For ML-AP, if fastreassoc command is received and scan
candidate is not found in host scan db.
Host will still send roam invoke request to FW, FW will
perform scan.
Change-Id: I8157cf396e3e0c19607d990307e07d31e879a2d4
CRs-Fixed: 3474257
When handling T2LM action request frame, clear the previous
ml peer level t2lm negotiation before applying new tid mapping
to FW.
Change-Id: Ie7f7a28fb7cf886778cd697605352c4ff47d18fd
CRs-Fixed: 3487089
Diag logging is used for connectivity logging of
BTM event.
Remove unused legacy connectivity logging mechanism for
BTM events.
Change-Id: I65dc5fe5f071fd74eeed2433894a238318ce76ab
CRs-Fixed: 3470401
Diag logging is used for connectivity logging of roam
candidate event and roam roam result event.
Remove unused legacy connectivity logging mechanism for
roam candidate and roam result events.
Change-Id: I826107fc764abdeec7408dbb6c52e580d1fe32f4
CRs-Fixed: 3470206
Scenario: There is dual sta present in HBS mode in DUT where one
sta is on wlan0 interface and it has 11be capability, so it has
formed ML connection with ML AP. The another STA is present on
wlan1 interface and this STA also supports 11be capability. But
as per current design, host supports 11be with MLO. It doesn't
support 11be alone. Also 11be with MLO is supported only in wlan0
interface. So for another connection which are present in wlan1
interface, even if it supports 11be host will downgrade to 11ax
and form connection.
During the formation of second connection, host will sends peer
create command(WMI_PEER_CREATE_CMDID) to firmware. Before sending
this command, host checks only eht cap in cm_create_bss_peer()
and fills the mld_mac addr. Because of that, DP assumes that
there is an ML connection and it creates MLD peer on wlan1
interface. Due to this, it causes ping failure on wlan1 interface
as the ICMP response might be coming on incorrect peer and it's
getting dropped.
As part of fix, check whether vdev is mlo supported or
not in cm_set_peer_mld_info()
Change-Id: Ieed67aa2735d200a140f9e771d791b3b9308a0f9
CRs-Fixed: 3474874
In case of wide band scan request, FW sends subband CCA busy info
to host via a new tlv wmi_cca_busy_subband_info added in existing
event WMI_CHAN_INFO_EVENTID.
Fix is to store CCA busy info coming from FW to HDD context.
Change-Id: Ic4b439ed1c270badbe265dc8543dc6d6a1612fab
CRs-Fixed: 3469467
Add support to stop t2lm timer during sta disconnection
and clear t2lm mapping.
Change-Id: I59e724331eda8b67fd1405031bdb00d12805ad09
CRs-Fixed: 3479496
When roaming happens, vdev CM changes from CONNECTED to ROAMING first,
vdev still keeps up, then RSO becomes ROAMING_IN_PROG later,
so need permit RSO cmd during CM ROAMING state.
Change-Id: If4902dcff6b8b4f169fb43da4a889b09e1c2875d
CRs-Fixed: 3482158
While mlo wlan0 sta connecting with vdev0 up and vdev1 not up yet, wlan1
vdev2 disconnected, enabled roaming on other connected sta, then vdev0 RSO
enabled before vdev1 up, mlme session for vdev1 wasn't created in F/W. When
roaming happened, vdev1 peer wasn't deleted, F/W asserted.
To fix it, for mlo sta, follow legacy sta, if not all of requested links
are up, abort RSO operation except deinit on it.
Change-Id: I7026614ab91cf0a603c97e962433dc5139b210dd
CRs-Fixed: 3471655
If PMKID present in AP expires, AP rejects the PMKID based association.
Current design is to clear the cache using BSSID/MAC address of link
in such cases to let a fresh SAE authentication to happen.
But host driver stores PMKSA with MLD address for ML BSS.
Adapt to the same and update PMKSA cache clear APIs to use
MLD address instead of link address.
Use the API which gives ML BSS peer's MLD address for
MLO VDEV or else legacy address for non MLO VDEV to remove
the correct entry in PMKSA cache based on association.
