAG_GROUP_RX
GTK key valid for RX only.
KEY_FLAG_GROUP_TX_DEFAULT
GTK key valid for TX only, immediately taking over TX.
KEY_FLAG_PAIRWISE_RX_TX
Pairwise key immediately becoming the active pairwise key.
KEY_FLAG_PAIRWISE_RX
Pairwise key not yet valid for TX. (Only usable with Extended Key ID
support.)
KEY_FLAG_PAIRWISE_RX_TX_MODIFY
Enable TX for a pairwise key installed with KEY_FLAG_PAIRWISE_RX.
KEY_FLAG_RX_TX
Not a valid standalone key type and can only used in combination
with other flags to mark a key for RX/TX.
This commit is not changing any functionality. It just adds the new
key_flag to all hostapd/wpa_supplicant set_key() functions without using
it, yet.
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
Markus Theil [Wed, 8 Jan 2020 23:03:02 +0000 (00:03 +0100)]
tests: Fix Python sleep function
Current Python versions have no os.sleep(), use time.sleep() instead.
module 'os' has no attribute 'sleep'
Traceback (most recent call last):
File "./run-tests.py", line 521, in main
t(dev, apdev)
File "/home/mtheil/hostap/tests/hwsim/test_pmksa_cache.py", line 356, in test_pmksa_cache_expiration
hapd.wait_ptkinitdone(dev[0].own_addr())
File "/home/mtheil/hostap/tests/hwsim/hostapd.py", line 282, in wait_ptkinitdone
os.sleep(0.1)
Signed-off-by: Markus Theil <markus.theil@tu-ilmenau.de>
Jouni Malinen [Wed, 8 Jan 2020 22:47:03 +0000 (00:47 +0200)]
nl80211: Pass set_key() parameter struct to wpa_driver_nl80211_set_key()
This is the function that actually uses the parameters, so pass the full
parameter struct to it instead of hiding the struct from it in the
simple wrapper.
This is in preparation for adding support to use a single WLAN netdev
with VLAN operations offloaded to the driver. No functional changes are
included in this commit.
Jouni Malinen [Wed, 8 Jan 2020 20:15:18 +0000 (22:15 +0200)]
driver: Move set_key() parameters into a struct
This makes it more convenient to add, remove, and modify the parameters
without always having to update every single driver_*.c implementation
of this callback function.
Jouni Malinen [Wed, 8 Jan 2020 18:52:26 +0000 (20:52 +0200)]
SAE: A bit optimized sae_confirm_immediate=2 for testing purposes
sae_confirm_immediate=2 can now be used in CONFIG_TESTING_OPTIONS=y
builds to minimize the latency between SAE Commit and SAE Confirm by
postponing transmission of SAE Commit until the SAE Confirm frame is
generated. This does not have significant impact, but can get the frames
tiny bit closer to each other over the air to increase testing coverage.
The only difference between sae_confirm_immediate 1 and 2 is in the
former deriving KCK, PMK, PMKID, and CN between transmission of the
frames (i.e., a small number of hash operations).
Vamsi Krishna [Wed, 11 Dec 2019 12:57:47 +0000 (18:27 +0530)]
Do not select APs found on disabled channels for connection
If a channel list changed event is received after a scan and before
selecting a BSS for connection, a BSS found on a now disabled channel
may get selected for connection. The connect request issued with the BSS
found on a disabled channel is rejected by cfg80211. Filter out the BSSs
found on disabled channels and select from the other BSSs found on
enabled channels to avoid unnecessary connection attempts that are bound
to fail.
The channel list information will be updated by the driver in cases like
country code update, disabling/enabling specific bands, etc. which can
occur between the scan and connection attempt.
Markus Theil [Wed, 8 Jan 2020 10:04:52 +0000 (11:04 +0100)]
tests: Fix undefined behavior in module tests
Test: wpa_supplicant module tests
../src/utils/utils_module_tests.c:933:7: runtime error: left shift of 1 by 31 places cannot be represented in type 'int'
Signed-off-by: Markus Theil <markus.theil@tu-ilmenau.de>
Jouni Malinen [Sun, 5 Jan 2020 19:18:18 +0000 (21:18 +0200)]
nl80211: Allow control port to be disabled with a driver param
This is mainly for testing purposes to allow wpa_supplicant and hostapd
functionality to be tested both with and without using the nl80211
control port which is by default used whenever supported by the driver.
control_port=0 driver parameter will prevent that from happening.
