When MSDU fails for checksum validation, do not aggregate those
packets and make sure current flow is flushed. Since checksum
failure packet data is not trust worthy it is not advisable to
build aggregated skb on top of checksum failure packets.
Change-Id: I09d8c4aeb656e6b0b5d268a60d72147534f2a2ab
CRs-Fixed: 3805053
Target may use association request frame to roam to AP in offload
roaming. Current API pe_set_rmf_caps only supports
reassociation frame parsing.
Add association request frame RSN IE extraction support in
pe_set_rmf_caps.
Change-Id: Ic8bc427a9d7a37dd3531cdfda243bce398eb7400
CRs-Fixed: 3264727
Currently the user configured MFP state that comes from the
userspace in connect request is not handled or processed.
Instead the RSN caps from assoc IE of connect request is inter-
sected with AP RSN caps and sent to Firmware using RSO command.
This RSN caps is used in FW in selecting a roam candidate, which
was causing the cross AKM (eg:SAE -> PSK) roam fail.
Hence, use the user configured MFP value in sending the RSN caps
to Firmware.
Change-Id: Ibf2d7bfba6cd17a98b9e4b1c8c468046ab2e7e62
CRs-Fixed: 3604149
Add support to update key management with higher security
after BSS create response.
Also, Currently if there are multiple AKM and ucast cipher.
Host overwrites AKM and ucast cipher value with the new one.
Instead of overwrite, add support to do ORing to keep all values.
Change-Id: I7a489c7d90a609c28be69a9fe9deb0b9b9c328af
CRs-Fixed: 3751018
Currently, host supports max 2 AKMs based on
NL80211_MAX_NR_AKM_SUITES.
Add support for max 5 number of AKMs in connect req.
Change-Id: I65d54cc0c09d07ae1ac4b8669a64543d332b75e6
CRs-Fixed: 3782253
Currently, STA doesn't support roam between WPA2 to WPA3
security or vice versa. To support this feature, host sends
list of allowed_authmode. So that Firmware will check and
roam on those authmode.
Fix, add support for allowed_authmode list in ap_profile.
Change-Id: I438a133a434ea12ec34680997ace358fd4910028
CRs-Fixed: 3782230
Add support for security score. On the basis of score,
host will select AP for initial connection and roaming.
Change-Id: I041a1b0c1456d7f01dd07e9b282996c56755655e
CRs-Fixed: 3782215
Host completes scan, however due to memory allocation
failure, driver cannot send the scan completed event to
the userspace. After timeout, supplicant sends vendor
scan abort to the driver. However, driver doesn't send
the failure status(because driver doesn't have any active
scan to abort). Since, the status code is sent as 0,
supplicant keeps waiting for the scan completed status
event from host.
Send the vendor abort failure reason to the userspace, so that
the supplicant can clear out scan-in-progress flag.
Change-Id: I771b7d3d76c7d7aea58ac3d02665bd43da00e222
CRs-Fixed: 3707469
Currently roaming_in_progress is set during ROAM_START event from
firmware and reset after ROAM_SYNC_COMPLETE or ROAM_ABORT or
ROAM_HO_FAIL event. But in suspend mode there can be scenario
that firmware sends the ROAM_HO_FAIL first being a wakable event,
followed by ROAM_START event which is non-wakable. Here the
roaming_in_progress bit gets reset and then set. But due to
HO_FAIL the host triggers the disconnection. And during this
currently the roaming_in_progress is reset during VDEV down in
all the cases except HO_FAIL. Hence reset the roaming_in_progress
during VDEV down in all cases irrespective of the VDEV stop type.
Change-Id: I53428b8a43769ed891bdc3e93b11207fe47a1939
CRs-Fixed: 3745192
Currently in the function hdd_send_roam_scan_channel_freq_list_to_sme,
the num_chan variable is declared as uint8_t and is incremented
for each nested attribute PARAM_SCAN_FREQ_LIST.
If the number of attributes sent by userspace is more than max value
of uint8_t, then an integer overflow occurs.
To avoid this issue, add a sanity check to see if num_chan has reached
SIR_MAX_SUPPORTED_CHANNEL_LIST before incrementing variable.
Change-Id: I601a73a118eb65ebb8575f6ed5ed1f29d915f59e
CRs-Fixed: 3568577
qdf_conc_list_lock from policy manager is destroyed as part of
idle shutdown, later when FW down indication comes it is trying
to acquire as part of ucfg_dp_bus_bw_compute_timer_stop.
So call ucfg_dp_bus_bw_compute_timer_stop function in
hdd_soc_recovery_cleanup after checking driver status.