Change-Id: I4af9814e377737eef4a54fdf15db9f6bac38d954
CRs-Fixed: 3433723
Add the following fixes to get Async TID to link map info:
1. Add APIs to register and unregister the T2LM callback.
2. Add HDD callbacks.
3. Send T2LM info, when receiving TID-to-link mapping response
from the associated AP MLD.
Change-Id: Ia98605b3244b24cc3cac41d2decf22ff6aa82f88
CRs-Fixed: 3431518
Need to add support for rTWT feature in host. Host needs to define rTWT
ini, and set rTWT bit in EHT Cap. rTWT ini will allow users to
enable/disable rTWT feature. We define rTWT ini and function to get ini
value. EHT Cap will advertise rTWT support to peer. To populate EHT Cap
we must validate rTWT is supported in driver. To do this, we check rTWT
and bTWT params.
Changes are as follows.
1) Define rTWT ini
2) Get rTWT ini value
3) Get rTWT support
4) Set rTWT bit in EHT Cap
Change-Id: Id5676c6b7c3d4f48a0010f2e9ae369d9dfa710e9
CRs-Fixed: 3446652
To process get_cu_for_each_subbw request from upper layer,
host needs to trigger scan for initial connection ch width
on every get_cu request.
Change-Id: I2bbf20c9207f46983b14dcd95bf81f32cfd7c060
CRs-Fixed: 3466958
Add the following fixes in case of peer-level tid-to-link
mapping negotiation.
1) Copy peer-level mapping to wlan_prev_t2lm_negotiated_info
from wlan_t2lm_onging_negotiation_info after negotiation
is completed as the former structure is used for mapping
reference.
2) Currently, the Host sends peer-level negotiated mapping
to FW, followed by an action frame response to the AP.
Reverse this order to align with the design flow.
Change-Id: I0e0db95c539cfd9aee27b481683e3e807a1f712e
CRs-Fixed: 3423125
Firmware sends beacon/probe response of the roam candidate to host
through roam_frame_event when it chooses a candidate which needs
SAE authentication to be performed. This is needed as the host
checks if corresponding scan entry is present in the scan db
before starting SAE authentication trigger.
But in MLO roaming case, firmware might send ML probe response
and the scan entry corresponds to re-assoc link might be present
in the per STA profile of the ML probe response.
So, extract all per-STA profiles and add them to scan db.
Also, use this API to extract and add the link specific
ML probe responses for the ML probe response received
in roam sync indication if extraction via roam info fails due
to some reason(currently it may fail if fw roams to a link but
got ML probe rsp from other link as link_info is not
present roam info).
Change-Id: I6c67927732a54568dfe618bdecaca8f37515e203
CRs-Fixed: 3454834
As part of commit(Id680d0f6193740b24e78115ac7dc7be26e28acff)
during roam driver notify the TDLS STA about disconnection.
During this driver sends the TDLS disable command to Firmware
without deleting the TDLS peers. As a result of this in FW the
TDLS delete instance is called and TDLS handle is marked as NULL.
And later in driver the TDLS peers are getting deleted and now
the teardown update command is sent to Firmware. But in FW as
the TDLS handle is already NULL, it has asserted and crashed.
Fix is to delete the TDLS peers in driver and update the teardown
to Firmware and then send the TDLS disable command to Firmware.
Change-Id: Ic425cc065a29bbc444531f523eaccb14014ae47a
CRs-Fixed: 3455117
Currently, MLO vdev flag is not cleared even when STA roams to a
legacy AP from an ML AP. This causes sending invalid info to
kernel as the os_if APIs still treat the connection as MLO
connection though STA roams to an ML AP.
Clear the assoc vdev MLO flag to avoid this.
Change-Id: Iaf013aebd1fe794202abdeeaa6b5db14987b6103
CRs-Fixed: 3466501
When a T2LM teardown action frame is received clear the
previously negotiated peer level t2lm mapping and update
FW with established mapping if valid or default mapping
applies.