Markus Theil [Fri, 3 Jan 2020 15:17:42 +0000 (16:17 +0100)]
nl80211: Report control port RX events
This allows EAPOL frames to be received over the separate controlled
port once rest of the driver interface is ready for this. By itself,
this commit does not actually change behavior since cfg80211 will not be
delivering these events without them being explicitly requested.
Signed-off-by: Markus Theil <markus.theil@tu-ilmenau.de>
Markus Theil [Fri, 3 Jan 2020 15:17:42 +0000 (16:17 +0100)]
Add no_encrypt flag for control port TX
In order to correctly encrypt rekeying frames, wpa_supplicant now checks
if a PTK is currently installed and sets the corresponding encrypt
option for tx_control_port().
Signed-off-by: Markus Theil <markus.theil@tu-ilmenau.de>
Brendan Jackman [Fri, 3 Jan 2020 15:17:41 +0000 (16:17 +0100)]
wpa_supplicant: Send EAPOL frames over nl80211 where available
Linux kernel v4.17 added the ability to request sending control port
frames via nl80211 instead of a normal network socket. Doing this
provides the device driver with ordering information between the
control port frames and the installation of keys. This empowers it to
avoid race conditions between, for example, PTK replacement and the
sending of frame 4 of the 4-way rekeying handshake in an RSNA. The
key difference between a TX_CONTROL_PORT and normal socket send is
that the device driver will certainly get any EAPOL frames comprising
a 4-way handshake before it gets the key installation call
for the derived key. By flushing its TX buffers it can then ensure
that no pending EAPOL frames are inadvertently encrypted with a key
that the peer will not yet have installed.
Update the RSN supplicant system to use this new operation for sending
EAPOL-Key frames when the driver reports that this capability is
available; otherwise, fall back to a normal Ethernet TX.
I have tested this on DMG (11ad/ay) devices with an out-of-tree Linux
driver that does not use mac80211. Without this patch I consistently see
PTK rekeying fail if message 4/4 shares a stream with other in-flight
traffic. With this patch, and the driver updated to flush the relevant TX
queue before overwriting a PTK (knowing, now, that if there was a message
4/4 related to the key installation, it has already entered the driver
queue), rekeying is reliable.
There is still data loss surrounding key installation - this problem is
alluded to in IEEE Std 802.11-2016, 12.6.21, where extended Key ID
support is described as the eventual solution. This patch aims to at
least prevent rekeying from totally breaking the association, in a way
that works on kernels as far back as 4.17 (as per Alexander Wetzel
extended Key ID support should be possible on 5.2).
See http://lists.infradead.org/pipermail/hostap/2019-May/040089.html for
a little more context.
Brendan Jackman [Fri, 3 Jan 2020 15:17:41 +0000 (16:17 +0100)]
nl80211: Control port over nl80211 helpers
Linux kernel v4.17 added the ability to request sending controlled port
frames (e.g., IEEE 802.1X controlled port EAPOL frames) via nl80211
instead of a normal network socket. Doing this provides the device
driver with ordering information between the control port frames and the
installation of keys. This empowers it to avoid race conditions between,
for example, PTK replacement and the sending of frame 4 of the 4-way
rekeying handshake in an RSNA. The key difference between the specific
control port and normal socket send is that the device driver will
certainly get any EAPOL frames comprising a 4-way handshake before it
gets the key installation call for the derived key. By flushing its TX
buffers it can then ensure that no pending EAPOL frames are
inadvertently encrypted with a key that the peer will not yet have
installed.
Add a CONTROL_PORT flag to the hostap driver API to report driver
capability for using a separate control port for EAPOL frames. This
operation is exactly like an Ethernet send except for the extra ordering
information it provides for device drivers. The nl80211 driver is
updated to support this operation when the device reports support for
NL80211_EXT_FEATURE_CONTROL_PORT_OVER_NL80211. Also add a driver op
tx_control_port() for request a frame to be sent over the controlled
port.