Change-Id: I27164a4b11467aee6c925225fe33d4fdbe0b4082
CRs-Fixed: 3689606
Currently host doesn't check if power type is supported
for connection channel while calculating best 6 GHz
power type for connection. This results in calculation
of incorrect power for connection values if 6 GHz power
type is not supported for that particular channel.
To address this issue configure power type for connection
only if that power type is supported for connection
channel.
Change-Id: I7e4ccdb962974cc8c415e8d174b2da0bf311f1a5
CRs-Fixed: 3697212
Find the best 6 GHz power type for connection
according to following regulatory policy:
1) SP power type is selected only if AP advertises
SP and client supports SP.
2) LPI power type is selected only if AP advertises
LPI and client supports LPI.
3) VLP power type is selected for the below cases,
a) AP advertises VLP and client supports VLP
b) AP advertises SP but client doesn't support
SP but supports VLP.
c) AP advertises LPI but client doesn't support
LPI but supports VLP.
Change-Id: I582fb582e1e11b731a1c6cda01f4fc366f166143
CRs-Fixed: 3456192
Currently the roam scan high RSSI delta is configured via INI.
Define an attribute to allow user configure high RSSI roam
trigger threshold. STA is expected to trigger roam if the current
connected AP's RSSI gets above this high RSSI threshold. STA's
roam attempt on high RSSI threshold aims to find candidates from
other better Wi-Fi bands.
This attribute value is given priority over the INI.
Use a new service bit WMI_SERVICE_5GHZ_HI_RSSI_ROAM_SUPPORT to
enable high RSSI roam trigger in 5 GHz as well.
Change-Id: Ide48ad2261b603de36bd1b31114b91c3a9d6606f
CRs-Fixed: 3586170
In present scenario, STA disconnects with AP if it receives
invalid channel in CSA IE. In this case STA shouldn't
disconnect with AP as this request may come from a spoof AP.
Ignore this CSA request as it might be from spoof AP and
if it is from genuine AP heart beat failure happens and
results in disconnection. After disconnection DUT may
reconnect to same or other APs.
Change-Id: I840508dd27d8c313a3e8f74c4e1f5aa64eecf6f9
CRs-Fixed: 3390251
Currently as STA, when populating channel width set, only max BW
advertised by target are considered. However issue is that due to
regulatory limitations, the max BW advertised may not be the real
operating BW used by our device. Under such conditions, there'll
be conflicts between AP's view of our capablities and the real
ones host configured to target.
Therefore when populating supported channel width set for VHT and HE
capabilities, take current session's channel width into account so
so that capabilities advertised could correctly reflect current
running configurations.
Change-Id: Ib072c3e4d36d5c3fbd5504d94a936175d1a92db0
CRs-Fixed: 3047338
Higher Bandwidth support 80/80+80/160 is enabled on 2.4 GHz Band.
Modififed the population of peer capability logic to disable
80+80/160 based on the band.
Change-Id: I8319ef3fcd93676bca1115a088a926a772545d24
CRs-Fixed: 3030837
Extract the 6 GHz AP power type from HE IE before
setting the power type in STA. By default, power type
is set to LPI.
CRs-Fixed: 3631944
Change-Id: I84c00bdca3358299c6ef919975328558b2928353
If country code for AP and STA are same and if STA does not
support VLP mode for a particular channel and if AP power
type is misconfigured to VLP mode, do not allow connection
to that AP as per regulatory guidelines.
Change-Id: I5c9759d1e8077e83f7b28985253be871213c8385
CRs-Fixed: 3375815
Initialize vht_cap and vht_op pointers in the
sch_bcn_update_opmode_change API.
Change-Id: If57723978d4c1dcced6b1fac4b50b541cc5bab04
CRs-Fixed: 3213123
After the connection is established with HE160/VHT160 AP,
the STA downgrades the channel width to VHT80/HE80, leading
to 50% TPUT reduction.
The associated beacon processing doesn't consider the new VHT OP
IE definition while parsing the channel width. Fix the channel
width processing in the beacon processing.
Change-Id: Icd7d36a8352be7574474fb5f37fa9b4ca291aa04
CRs-Fixed: 3211833
Currently host driver does not free the memory for HE
operation from cache connection info when it receives change
interface from STA to SAP mode which results in mem leak.
To address above issue, free HE operation element memory on
STA to SAP interface change.
Change-Id: I837640814459a62d819e606fc44adc9bfc60fe5c
CRs-Fixed: 3045848
User disables all the channels when the SAP is in
operation, and policy manager restart sap functionality
cannot get any valid channels and SAP also doesn't stop.
Later, kernel stops the SAP, since it is operating on an
invalid channel.
