Currently DP RX threads are using the same wait_q for all operations.
The problem with this is that when there is traffic for only one
threads, we end up waking up other threads as well moementarily.
This wastes power and is in-efficient.
Use different wait queues for different threads.
CRs-Fixed: 2495719
Change-Id: I689659b7aa0ab93b7e2f009d2dc7fe741b66ee78
Clear recovery flag after idle restart such that next time wifi can
turn on properly.
Change-Id: I7063a9edd08334a84068f88c0564dd4f2aabf666
CRs-Fixed: 2507899
Update the mac, phy, mcs, nss caps of bands
2.4ghz, and 5ghz for HE capabilities in the
wiphy structure to inform userspace about
the driver capability.
Change-Id: I9fb1233e5396bf3ea411c057af9cbcd5b358271c
CRs-Fixed: 2486266
Currently if the driver does not find any scan results
in the ACS scan, it selects a best channel as the PCl
channel by filtering the PCl list based upon the start
and end channel, which may not be correct everytime
as start and end channel does not specify the channels
to be selected.
Fix is to select a default operating channel from the
acs list itself, if no scan results are present.
Change-Id: I9a76957087b9349da66545e0fcaede2355f732cd
CRs-Fixed: 2504796
When CONFIG_FEATURE_HL_GROUP_CREDIT_FLOW_CONTROL is enabled each group
has its own credit limit.
It may happen that when the High Latency TX Scheduler selects a category
the txq at the head may belong to a group which has credits less than the
"credit_reserve" of that group. In this case the scheduler will return
without downloading any frames although the other group may have credits
and also frames to be downloaded.
The scheduler will be called again if there is a credit update from FW or a
packet arrives from network stack and the next txq will be picked up which
belongs to the group which has sufficient credits.
It is seen that sometimes there is no credit update from FW (since the
host has sufficient credits) and the network stack also does not
transmits packet since it has already queued packets in driver's queue.
In such case the scheduler is not called and throughput drop to zero is
seen although there are enough credits on host.
To avoid such a situation, in case scheduler is unable to download
packets from a txq since its group does not have enough credits, iterate
over to the next txq in the chosen category and download its packets.
Exit from the schduler is case able to download from any txq OR not able
to download from any txq.
Change-Id: I6143d5c3aa40761d1997846896e5e77435252b26
CRs-Fixed: 2485819
Convert channel to freq in vdev start structure to
avoid duplicate channel numbers in 6ghz, and 2.4ghz
operation.
Change-Id: Ie78d93a34d1d558b9c94fc8a54aa95ee07bda55e
CRs-Fixed: 2508040
Currently the host driver blocks roam scan offload command to
the target if the connected akm is OWE or SAE.
Add changes to read the WMI_SERVICE_WPA3_SAE_ROAM_SUPPORT and
WMI_SERVICE_WPA3_OWE_ROAM_SUPPORT service capability advertised
by the target and set it to the mlme fw_akm_bitmap. Read this
bitmap in csr_roam_offload_scan() to allow RSO command.
Also the connected akm is sent to the target via the
WMI_AP_PROFILE command. Translate the CSR akm type for OWE and
SAE akm also.
Change-Id: I68527722ab099e0c8e1681ed64ea416b9a6e9936
CRs-Fixed: 2491021
Currently, the cdp API cdp_clear_peer takes as input the sta_id. As a
part of cleaning up the usage of sta_id, replace it by peer mac address.
Change-Id: I4f8bee21ae42d8c9929da8fe2fcfa534778647ed
CRs-Fixed: 2503139
Cleanup ol_txrx_clear_peer to be peer mac address based
from local peer id based.
Change-Id: I63154508e6a08f973a4c602de58217e6bf23d683
CRs-Fixed: 2503737
For WPA3 SAE LFR-3.0 roaming, authentication is offloaded to the
wpa_supplicant via the host driver due to the target crypto
inavailability limitations. The target does candidate selection
and sends the WMI_ROAM_PREAUTH_START_CMDID to the driver.