Jouni Malinen [Sun, 5 Jan 2020 16:32:10 +0000 (18:32 +0200)]
driver: Remove unused send_ether() driver op
This was used only for FT RRB sending with driver_test.c and
driver_test.c was removed more than five years ago, so there is no point
in continuing to maintain this driver op.
Bilal Hatipoglu [Fri, 3 Jan 2020 08:58:26 +0000 (11:58 +0300)]
WPS: Add application extension data to WPS IE
Application Extension attribute is defined in WSC tech spec v2.07 page
104. Allow hostapd to be configured to add this extension into WPS IE in
Beacon and Probe Response frames. The implementation is very similar to
vendor extension.
A new optional entry called "wps_application_ext" is added to hostapd
config file to configure this. It enodes the payload of the Application
Extension attribute in hexdump format.
Signed-off-by: Veli Demirel <veli.demirel@airties.com> Signed-off-by: Bilal Hatipoglu <bilal.hatipoglu@airties.com>
Jouni Malinen [Sat, 4 Jan 2020 21:11:28 +0000 (23:11 +0200)]
P2P: Move p2p_long_listen into struct wpa_global
This variable is not specific to any P2P group interface and since it
was already used through global->p2p_init_wpa_s, it is cleaner to simply
move this to the global structure so that there is a single variable
instead of per-interface variables and need to pick the correct
interface.
Benjamin Berg [Fri, 3 Jan 2020 21:18:52 +0000 (22:18 +0100)]
tests: Test p2p_long_listen longer than remain-on-channel
This tests an error, where the p2p_long_listen information from the
wrong device was used internally in wpa_supplicant when using the
separate P2P Device interface.
Benjamin Berg [Fri, 3 Jan 2020 21:18:51 +0000 (22:18 +0100)]
P2P: Always use global p2p_long_listen
The p2p_long_listen value was set on the control wpa_s struct while in a
lot of cases it operated on the p2p struct. Explicitly use the global
p2p_init_wpa_s struct in cases where we might not be operating on it
already.
Without this, simply starting a p2p_listen operation (e.g., using
wpa_cli) will not work properly. As the p2p_long_listen is set on the
controlling interface and wpas_p2p_cancel_remain_on_channel_cb() uses
p2p_init_wpa_s, it would not actually work. This results in
wpa_supplicant stopping listening after the maximum remain-on-channel
time passes when using a separate P2P Device interface.
Jouni Malinen [Sat, 4 Jan 2020 18:23:05 +0000 (20:23 +0200)]
nl80211: Use monitor interface for sending no-encrypt test frames
Since NL80211_CMD_FRAME does not allow encryption to be disabled for the
frame, add a monitor interface temporarily for cases where this type of
no-encrypt frames are to be sent. The temporary monitor interface is
removed immediately after sending the frame.
This is testing functionality (only in CONFIG_TESTING_OPTIONS=y builds)
that is used for PMF testing where the AP can use this to inject an
unprotected Robust Management frame (mainly, Deauthentication or
Disassociation frame) even in cases where PMF has been negotiated for
the association.
Jouni Malinen [Sat, 4 Jan 2020 11:05:26 +0000 (13:05 +0200)]
Allow testing override for GTK/IGTK RSC from AP to STA
The new hostapd gtk_rsc_override and igtk_rsc_override configuration
parameters can be used to set an override value for the RSC that the AP
advertises for STAs for GTK/IGTK. The contents of those parameters is a
hexdump of the RSC in little endian byte order.
This functionality is available only in CONFIG_TESTING_OPTIONS=y builds.
This can be used to verify that stations implement initial RSC
configuration correctly for GTK/ and IGTK.
Jouni Malinen [Fri, 3 Jan 2020 16:06:01 +0000 (18:06 +0200)]
SME: Postpone current BSSID clearing until IEs are prepared
sme_send_authentication() could fail before actually requesting the
driver to authenticate with a new AP. This could happen after
wpa_s->bssid got cleared even though in such a case, the old association
is maintained and still valid. This can result in unexpected behavior
since wpa_s->bssid would not match the current BSSID anymore.
