Add support to enable the "FT over DS" via
QCA_WLAN_VENDOR_ATTR_CONFIG_FT_OVER_DS vendor config
command.
This configuration helps the firmware to support
“FT over DS” instead of “FT over Air”
Change-Id: I493bf15f83bc39135aaeefa9d8ef328f0fc0cda3
CRs-Fixed: 2968680
During FT-EAP roaming after session timeout if the FTIE received
in the EAPOL of full EAP is improper, the supplicant would
disconnect the AP with the reason FTIE mismatch.
During this disconnection supplicant flushes all the existing PMK
cache entries and eventually the framework would add these cache
entries again (apparently, the current entry would be with the
older PMK). On this addition to the driver, there is no connected
profile and thus no MDID is associated to this PMKSA entry(AP1-PMK1).
Further, on the fresh (Full EAP) connection to the AP2 with the
same MDID, a new cache entry gets added and thus MDID from the
connected profile gets associated with the PMKSA cache entry. At
this point of time, there are two PMKSA cache entries (AP1-PMK1
without MDID and AP2-PMK2 with MDID).
Adding further, if the incorrect PMK (without MDID / stale entry)
i.e. PMK1 is given to the firmware on the RSO command for the
specific BSSID (here AP1) during roam, further roams shall be
calculated with this PMK1 and result in to the invalid pmkid
issue.
Thus, avoid this situation by removing that stale PMKSA entry in
the driver and this entry can be deleted only when it is associated
with the MDID. Driver on the new pmksa addition shall check for the
matching pmksa cache entry with the MDID and delete the earlier ones.
To associate the MDID, this commit checks for the same from the scan
entry of the BSSID, if there is no connected profile for the same
(this happened when the framework has added the cache entry when in
the disconnected state).
Change-Id: If81d66059173dda8d7a2940ca10dc96ef16e125c
CRs-Fixed: 2972115
Fils info is stored per vdev. In RSO scan mode command, in FILS
TLV, the next ERP sequence number to be used in the next FILS
connection is sent to firmware and is filled from the vdev fils
info. But after roam sych event is received, the erp sequence
number is not updated in the vdev fils info, causing FILS roaming
failure.
Update the fils erp sequence number to vdev fils info after
roaming.
Change-Id: Ibc624d9c4b0fc535c430881d3033169b002c8419
CRs-Fixed: 2988113
When HO failure happens or roam synch failure happens due to
north bound disconnect being already processed, host
transitions to RSO STOP-> Roam Deinit as part of the disconnect.
But after this roam abort is done, which tries to reset the
state to ROAM enabled. This causes roam init to be sent
immediately and in STA + STA scenario where STA2 roaming
is also enabled, it causes fw to go to invalid state
as STA 1 vdev is deleted.
Allow roam abort to transition to RSO Enabled only in
Roam Synch in progress state or Roam in progress state
and not in RSO deinit state.
Change-Id: I1f07d4a3068f719b8d05e111bc8d45626c6a2468
CRs-Fixed: 2969507
For STA+STA case, if disconnect is started on one vdev,
roaming is getting disabled on other vdev to fix this
skip updating RSO state.
Change-Id: If527555c3ea3d50516be3a743bb03914b3a2359f
CRs-Fixed: 2981648
Add partner links information in join request to proceed for ML
connection in best links selected in ML scoring.
Change-Id: I463264d9c78f3ab061d54a3245b15174d0940910
CRs-Fixed: 2961881
With connection manager enabled, the send IE request is
not sent after connection.
Add support to send set IE request after connection is
successful in connection manger.
Change-Id: I2f30eb67ba295e0ca416ccb40e12a376e0d7a47e
CRs-Fixed: 2976734
In case of WTC_BTM roam trigger reason, if FW reports rssi
threshold for 5G and 6G AP via wmi_roam_trigger_reason,
Host should print it in kmsg.
If wtc_candi_rssi_ext_present is set (1), It means fw
reports 5g and 6g candidate rssi (new firmware case).
If wtc_candi_rssi_ext_present is clear (0), It means fw
doesn’t report 5g and 6g candidate rssi (old firmware case).