Once the SAE authentication is done, the supplicant sends the
external auth NL command with the authentication status, PMKID
derived. On receiving this event, send the PMKID, bssid, auth
status to the target over WMI_ROAM_PREAUTH_STATUS_CMDID command.
The host driver decides if the external authentication event is
received for preauth candidate or for initial connection by
checking the connected state information of the csr session.
When the authentication fails with the first candidate, the
target triggers the preauth start for the next candidate. This
happens till there are no candidates available to roam.
There could be a situation where the firmware sends preauth
event for candidate1 and before the supplicant could send the
external auth command to the host driver, the candidate1 preauth
timer could have expired and firmware sends the preauth start for
second candidate. After SAE is done with second candidate also,
the supplicant sends external auth event for second candidate. So
in this situation the firmware should decide to choose the
latest candidate. The host driver here is a mere passthrough and
doesn't maintain timer for the preauth event for each BSSID to
drop the expired candidate preauth events.
Change-Id: I799a05e55a59677dfc2ddf5b55ebc888ccfea03e
CRs-Fixed: 2507003
In case Device is P2P GO and SSR is triggered due to any reason,
Stop_adapter is called. Inside stop adapter adapter->session.ap.beacon
is freed. When P2P GO negotiation is initiated after SSR,
__wlan_hdd_cfg80211_stop_ap function tries to access
adapter->session.ap.beacon which freed during SSR.
Failure is returned in case adapter->session.ap.beacon is NULL,
Which causes failure of p2p negotiation.
Fix is to return success to upper layers in case SAP is already
down, So that beacon interval can be cleaned up.
Change-Id: Ie7a30475b2a54860843b1430964585d22f02c110
CRs-Fixed: 2503216
The Linux Coding Style frowns upon mixed-case names so rename
identifier tpAddBssParams to be compliant.
The Linux Coding Style enumerates a few special cases where typedefs
are useful, but stresses "NEVER EVER use a typedef unless you can
clearly match one of those rules." The tAddBssParams typedef does
not meet any of those criteria, so replace it (and the "tp" variant)
with a properly named struct.
Change-Id: Icf7fe19cfe6eb6fbac5689e3b5474d38c520b311
CRs-Fixed: 2506539
Firmware offloads WPA3 SAE authentication to host due to the
crypto library availability limitation. Once the roam candidate
selection is done in the firmware, firmware sends the
WMI_ROAM_PREAUTH_START_EVENTID. Once this event is received, the
host driver triggers the SAE.
Add support to offload the SAE authentication from firmware to
wpa_supplicant.
Change-Id: Ie3a9e731194e1edfcd9f33ef0d601767b9485d46
CRs-Fixed: 2490201
Create bss peer in lim_process_mlm_post_join_suspend_link
when sta connect, don't need pose message to wma queue.
Change-Id: I23325c9d41b68f67ea9903008521f026e46395c4
CRs-Fixed: 2505558
In case of high latency data path, addba/delba messages
needs to be handled to support partial reorder.
This reverts commit b8919e14c5
and fix below security isses in ol_tx_addba_handler.
1. handle memory allocation failure scenario.
2. Free array memory before assign new memory to avoid
memory leak.
Change-Id: I4f577c9e8bcb40f70ffb1b305659a059eac68d8d
CRs-Fixed: 2488966
Wifi turn-on will timeout and fail if system suspend occurs during idle
restart. To solve this, abort system suspend during idle restart.
Change-Id: I0a4a1fc9525b5fb19f51a289ca0b8a76c73038e6
CRs-Fixed: 2506716
Currently the host driver updates the okc, pmksa flags
only after the set key management command is received from the
wpa supplicant. This is not needed as set_pmksa command is
received for 802.1x/SAE/OWE and key management needs to be
offloaded to firmware always for it to be able to include PMKID
in reassoc request/responses.
Receive the pmkid_modes from ini config parsed to mlme and
update in 11i tlv param of ROAM_SCAN_MODE command.