To fix this, restart the SAP on a valid channel only if
atleast one valid channel is available. Otherwise, stop
the SAP instead of waiting for kernel to stop the SAP.
CRs-Fixed: 3617706
Change-Id: Ia613b8e9585ff19aae9bf97ea5d0fc230ae41c5f
Currently when host receives disable channels command, it
disables the channels and if SAP is on it checks if SAP is
on one of the disabled channels in that case host stops the
SAP.
Based on current requirement host should restart the SAP,
with this change add support to restart the SAP if it is
operating on disabled channel.
Change-Id: I89ce21a8707e2c0af0737116b883be13503044e0
CRs-Fixed: 2885308
Offload TX data packets such as ARP response, EAPOL during roaming are
sent by firmware through HTT msg if packet capture mode is enabled.
Whenever any such packet is received via HTT msg, host inspects the
ether type of the packet and matches with the TX filter set by user
via vendor command. If the ether type matches with the TX filter set
by user, then host forwards that packet to packet capture mode interface
otherwise, drops it.
To inspect the ether type of any packet, host uses generic API which
expects packet to be in SKB format. Currently, whenever any offload
TX data packet is received in HTT msg, host wrongly passes the buffer
received in HTT msg instead of SKB to APIs expecting SKB buffer.
This leads to undefined behavior.
So, to fix above issue, whenever any offload TX data packet received,
first allocate the SKB, copy the payload buf of HTT msg which is TX
packet to SKB data and then pass that SKB to the generic APIs to get
the ether type.
Additionally, this change fixes the minor logging error.
Change-Id: If09d49d8a1dcc04ca81454fc262bb5789a0f56be
CRs-Fixed: 3613594
User can set following combination of configs in vendor command for
packet capture mode:
1. PKT_CAPTURE_MGMT_CONNECT_NO_BEACON: to receive all mgmt frames but
no beacons
2. PKT_CAPTURE_MGMT_CONNECT_NO_BEACON + PKT_CAPTURE_MGMT_CONNECT_BEACON +
connected_beacon_interval : to receive all other mgmt frames and only
connected SSID beacons at particular intervals
3. PKT_CAPTURE_MGMT_CONNECT_NO_BEACON +
PKT_CAPTURE_MGMT_CONNECT_SCAN_BEACON: to receive all other mgmt
frames and beacons only during scan.
But with current condition connected SSID or scan beacons config will
not be sent to FW as host checks for PKT_CAPTURE_MGMT_CONNECT_NO_BEACON
config only. Also, on reception of any beacon, host checks for only
PKT_CAPTURE_MGMT_CONNECT_NO_BEACON and if it is set, host drops the
beacon which is wrong.
So, enhance the conditions to send config to FW as well as
remove the condition on reception of any beacon so that connected SSID or
scan beacons are forwarded to packet capture interface when connected
beacon interval config or PKT_CAPTURE_MGMT_CONNECT_BEACON or
PKT_CAPTURE_MGMT_CONNECT_SCAN_BEACON is set by user in vendor command
along with PKT_CAPTURE_MGMT_CONNECT_NO_BEACON config.
Change-Id: I246b175f1c88ed45214527880ba14cdc17bf8206
CRs-Fixed: 3604708
Currently driver uses iface name to get netdev when it processes
ndp responder request. There is a possibility that framework does
not know the iface name when it sends the error response.
if driver does not receive iface name it fails to get the netdev
and fails to perform the dsc operation, because of which driver
drops the ndp responder request command.
To address this issue, use wdev to get netdev instead of iface name.
Change-Id: Ie230f19dbd895f81a4730160cbc9a2de3cc3179d
CRs-Fixed: 3136089
AP advertises TPE IE Tx power as 8-bit unsigned int. STA
needs to convert it into an 8-bit 2s complement signed
integer in the range –64 dBm to 63 dBm with a 0.5 dB step.
Thus, halve the tx power received from AP's TPE IE.
Change-Id: Ibd3227a2f11f230b164af3c65a65f5e61879e25d
CRs-Fixed: 3313617
Currently, wakelock with timeout is acquired when psoc idle timer is
started to prevent suspend but it is not released when psoc idle timer
is stopped. So, in scenarios where wlan is turned off/on frequently,
psoc idle timer will be started and wakelock with timeout will be
acquired as part of wlan off but after wlan is turned on suspend will
not be allowed until wakelock timeout occurs as wakelock is not released
when psoc idle timer is stopped.
To allow suspend after wlan is turned on, release wakelock which was
acquired as part of wlan off in psoc idle timer start.
Change-Id: I85ad9ce4ae9618a88601295d2194b4033be88c2d
CRs-Fixed: 3578872