Host use flag "wtc_candi_rssi_ext_present" to keep backward
compatibility.
Change-Id: Idbaac2e9d470525641c152b024b8786d262e600c
CRs-Fixed: 2965025
Currently, if it's 11r connection, zero out reassoc_ft_ie in
cm_preauth_success(), it will cause FT roaming failed due to
ft ie loss with reassoc request frame.
Fix is to only zero out reassoc_ft_ie when connection is 11r
and open mode.
Change-Id: I8c1d6ca743b528c4ab0211723a792b14abe5acab
CRs-Fixed: 2976385
Fix LFR2 roaming related memory leak issues in function
lim_send_reassoc_req_with_ft_ies_mgmt_frame() and
wlan_cm_host_roam_start().
Change-Id: I83271b57d0651d680324b67f0e6e6a3694c9ab8c
CRs-Fixed: 2976469
With commit I84c77644c6023322ac77e307b06be70ec0f58ed5, Host
allows dual sta roaming only if two STA's sessions for are
active (STA + STA) and concurrency is in DBS.
Currently, before the increment of the number of active
sessions in the policy manager, the host tries to enable
dual-band roaming. This results in dual sta roaming failure
with the reason number of active sessions equal to one.
The fix is to make sure the host should try to enable dual-band
roaming after incrementing the number of active sessions in the
policy manager.
Change-Id: I150926897cb773db979279edc54e5fb731d98e43
CRs-Fixed: 2971831
EAPOL handshake is done by wpa_supplicant in initial connection
and WPA3 roaming as well. Driver is supposed to wait for EAPOL
handshake/key complete timeout after getting roam sync indication
from firmware. But this wait is skipped with the change
change-id : Ie4f06cfcb066ae245de024b62da586aade783aec.
Add the same check again.
Change-Id: I437c6be15d7c1a3775a177c2da6c581d417c2533
CRs-Fixed: 2963548
Refactor csr_roam_pmkid_req_callback to use
mlme_cm_osif_pmksa_candidate_notify and move other
unused API under CM flag.
Change-Id: I835d54804f733eddff48de38e4c545a87f34bb30
CRs-Fixed: 2965602
For dual sta case, if vdev0 disconnected before vdev2 disconnected, vdev0
PCL cmd retained vdev level, after vdev0 reconnected first, just allowed
intra band in PCL, only can roam in intra band.
Fix: When vdev disconnect, reset it's PCL cmd to pdev level.
Change-Id: I3ed3ec398cfe1a17d07ce716c27026b9660305bd
CRs-Fixed: 2960194
The WTC parameters are stored in MLME vdev private object.
So the parameters are cleared once vdev delete is done.
Move the WTC parameters to psoc->mlme_cfg to retain the
values across disconnections.
Change-Id: Icecfbef19af994f8a05622501d435bfb63beca50
CRs-Fixed: 2959715
Existing INI disable_4way_hs_offload is defined as a boolean
that controls the EAPOL Handshake to be offloaded to Firmware or
Not. There is no way to specify the EAPOL offload per AKM.
Hence extended the support of EAPOL handshake offload to firmware
per AKM by converting this INI to a bitmap integer.
With ini value as 0 by default Firmware takes care of 4-way HS
during roam for all AKMs except for SAE and OWE. If
this ini value is 1 then Host takes care of EAPOL for all AKMs.
Having the backward compatibility, added a new value to ini.
If this ini value is set to 2 and also if the SAE EAPOL offload
service bit is set by Firmware then the behavior is same as ini
with 0 except that Firmware takes care of 4-way HS for SAE roam
as well. And for OWE it is still handled by supplicant.
To achieve this, new roam offload flag is introduced to indicate
Firmware about SAE roam EAPOL offload, So that in Roam sync
indication Firmware will inform the status as connected but not
authenticated so that supplicant can take care of 4-way EAPOL HS.
Change-Id: I8b65a1768fad91fac7e427712d3ccf4586176f5f
CRs-Fixed: 2955029
Fix some LFR2.0 issues for connection manager path as below:
1: Add ptr null pointer and len check before access
rsp->connect_ies.bcn_probe_rsp.