Change-Id: If169b402731e3713479e90b627060df470e5dab2
CRs-Fixed: 2506580
The host driver drops any authentication frames that doesn't
belong to current pe session. But, for WPA3 SAE roaming
pre-authentication is offloaded to user space, and as part of
that, the host driver should handle the SAE commit/confirm
messages received from the target AP for which there is no
pe_session.
Add changes to check the auth algorithm in the authentication
frames received for which no pe_session exist and if auth algo is
SAE, indicate that frame to wpa_supplicant.
Change-Id: Ie911524475423704246c1b4f694b83abdbf5129a
CRs-Fixed: 2505556
RDDM can be collected during kernel panic and pld_collect_rddm is
a blocking call that will cause WLAN watchdog if the caller is from
scheduler thread. Hence revert this change and collecting FW dump
inside kernel panic handler.
CRs-Fixed: 2506844
Change-Id: I8667896e61b856fc2e03e1c8278a6451b6a04c24
Currently driver passes sta_id through out the call for getting the TSM
stats. Although passed, the sta_id is not used anywhere in the call
stack. As part of removing the sta_id usage, this sta_id can also be
removed safely.
Remove the usage of sta_id in get TSM stats path.
Change-Id: I1efbb1bfdc8df84e9b6a71dcd5bd39ea93ced21a
CRs-Fixed: 2484701
Increase the max number of MSDU in AMSDU limit to 15 and set the
default value to 0 to config the FW in auto mode selection in which
FW picks the maximum number of MSDUs in AMSDU based on connection mode.
Change-Id: I6c1f1292a31fc9ea857e60e7a81d10b6def5cbc1
CRs-Fixed: 2502544
Currently, the cdp API cdp_register_peer takes as input the sta_id. As a
part of cleaning up the usage of sta_id, replace it by peer mac address.
Change-Id: I0e92258777aff02940dd619c4d2ad784aa3e7f31
CRs-Fixed: 2502996
Cleanup ol_txrx_register_peer to be peer mac address based
from local peer id based.
Change-Id: I59fd19c0185d8fe89563ac78bc9e8a8117c12ed1
CRs-Fixed: 2503182
In standard linux kernel, struct dma_iommu_mapping is not defined for
arm64, while in qualcomm linux kernel, struct dma_iommu_mapping is
defined for arm and arm64.
To fix build error in standard linux kernel, use struct dma_iommu_mapping
only when ENABLE_SMMU_S1_TRANSLATION is defined.
Change-Id: Ib02073163f2a38c47f864c6592fe2470046868ce
CRs-Fixed: 2496733
Currently, the IPA AIP ucfg_ipa_wlan_evt takes as input the sta_id. As a
part of cleaning up the usage of sta_id, remove the parameter. Since
mac_addr is already present in the list of parameters, no need to add.
Change-Id: Iacc4a05de88973014eb4bcc1dfc97b7dc4a0d4fa
CRs-Fixed: 2503944
Currently driver passes sta_id through out the call for getting the SNR
stats. Although passed, the sta_id is not used anywhere in the call
stack. As part of removing the sta_id usage, this sta_id can also be
removed safely.
Remove the usage of sta_id in get SNR stats path.
Change-Id: I39482d8f6d8e3bd32e940a85d5592dada8f0460b
CRs-Fixed: 2503956
Currently the function hdd_get_peer_idx is not being called from
anywhere. As the use of sta_id is being depreceted, the function can be
safely removed.
Clean up the unused function.
Change-Id: I6896d03ad2e8b427147e2259b737836cd3f44ca0
CRs-Fixed: 2484699
When FEATURE_WLAN_AP_AP_ACS_OPTIMIZE enabled, it will allocate
memory in INIT domain (wlan_hdd_startup), and free in ACTIVE
domain (hdd_wlan_exit).
Move to wlan_hdd_start_modules/wlan_hdd_stop_modules to avoid.