Fix this by postponing clearing of wpa_s->bssid until the IE preparation
has been completed successfully.
Ben Greear [Tue, 19 Mar 2019 21:34:24 +0000 (14:34 -0700)]
nl80211: Don't set offchan-OK flag if doing on-channel frame in AP mode
I saw a case where the kernel's cfg80211 rejected hostapd's attempt to
send a neighbor report response because nl80211 flagged the frame as
offchannel-OK, but kernel rejects because channel was 100 (DFS) and so
kernel failed thinking it was constrained by DFS/CAC requirements that
do not allow the operating channel to be left (at least in FCC).
Don't set the packet as off-channel OK if we are transmitting on the
current operating channel of an AP to avoid such issues with
transmission of Action frames.
Signed-off-by: Ben Greear <greearb@candelatech.com>
Jouni Malinen [Fri, 3 Jan 2020 13:23:49 +0000 (15:23 +0200)]
nl80211: Rename send_action_cookie to send_frame_cookie
This is to match the NL80211_CMD_ACTION renaming to NL80211_CMD_FRAME
that happened long time ago. This command can be used with any IEEE
802.11 frame and it should not be implied to be limited to Action
frames.
Jouni Malinen [Fri, 3 Jan 2020 13:18:46 +0000 (15:18 +0200)]
nl80211: Clean up nl80211_send_frame_cmd() callers
Replace a separate cookie_out pointer argument with save_cookie boolean
since drv->send_action_cookie is the only longer term storage place for
the cookies. Merge all nl80211_send_frame_cmd() callers within
wpa_driver_nl80211_send_mlme() to use a single shared call to simplify
the function.
Jouni Malinen [Fri, 3 Jan 2020 12:53:37 +0000 (14:53 +0200)]
nl80211: Get rid of separate wpa_driver_nl80211_send_frame()
Merge this function into wpa_driver_nl80211_send_mlme() that is now the
only caller for the previously shared helper function. This is a step
towards cleaning up the overly complex code path for sending Management
frames.
Jouni Malinen [Fri, 3 Jan 2020 11:49:10 +0000 (13:49 +0200)]
nl80211: Support no_encrypt=1 with send_mlme()
This allows send_mlme() to be used to replace send_frame() for the test
cases where unencrypted Deauthentication/Disassociation frames need to
be sent out even when using PMF for the association. This is currently
supported only when monitor interface is used for AP mode management
frames.
Jouni Malinen [Fri, 3 Jan 2020 11:34:37 +0000 (13:34 +0200)]
Make hostapd_drv_send_mlme() more generic
Merge hostapd_drv_send_mlme_csa() functionality into
hostapd_drv_send_mlme() to get a single driver ops handler function for
hostapd. In addition, add a new no_encrypt parameter in preparation for
functionality that is needed to get rid of the separate send_frame()
driver op.
Jouni Malinen [Fri, 3 Jan 2020 11:22:32 +0000 (13:22 +0200)]
P2P Manager: Use send_mlme() instead of send_frame() for Deauthentication
send_frame() is documented to be used for "testing use only" and as
such, it should not have used here for a normal production
functionality. Replace this with use of send_mlme() which is already
used for sending Deauthentication frames in other cases.
Jouni Malinen [Fri, 3 Jan 2020 10:21:36 +0000 (12:21 +0200)]
IBSS RSN: Use send_mlme() instead of send_frame() for Authentication frames
send_frame() is documented to be used for "testing use only" and as
such, it should not have used here for a normal production
functionality. Replace this with use of send_mlme() which is already
used for sending Authentication frames in number of other cases.
Jouni Malinen [Fri, 3 Jan 2020 10:08:58 +0000 (12:08 +0200)]
nl80211: Simplify hapd_send_eapol() with monitor interface
Call nl80211_send_monitor() directly instead of going through
wpa_driver_nl80211_send_frame() for the case where monitor interface is
used for AP mode management purposes. drv->use_monitor has to be 1 in
this code path, so wpa_driver_nl80211_send_frame() was calling
nl80211_send_monitor() unconditionally for this code path and that extra
function call can be removed here to simplify the implementation.