2: Remove aged AP threshold check so that roam process can find
candidate AP with more probability.
3: Set cm_ctx->vdev to cmd_info.vdev in
cm_remove_preauth_cmd_from_serialization().
4: Add more debug logs for preauth part.
Change-Id: I4b0766bba40e3ebc3d01bdd4ae59aec1ec3a21d9
CRs-Fixed: 2961151
Remove duplicate code in connection manager and call
same API from connect complete and roam complete.
Change-Id: If7e9e1069acb1dede9433d16ef96f37029ff716d
CRs-Fixed: 2960921
Send TDLS STA notify only when connect is success.
This is required as incase connection, it will increment
tdls_soc_obj->set_state_info.set_state_cnt.
It that connection results in failure, then also, without
this check this count will increment.
For subsquent connections, in tdls_send_update_to_fw()
because of incremented count, it will be treated as
FW Set State is already sent and will not send it
which will result in FW crash while adding TDLS peer.
Change-Id: Ic3bedfb1357ac703f98d2d0285dcc85605bab97e
CRs-Fixed: 2955401
Currently in WTC mode the candidate AP RSSI threshold
configuration is applied for candidates on all band.
Add support to configure the candidate AP RSSI threshold for
different bands.
Change-Id: I94442c0f1bd3ea20dc2f5ebe893a9ead7fb8d890
CRs-Fixed: 2951610
Only do cleanup roc request and notify tdls sta disconnect
when sta disconnection operation is from osif or north bound.
Change-Id: I1d0aa4f1cd8a23baca890228485f22124eeabf78
CRs-Fixed: 2954772
Roam offload init command should be sent to firmware only
for STA mode. For P2P vdev, if roam init command is sent,
it causes abnormal target behavior.
Don't send roam init command if the vdev opmode is not
Station.
Change-Id: I0bfc3c6b4bc2eef7459b092101bc4d72c68761e5
CRs-Fixed: 2924563
If HO fail/ roam abort is received without any roam start indication
the roam command will not be present in connection manager, so
continue with cleanup even when roam command is not present in these
cases.
Change-Id: I0ff3307b4d4ba96cc3b0c413fd0a41be47e9c588
CRs-Fixed: 2948064
When updating the scan filter with valid channels, refer to the MLME
channel list instead of the policy manager list. The policy manager list
is used mainly for SAP related APIs.
Change-Id: I5530e5bcfd6b5c04063d7530fbc996ffca91df3f
CRs-fixed: 2946365
In roam sync handling, pmksa memory is freed right after
set pmka entry is updated in crypto priv params instead
allocate separate buffer for crypto.
Change-Id: I337efcb29d2233b016f93cae7f4f23d3203786a6
CRs-Fixed: 2931162
When the AP channel is changed, fw reports a suitable AP.
But it's skipped by host as it has same bssid as the one
currently associated. The same process repeats continually,
which result in fail to reconnect with the AP. To avoid this,
send BMISS indication to upper layer and restart roam.
Change-Id: Idfcb69f81ce72a4ab7ef0c105f3af40cb92c53d5
CRs-Fixed: 2913663
Consider two APs: AP1, AP2
Both APs configured with EAP 802.1x security mode
and OKC is enabled in both APs by default. Initially
DUT successfully associated with AP1, and generated
PMK1 by performing full EAP and added an entry for AP1
in pmk table. At this stage, pmk table has only one
entry for PMK1 (1. AP1-->PMK1).
Now DUT roams to AP2 using PMK1 (as OKC is enabled)
but session timeout happens on AP2 just before 4 way
handshake completion in FW. At this point of time DUT not
in authenticated state. Due to this DUT performs full EAP
with AP2 and generates PMK2. As there is no previous
entry of AP2 (AP2-->PMK1) in pmk table. When host gets
pmk delete command for BSSID of AP2, the BSSID match
fails. Hence host will not delete pmk entry of AP1 as well.