Change-Id: I0fd93baa306a136af2faa7af9dcccfb709984ef9
CRs-Fixed: 2505462
If firmware doesn't support 5GHz band then ieee80211_supported_band
structure for HDD_NL80211_BAND_5GHZ is not allocated in wiphy and in
wlan_hdd_update_ht_cap() trying to derefer member-fields of
ieee80211_supported_band[HDD_NL80211_BAND_5GHZ] without NULL check is
causing NULL pointer exception.
To address this, add NULL check for 2GHz and 5GHz band pointers.
Change-Id: I8aa93b9cbe26ce674563505e222f5cb2aa970051
CRs-Fixed: 2503133
Remove operationChannel from structure csr_roam_profile, remove
the code where value assigning to operationChannel take place.
Change-Id: If7cd64d4d7513000181f92faabd6c863341c71f9
CRs-Fixed: 2503043
1) Remove channel from struct hdd_mon_set_ch_info. All client
code will use freq within the same struct from now on.
2) Replace the usage of channel ID(op_chan) with channel
frequency(op_freq) in function hdd_mon_select_cbmode's
parameter list.
3) Fix a number of camel cases detected in function
hdd_select_cbmode.
Change-Id: Ib7c1ee453b34647e983749a713b33017b92103ac
CRs-Fixed: 2504779
After got restart channel rsp from fw, vdev state will change
as: ST-RESTART_PROG->ST-CONN_PROG.
In lim_send_sme_ap_channel_switch_resp, if channel is non-dfs,
vdev state will change to up; if dfs channel, later it will
change to DFS CAC when starting cac timer.
Current issue is if there is ap stop after restart channel rsp,
while processing WLAN_VDEV_SM_EV_DOWN, there is no hanler in
ST-CONN_PROG > ST-DISCONN_PROG.
From design perspective, In SAP, ST_CONN_PROG is a dummy state,
ideally, SAP state should change as below without preemption :
RESTART_PROGRESS->CONN_PROGRESS->UP
RESTART_PROGRESS->CONN_PROGRESS->DFS_CAC_WAIT
To fix issue, change vdev state to DFS_CAC_WAIT in
lim_send_sme_ap_channel_switch_resp; So WLAN_VDEV_SM_EV_DOWN will
be handled in DFS_CAC_WAIT state.
At the same time, set DFS_CAC_WAIT when starting SAP, and clear
sap_move_to_cac_wait_state in sap state machine.
Change-Id: Iee89521471e456a553f40577da6d1e69aef3b803
CRs-Fixed: 2501339
Currently the driver choses a default channel for SAP
in case of failures, logic of which is incorrect as
it does not take into account the hw mode, and all
other checks such as SRD, unsafe, DFS etc.
The channel chosen after this maybe not as per
expectation.
Fix is to chose a default channel from the acs channel
list itslef so that the checks of SRD etc. ae honored.
Change-Id: Ife8a6b88938fd566bed92a5bde1cb0fe34a45462
CRs-Fixed: 2476351
Currently the driver sends the CSA IEs in the
beacon every beacon interval, and updates the
CSA IE count in every beacon.
If the wlan gets suspended in between the
updation of CSA IEs, the CSA is delayed
till the next resume, which could lead to
STA kickout event, if there is delay between
the CSA period, and the channel switch time.
Fix is to take a wakelock till CSA is completed
in order to avoid the STA kickout.
Change-Id: Iff03476433c755cbddc7568ffbd24ddb81fd1c90
CRs-Fixed: 2504039
Currently the driver checks whether the device
supports antenna sharing, and if the AP is added
in the OUI framework, then the driver modifies the
nss value to 1 to avoid sending SMPS to the peer AP.
Now suppose the device does not support Antenna sharing,
but supports DBS and is helium HW, then going to DBS HW
mode would result in peer sending a SMPS frame to the peer
as the helium HW only has two antennas, and one is needed
by each MAC now.
Fix is to add a third param in force 1x1 ini which would
decide the driver should consider the antenna sharing as
mandatory or not.