Change-Id: I7951c76fe3d6c1bf231300675f2b954ff7697b39
CRs-Fixed: 3420421
Currently, MLO flags are cleared before issuing disconnect on
the link vdev when DUT roams to a single link ML-AP. But these
flags get cleared as part of disconnect handler and not supposed
to be cleared anywhere else.
Clearing these flags before issuing disconnect on link causes
sending disconnect done indication to kernel as the flags
got cleared, which make the vdev as non-MLO. This race condition
is possible when a disconnect is queued from userspace on
link vdev while one disconnect gets processed on it(due to
single link roaming).
Change-Id: I8206c1b8197bf26ad852eafb631c8585bc749846
CRs-Fixed: 3439175
Host received deauth and roam sync back-to-back after WoW, sent vdev
RSO_DISABLE before RSO_STATUS event, F/W assert.
1. Host received deauth event, limSmeState went out of link est state, sent
async event of cm_send_sb_disconnect_req.
2. Host received and handled roam_sync, CM_RSO: vdev0:
[ROAM_RSO_ENABLED(2)] --> [ROAM_SYNCH_IN_PROG(5)], roam sync failed since
not in link est state.
3. Host aborted roaming, ent back to CM-VDEV-0: CONNECTED, removed roaming
cmd from serialization queue.
4. Host sent RSO stop to F/W.
5. Host handled cm_send_sb_disconnect_req, dropped RSO stop during roam
sync, called cm_disconnect_continue_after_rso_stop without waiting for
RSO_STATUS event.
6. Host sent vdev RSO_DISABLE and vdev stop before received RSO_STATUS
event, F/W asserted.
To fix it, if host roam sync failed, need wait for HO_FAIL event and handle
it before handling other cmd, can't remove roaming cmd from assoc vdev
serialization queue to unblock other pending cmd before handling
HO_FAIL, just send assoc link RSO stop to F/W one time with
REASON_ROAM_SYNCH_FAILED, then wait for 2 events: FW_ROAM_EVT RSO status
and HO_FAIL. then handle HO_FAIL, at last remove roaming cmd, unblock other
cmd and trigger disconnect.
Change-Id: I03dd3ed21007ee67613c2d90b5567348c240f375
CRs-Fixed: 3445233
Reserve log like "[18:11:32.086000] [ROAM_TRIGGER]: VDEV[0]"
for common roam and special roam.
Some customers are used to get log from serial port, reserve
this log can better support debug.
Change-Id: I79565c87cfcf42ae5de3d6895c27666d2676370b
CRs-Fixed: 3451003
Rename wlan_cm_set_offload_ssid() to
wlan_cm_update_offload_ssid_from_candidate().
Also, add separate APIs for get and set roam_offload_bssid.
Change-Id: I64d6661a265aeafde7c09fa59bc94cf7a49baf0b
CRs-Fixed: 3445337
Currently, OWE assoc link scan entry is cached post roam success
but link scan entry is not cached. This causes join failure while
non-assoc link tries to get connected post assoc link roam.
So, cache scan entries of all links while processing first link
itself to avoid these failures.
This also fixes adding scan entry properly for non-MLO OWE
cases as well.
Change-Id: I3188f98f2682acceaae28f226e6c8abe0c81fff9
CRs-Fixed: 3437632
Currently, when ML probe response is received in non-assoc link
channel, right link_id is not passed to the link_probe rsp
generation API as it loops through all available per STA profiles.
It fails when probe rsp received link id is given as corresponding
per STA profile is not present in the probe rsp.
Skip calling the ML probe rsp generation API for
the link id on which it's received.
Change-Id: Ib636c6d5909b7f2760c7e57ec4f14f8ed060a910
CRs-Fixed: 3428521
Since we fill roam information into a ring buffer, when ready to
write a buffer, we need clear the dirty data first.
Change-Id: I94dac785803f15451df816c303f8a594ee688ff7
CRs-Fixed: 3437152
In roam invoke scenarios, if FW fails to roam
it will send roam invoke failure.
Until roam invoke failure is received FW is trying
to roam.