Jouni Malinen [Thu, 2 Jan 2020 21:16:22 +0000 (23:16 +0200)]
nl80211: Don't accept interrupted dump responses
Netlink dump message may be interrupted if an internal inconsistency is
detected in the kernel code. This can happen, e.g., if a Beacon frame
from the current AP is received while NL80211_CMD_GET_SCAN is used to
fetch scan results. Previously, such cases would end up not reporting an
error and that could result in processing partial data.
Modify this by detecting this special interruption case and converting
it to an error. For the NL80211_CMD_GET_SCAN, try again up to 10 times
to get the full response. For other commands (which are not yet known to
fail in similar manner frequently), report an error to the caller.
Jouni Malinen [Thu, 2 Jan 2020 18:29:24 +0000 (20:29 +0200)]
Test functionality to override driver reported signal levels
"SET driver_signal_override <BSSID> [<si_signal< <si_avg_signal>
<si_avg_beacon_signal> <si_noise> <scan_level>]" command can now be used
to request wpa_supplicant to override driver reported signal levels for
signal_poll and scan results. This can be used to test roaming behavior.
Jouni Malinen [Thu, 2 Jan 2020 18:33:01 +0000 (20:33 +0200)]
Fix signal_poll based roaming skip
Fix a rebasing issue in the signal difference calculation. The older
patch was not updated to use the new cur_level local variable to get the
possibly updated signal level for the current BSS.
Jouni Malinen [Thu, 2 Jan 2020 14:50:50 +0000 (16:50 +0200)]
nl80211: Fix libnl error string fetching
libnl functions return a library specific error value in libnl 2.0 and
newer. errno is not necessarily valid in all error cases and strerror()
for the returned value is not valid either.
Use nl_geterror() to get the correct error string from the returned
error code.
Jouni Malinen [Thu, 2 Jan 2020 14:50:50 +0000 (16:50 +0200)]
mac80211_linux: Fix libnl error string fetching
libnl functions return a library specific error value. errno is not
necessarily valid in all error cases and strerror() for the returned
value is not valid either.
Use nl_geterror() to get the correct error string from the returned
error code.
Jouni Malinen [Thu, 2 Jan 2020 09:51:04 +0000 (11:51 +0200)]
Maintain BSS entries for 5 seconds after interface is disabled
This is targeting the case of MAC address change for an association
which may require the interface to be set down for a short moment.
Previously, this ended up flushing the BSS table that wpa_supplicant
maintained and that resulted in having to scan again if the MAC address
was changed between the previous scan and the connection attempt. This
is unnecessary extra latency, so maintain the BSS entries for 5 seconds
(i.e., the same time that the old scan results are consider valid for a
new connection attempt) after an interface goes down.
Matthew Wang [Tue, 17 Jul 2018 17:56:21 +0000 (10:56 -0700)]
Improve roaming logic
Currently, wpa_supplicant may roam too aggressively; the need_to_roam()
function will return early with a roaming decision if the difference in
signal level or throughput between the current and selected APs is
"sufficiently large." In particular, if the selected AP's estimated
throughput is more than 5k greater than the current AP's estimated
throughput, wpa_supplicant will decide to roam. Otherwise, if the
selected AP's signal level is less than the current AP's signal level,
or the selected AP's estimated throughput is at least 5k less than the
current AP's estimated throughput, wpa_supplicant will skip the roam.
These decisions are based only on one factor and can lead to poor
roaming choices (e.g., a roam should not happen if the selected AP's
estimated throughput meets the threshold but the current signal and
throughput are already good, whereas a roam should happen if the signal
is slightly worse but the estimated throughput is significantly better).
This change standardizes the roaming heuristic for signal strength
difference requirements and will hopefully improve user experience. The
change can be summarized as follows: based on the current signal level,
a certain roaming difficulty is assigned. Based on the selected AP's
estimated throughput relative to the current AP's estimated throughput,
the difficulty is adjusted up or down. If the difference in signal level
meets the threshold, a roam happens.