Change-Id: I3ae00fcbd642c7780952d66ccbf1208335fcb077
CRs-Fixed: 2496831
After sending the wmi_peer_delete command to FW. If there
is a change in channel info of HT cap from connected AP.
Then TDLS peers in lim will be deleted and dph entry will be
freed.
While processing the tdls del sta response from FW
if dph entry is NULL then WLAN_SER_CMD_TDLS_DEL_PEER is
not released and that can lead to active command timeout.
Release the serialization command in case of NULL dph entry.
Change-Id: Ib56806875cc84a0ecb9bcab16e9e0bb9d118efbb
CRs-Fixed: 2503993
Check if the configured tdls offchannel through IOCTL
is supported or not. This is required if firmware
does not support 5G channels or any invalid channels
Change-Id: I49e3b7b758fddf22fbfe1b64dda2e4f47a27cf3b
CRs-Fixed: 2503913
On wlan suspend, kck sent as part of gtk offload request
is not populated properly since the kck_len is not
updated when caching the gtk request resulting in all
zeroes kck.
Fix is to update kck_len when saving gtk offload
request.
Change-Id: I9e4430565a04cff6da7f06d23a9542a2f53568b0
CRs-Fixed: 2502131
Currently the driver sends a unicast probe request
to the BSSID which is blacklisted, which in turns
violate the SPEC, as the BSSID maybe rejected as
part of OCE assoc reject.
Fix is to check if the BSSID is blacklisted, don't
go for unicast probes.
Change-Id: I14f283d7425d40706cef29db38b48333ee9a7fff
CRs-Fixed: 2500273
If no channel is found suitable for ACS, don't scan
all the channels as done currently by the driver, and
send ACS fail code to the userspace as SAP cannot be
started on the band specified by the use.
Change-Id: I18edff6aaf384384841487d06474fd5fba28fc2f
CRs-Fixed: 2476347
Currently for NDI interface bus bandwidth timer is not
started. As part of this change start the timer when new
ndp peer is created and stop the timer when last ndp peer
connected is deleted.
Change-Id: Ia710ec9872636ee3cefa30815692c2f8c9816687
CRs-Fixed: 2499548
Fix the compilation issues passing proper parameter to the function
wmi_unified_thermal_mitigation_param_cmd_send.
CRs-Fixed: 2501080
Change-Id: I9ce0038ef2844536612f0bf71314dbc214f354d9
Currently, get_tx_power is given for all interfaces but FW may
not respond for interfaces like p2p device mode or NAN.
Allow cfg80211_get_tx_power only for STA/P2P_client associated
state and SAP/GO started cases. In rest all cases do not give
this command to FW.
Change-Id: I01ff4f11c35ca00183002d2d1c2c481f3bea77c6
CRs-Fixed: 2502905
Get station stats API will be periodically called from upper layer so
in order to save power, instead of waiting for inactive timer to fire,
suspend right after everything is done.
Change-Id: I452ac3631d6799ffdcb1d09ff71afb2730b92b59
CRs-fixed: 2502391
Modify the nud tracking ini so that the driver
does a disconnection/roaming after NUD failure
is detected.
According to the ini specified, the driver would
either disconnect or do roaming.
Change-Id: I6af1c819322a702605d92ada735298a124127533
CRs-Fixed: 2497153
Associating peer's phymode is not allowed to be highter
then sap's. If it is, set the peer phymode equal to sap
phymode.
Change-Id: Ia8152ce27fb8ddf2bbff5139d3b8d89c811d7ed0
CRs-Fixed: 2500383
Do more frequency list population in a number of functions
to make sure the resulting tCsrChannelInfo structures
have the right channel frequency info.
Meanwhile, enlarge the buffer allocated for freq_list in
csr_roam_copy_profile to make sure all channel frequencies
can be replicated.
Change-Id: Ia8b90581211d657851e1d8fc2556dcb633c8001a
CRs-Fixed: 2500236