Currently if FW aborts roam, host will trigger
disconnect and rso stop is sent due to this FW
is unable to retry roaming.
Fix is to not trigger disconnect at roam abort
and wait until FW sends roam invoke failure.
To handle this store roam invoke params in rso cfg.
Change-Id: I32eeda70ef9f4d8a049fec53053f9e0c6eb36e51
CRs-Fixed: 3433320
SAE authentication happens with MLD address rather than link
addresses as it's common authentication for all links in an ML
connection. But OTA exchange of authentication frames happens
with a chosen link address from available links.
As userspace takes care of the SAE authentication, host driver
has to replace the rx SAE auth frame SA with peer MLD and DA with
self MLD address. Similarly for tx SAE auth frames, replace self
SA with link address and DA with peer link address.
Current change is to support the same for roaming case. Fetch the
MLD from scan cache corresponds to link address provided in roam
auth event and replace as mentioned above.
Change-Id: Ide15caa122f78301391b5ea891e2e6eaf4af8974
CRs-Fixed: 3441483
Obsolete TWT functions are still called in driver. As part of
removal, Remove these functions and use component function.
Change-Id: I9f8b4186753ec9d446399be9e1cfc03a63a17391
CRs-Fixed: 3294868
When roam from mlo to legacy, self mac need change from link addr to
interface addr.
When roam from legacy to mlo, self mac need change from interface addr
to link addr.
Self mac in vdev is updated by mlo_roam_update_vdev_macaddr already,
self mac in pe session need update too.
Remove useless self_mac in roam_offload_synch_ind.
Change-Id: I7bdfe35d5a219cb2c0eba30a4995babc8ac6199a
CRs-Fixed: 3442785
Since special vendor roam log is not need for common roam,
add flag to filter the log.
Change-Id: I5eb47f68048a48f817037e5e98d3e7fe51ee6565
CRs-Fixed: 3437627
Disconnection is logged even if keepalive indication is
sent to host after firmware sending final beacon miss
event.
Add two new roam failure reason code to indicate roam
failure and send final beacon miss.
ROAM_FAIL_REASON_NO_AP_FOUND_AND_FINAL_BMISS_SENT: Indicate
roam failure when no ap is found and final beacon miss
is sent.
ROAM_FAIL_REASON_NO_CAND_AP_FOUND_AND_FINAL_BMISS_SENT:
Indicate roam failure when no candidate ap is found and
final beacon miss is sent.
Change-Id: I6d7a6034f8be5388dd35dd3c05fecfd40a8d888d
CRs-Fixed: 3437060
In SAE roam offload case, the SSID of the candidate is
retrieved from the roam candidate frame, and this SSID
is sent to the userspace for initiating SAE authentication.
However, if roam candidate frame is an MBSS beacon and the
candidate BSSID is present in non-tx profile, still the
SSID from tx profile is sent to the userspace, which leads
to authentication failure.
To fix this:
a) Save the BSSID in addition to the SSID during the roam
candidate frame extraction.
b) During roam auth offload event, get the SSID from the
scan cache only if the existing BSSID cache(filled in #a)
doesn't match with the BSSID received in roam offload event.
This would prevent the time expensive scan table lookup
operation for non-MBSSID cases.
Change-Id: Ib8cc0408cba7ccfc8c2b3e2d657f7b40f8a4d421
CRs-Fixed: 3437698
When fast reassoc command is received with bssid
and channel frequency set to zero, if no candidate
found, STA DUT should disconnect.
Currently STA is disconnecting from current AP but
the also when FW aborts roam.
To fix this add additional check for roaming source and bssid.
Change-Id: I6ee3f5cea6ef3976b9cc5549ab113e4af30484d7
CRs-Fixed: 3441770
In dual STA case, driver will avoid selecting candidates causing
MCC, but the current hw mode may have some delay updated from firmware.
So the API policy_mgr_are_2_freq_on_same_mac may not get correct
DBS or SBS channel state of two channel frequency.
Fix it by do not depending on current hw mode by using API
policy_mgr_2_freq_always_on_same_mac to check potential MCC or
not of two channel frequency.