The hard-coded values were selected purely based on the previous version
of this function. They may eventually need to be fine-tuned for optimal
performance.
Signed-off-by: Matthew Wang <matthewmwang@chromium.org>
Jouni Malinen [Wed, 1 Jan 2020 15:09:18 +0000 (17:09 +0200)]
Allow roam to lower signal level if throughput benefit is significant
Do not prevent roam to a different BSS based only on the signal level
with the current BSS being higher than with the selected BSS. If the
estimated throughput is significantly higher (> 20%), allow roaming if
the following conditions are met.
Jouni Malinen [Wed, 1 Jan 2020 14:56:40 +0000 (16:56 +0200)]
Skip roaming based on signal level difference if current SNR is good
If the current SNR with the associated BSS is sufficiently good (better
than GREAT_SNR = 25), there is limited benefit from moving to another
BSS even if that BSS were to have a higher signal level. As such, skip
roaming based on the signal level difference between the selected BSS
from scan results and the current BSS for such cases.
Jouni Malinen [Wed, 1 Jan 2020 11:38:11 +0000 (13:38 +0200)]
Use signal_poll noise information for roaming, if available
Using average signal strength from the driver and hardcoded noise floor
does not look like an ideal design since there can be significant
differences in the driver-reported noise floor values. Furthermore, even
though the current noise floor is a snapshot from the driver, it is
common for drivers to use a noise floor value from a longer calibration
step and that should not prevent the driver provided value from being
used. This makes the comparisons of the signal strengths between the
current AP (signal_poll) and other APs (scan) more accurate.
As an example, test runs in home environment showed 5 dB difference
between the driver reported noise floor and the hardcoded value and this
could result in significant differences in estimated throughput
calculation.
Jouni Malinen [Wed, 1 Jan 2020 09:26:52 +0000 (11:26 +0200)]
RSN: Do not add PMKSA candidates unnecessarily
Add PMKSA candidates from scan results only if they advertise an AKMP
that is used with RSN pre-authentication. Previously, candidates were
added but then ignored later if the AKMP was not suitable.
Jouni Malinen [Wed, 1 Jan 2020 10:03:42 +0000 (12:03 +0200)]
Clear last Michael MIC error timer on FLUSH
TKIP countermeasures were already terminated on FLUSH, but the timer for
detecting two Michael MIC errors within 60 seconds was left behind. This
resulted in test case failures with following test sequence:
ap_cipher_tkip_countermeasures_sta ap_cipher_tkip_countermeasures_sta2
Matthew Wang [Fri, 6 Dec 2019 22:27:48 +0000 (14:27 -0800)]
wpa_supplicant: Fall back to avg_signal in roaming decision
Some drivers (e.g. Marvell WiFi) don't report avg_beacon_signal, but
it's still useful to poll for the signal again when a roaming decision
needs to be made. Use si.avg_signal when si.avg_beacon_signal is not
available.
Signed-off-by: Matthew Wang <matthewmwang@chromium.org>
Update throughput estimate for the current BSS based on signal poll
We saw that on certain platforms in certain places we keep switching
between two APs and eventually get the same RSSI. Debugging showed that
we have a very big difference between the two antennas.
Ant A can hear AP A very well (-60) but AP B very bad (-80)
Ant B can hear AP B very well (-60) but AP A very bad (-80)
When the device associates to AP A, it'll learn to use Ant A. If the
device uses one single antenna to receive the scan results, it may hear
the AP it is currently associated to on the second antenna and get bad
results. Because of that, the wpa_supplicant will roam to the other AP
and the same scenario will repeat itself:
Association to AP A (Ant A reports -60).
Scan on Ant A: AP A: -60, AP B: -80
Scan on Ant B: AP A: -80, AP A: -60 ==> ROAM.
Association to AP B (Ant B reports -60)
Scan on Ant A: AP A: -60, AP B: -80 ==> ROAM
Etc...
Improve this by querying the signal level of the current AP using
drv_signal_poll() instead of relying on the signal level that we get
from the scan results. Also update the throughput estimate based on the
likely more accurate values for the current association.
projects / thirdparty / hostap.git / log