At this point of time, the PMK table has two entry
1. AP1-->PMK1 and 2. AP2 --> PMK2.
Ideally, if OKC is enabled then whenever timeout occurs in
a mobility domain, then the driver should clear all APs
cache entries related to that domain but as the BSSID
doesn't exist yet in the driver cache there is no way of
clearing the cache entries, without disturbing the legacy
roaming. Now security profile for both APs changed to FT-RSN.
DUT first disassociate with AP2 and successfully associated
with AP2 and perform full EAP and generates PMK3. DUT first
deletes PMK entry for AP2 and then adds a new entry for AP2.
At this point of time pmk table has two entry AP2--> PMK3
and AP1-->PMK1. Now DUT roamed to AP1 using
PMK3 but sends stale entry of AP1 (PMK1) to fw via RSO
command. This override PMK for both APs with PMK1
(as FW uses mlme session PMK for both APs in case of FT
roaming) and next time when FW try to roam to AP2 using
PMK1, AP2 rejects PMK1 (As AP2 is expecting PMK3) and
initiates full EAP with AP2, which is wrong.
Fix is to update pmk table entry for roamed AP1 with PMK3 value
comes to host via roam sync indication event.. By this host
override stale entry (if any) with the latest valid pmk for
that AP at a point of time.
Also add new api to get the sae single pmk ini value and
update the bss desc of single_pmk cap after intersecting the ini
value.
Change-Id: I12fecbae69d0d2a11f2a39eee6be960752d831bd
CRs-Fixed: 2889847
Address below corner scenarios for dual sta roaming:
1. Initialize primary interface vdev id
with WLAN_UMAC_VDEV_ID_MAX.
2. Enable roaming while processing set primary interface
vendor command irrespective of dual sta roam policy.
3. Disable mcc_adaptive_scheduler before sending
WMI_RESMGR_ADAPTIVE_OCS_ENABLE_DISABLE_CMDID to FW and
enable it if host receives dual sta policy as unbiased.
Change-Id: I8e5254c6f9053bd5fe7f925af4b577e76c9b9a2e
CRs-Fixed: 2928870
Presently the roaming design is to support roaming on
both the STA in DBS mode (with different bands).
With commit I60f6bdb7df4d9eece4fd14abe1fab2d5644d47f1,
Driver allow STA2 connection to any available AP(s).
So in case if STA2 connectes in MCC mode then host
should not allow roaming on non primary STA iface.
Change-Id: I84c77644c6023322ac77e307b06be70ec0f58ed5
CRs-Fixed: 2928804
If there is more than one STA iface concurrently active
and one of them is marked as a primary iface. The host
received CONCURRENT_DUAL_STA_POLICY vendor command with
policy PREFER_PRIMARY. Then Don’t consider PCL weightage
for an STA connection. Due to this host/fw allows a new
connection either in DBS, SCC or in MCC.
Change-Id: I400165f8dd3ab7b94b2cb808f8b34b34d6d42fee
CRs-Fixed: 2929015
Reset scan reject params on roam abort and also if connect
lead to roam invoke release the wake lock as we do in legacy
case, if required roam invoke can use new wake lock.
Change-Id: Ieb870e2598687a1418e896d761f19a198a03a98f
CRs-Fixed: 2929311
Currently below implementation is missing in the driver. So driver
is failing to associate in 11be. Add this implementation to support
11be association in the driver
- Filling EHT IEs in probe response
- Sending EHT mcs capabilities while creating the peer
- Fix compilation issues when 11BE feature is enabled
- Add the the 320MHZ to wma_chan_phy_mode
Change-Id: Ib535db413d5578840feed7fd18bb00e5e28cfb6f
CRs-Fixed: 2926471
Currently, the driver restricts the connection of the 2nd
STA to channels on the other band of the 1st STA connected
band. As per the new requirement, the host should allow STA2
connection to any available AP(s), if primary iface is
configured.
Change-Id: I60f6bdb7df4d9eece4fd14abe1fab2d5644d47f1
CRs-Fixed: 2915733
Add logic similar to I5d8ec9e9d4f8a44178b113d370799fe2bc0373fb
in connection manager to flush PMID on HB failure.
Change-Id: Id87b5774a18a377719c56d3f34dc31284f0a39a3
CRs-Fixed: 2928282