Change-Id: I7d5382e786c421d5f99578e08b62b07864f48974
CRs-Fixed: 3441024
Subnet change status field is obtained from bits 4 & 5 of the
roam_reason field in roam synch indication. But in
cm_fill_roam_info() the subnet_change_status is received from
the masked roam reason field and is always 0. So after
roaming to AP with different subnet wrong subnet change
indication is sent to userspace and ultimately NUD failure is
triggered.
Get the subnet_change_status directly from the roam synch
indication
Change-Id: I708276750c0aa3f1586cf9501f55f742d42ebf3f
CRs-Fixed: 3438725
When fast reassoc command is received with bssid
and channel frequency set to zero, if no candidate
found, STA DUT should disconnect.
Currently STA DUT stays on the same AP.
Fix is to add bssid zero check and disconnect from
current AP.
Change-Id: Id6a26703929f7783e852bb42eebd57f5d5ba8b13
CRs-Fixed: 3433308
Function wlan_cm_is_sta_connected() was removed by the following patch:
qcacld-3.0: Code cleanup for connection manager CSR
Change-Id I156b7ee286f666ee5d660312bf5ba2fe364165c4
However, the prototype was not removed, so remove it now.
Change-Id: Id23f14f1390c72e5f39cef3fbab2eacd6752dc1d
CRs-Fixed: 3430410
In case supplicant enable's roaming before link vdevs are up
drop the request as once all links are up host will enable RSO.
Change-Id: Iba10b6da7c0d24f6e7b1601eefab1dbee0e3af46
CRs-Fixed: 3429178
For mlo roaming, roam sync will be handled on
link vdev/s first and then assoc link vdev.
In case of mlo self roaming, an issue is observed
during peer creation as one already exists on other link.
To solve this all link peers will be deleted
while handling roam synch on first link.
Change-Id: I4d3ff71dd7c8b54f3f06b0147ce23ac8bfe5f49b
CRs-Fixed: 3421698
Notif_param of Roam event ID carries the trigger reason as part
of roam scan start notif. Currently host doesn't read this param.
This commit reads the trigger reason and sends it to user-space
via vendor attribute whenever the roam event vendor command is
enabled.
Change-Id: If9fdbc31df149f4f1c9ca196ee74e2fcb3709a1c
CRs-Fixed: 3161803
As part of commit I6d04f0946e6735613a4cf3d4361bf31052314fa0
a vendor command was added to enable roam stats event.
This change registers callback to send the added vendor
command event to userspace.
Change-Id: I3c504797b1b9619d3e4ecc1c37256dea75127e14
CRs-Fixed: 3403536
Host driver may abort the roam sync due to multiple reason.
The expectation is to revert the state of vdevs/links to original
state i.e. before receiving the roam sync indication.
Currently, there is a cleanup of vdev MLO bitmap assuming that
roam abort should reset whatever is set as part of roam sync
handling. But these flags might get set in previous connection
and are not supposed to reset upon roam sync abort.
Driver does the disconnection as part of HO_FAIL event handling
and corresponding vdev MLO flags gets cleared as part of that.
So, don't cleanup any vdev flags as part of roam sync abort.
Cleanup wlan_connect_req_links which is set only when roam sync
starts.
Change-Id: I5ca0d789a49fcf954dac332f494650866913f94b
CRs-Fixed: 3418430
Support mld self roam between DBS(2 GHz+5/6 GHz) and HBS(5 GHz+6 GHz).
Change high rssi config.
Change-Id: Icb65f26509848a09fe46bf01154a0f42efa073f2
CRs-Fixed: 3416549
3-link ML connection/roaming expects link id to generate link
specific (re)assoc response from received (re)assoc response.
This info is fetched from pe_session but it's not populated
properly in case of roaming. Roam sync indication carries all
links info whereas the assoc rsp generation API expects the
link info for which it has to be generated.
Copy partner link info only from roam sync indication to
pe_session to cater the same.
Change-Id: I75c9a42688f7613ca3f1afab8afe77622bb659aa
CRs-Fixed: 3412939
