Some p2p device include invalid RSN IE in wildcard probe respone
frames in p2p discovery. Do not drop it because the wildcard
SSID entry will not be used to do association.
Change-Id: I1e390509931094d9d9a55d2caef825c5772a086b
CRs-Fixed: 3282030
If get scan filter for link vdev, that means the MLO STA has
more than 1 link, do not filter the link vdev's scan filter
channel list. Link that is causing MCC with legacy STA can
be disabled in post conenction.
Change-Id: I4aafb7168e60bf158904d3b1c4a4c42d4eaeb126
CRs-Fixed: 3285312
In reg_update_sup_ch_entry_for_mode function, mas_chan_list_6g_ap is
copied to super chan list when AFC power event is not received.
mas_chan_list_6g_ap would have SP channels enabled based on the regulatory
rules and this is propagated to super channel list. In outdoor deployment,
the SP channels can be marked enabled when AFC power event is not received
as there is REGULATORY_CHAN_AFC_NOT_DONE flag set for such channels. But
in indoor deployment, as this flag is not applicable, SP channels should
not be marked enabled.
To fix this, return from reg_update_sup_ch_entry_for_mode function before
enabling the SP channels if the deployment type is not outdoor.
Change-Id: I3c5ccdc4826c33002bf1bb33f686404fa0ffcc65
CRs-Fixed: 3270672
Host driver sends RSO stop request to firmware as part of
disconnect and waits for RSO stop response. It pauses the
disconnection and resumes once RSO stop response is received
from the firmware.
But cm_lock is not taken when disconnection is resumed from RSO
stop response path. This causes synchronization issues with
north bound thread which issued the disconnection.
Acquire cm_lock by using cm_sm_deliver_event() API to provide
synchronization between disconnect initiator thread and
scheduler thread.
Change-Id: I3c287a020e0a12abab2ad6237594e022adc904a7
CRs-Fixed: 3287190
When roam from legacy to mlo, vdev1 is updated first, ml peer is created,
but failed to attach since it is link vdev, then vdev0 is updated, ml peer
is created again, ref count is 1.
When disconnect, stop vdev1 first, in stop event handler, ml peer ref count
becomes 0, mlo_peer_cleanup is called unexpectedly.
mlo_peer_cleanup should be called after MLO vdev0 stopped.
To fix it, change wlan_mlo_peer_create, mlo_dev_mlpeer_attach is called if
the ml peer not attached and can't be found.
Change-Id: Iae3b2b498849646ae71154484b555a7fc9a36017
CRs-Fixed: 3277886
Per spec 11be_D2.1.1, the TSF Offset subfield of the STA Info field
indicates the offset (Toffset)between the TSF timer of the reported
AP (TA) and the TSF timer of the reporting AP (TB) and is encoded
as a 2s complement signed integer with units of 2 µs. Toffset is
calculated as Toffset= Floor((TA – TB)/2).
Change-Id: I7810568f6308e369dcf2ff26bdfd1246783466d4
CRs-Fixed: 3276836
The macro QCA_DFS_BW_PUNCTURE depends on the macro
WLAN_FEATURE_11BE. This creates compilation issue and fix
is introduced in this patch.
CRs-Fixed: 3283332
Change-Id: Ie00f7384142b39c309845408c6268ae6346f4734
The macro QCA_DFS_BW_PUNCTURE depends on the macro
WLAN_FEATURE_11BE. This creates compilation issue and fix
is introduced in this patch.
CRs-Fixed: 3282072
Change-Id: I680670e58bd57d104bc69f2b4e7a56010b7584a5
When radar is found in any DFS channel, a radar bitmap is
generated based on the radar infected channels. For example,
if operating channel is 100 HT160 and radar found in 112,
then the generated bitmap is B0000_0010. This proposed bitmap
is compared with the valid puncturing patterns. In this case,
the proposed bitmap is valid puncturing pattern.
In case of a valid puncturing pattern, send the final
bitmap to DFS module and the module updates the DFS channel
puncturing pattern. If an invalid puncturing pattern is
found, return value zero and do not update DFS channel
puncturing pattern.
CRs-Fixed: 3273800
Change-Id: Iee23ac267bfea1b4815a72c2728555f96bed6466
When teardown completeion event is received from FW, teardown
completion handler resets pdev link to NULL, when the expectation
is only to put pdev state in teardown, such that when soc goes
down the list gets cleared.
Change-Id: Ief490eabe0546207f0ef649cb6d5cde1faf582d8
CRs-Fixed: 3280671
There are two different name with different range of value for
aliasing INI's. So, it should pick the valid range of value wrt INI
name.
Change-Id: I1b3f8b276a65f2e1659dd4e3ede8512513e3da71
CRs-Fixed: 3268728
When radar is found on any DFS channel, generate a
bitmap which represents the channels that are
radar infected. For example, if the current channel
is 100 HT160 and radar is found in chan 104, then the
bitmap generated have the value B0000_0010. The bit 1
represents radar infected channels.
The radar bitmap calculation also checks for any existing
punctured patterns. If already punctured patterns are
available then it includes the existing pattern and
also current pattern in the bitmap calculation.
CRs-Fixed: 3273737
Change-Id: Icfc5d4e22980d3d2375152c12d244c7a8f5e318d
When set_fcc_channel is set, host driver sets tx_power info about
channels 12, 13 to hardcoded values. Now firmware sends information
about channels 12, 13 to host driver and currently there is no
support to extract and store this information and use it to
modify tx_power of channels 12, 13 in host driver.
Add support to store the tx_power of channels 12, 13 received
from firmware and use this to modify tx_power of fcc channels.
Change-Id: Ie7b20a75bb09956a70b7b133ce7ce9beb8db138b
CRs-Fixed: 3270676
Add the global dispatcher function pointers to enable/disable
the DFS puncturing feature.
Add a new dfs member "dfs_use_puncture" to remember if the
feature is enabled or disabled.
CRs-Fixed: 3272765
Change-Id: I1f45a9f44ede7f8606f770365220c5f7508847cf
Issue: The offset parameter in bangradar unit test parameter is only one
byte long. Whereas radar offsets has a range of [-160, 80] MHz if host
supports 320 MHz bandwidth. but with bangradar unit test command only till
-128 MHz is possible
To fix this, Add frequnecy offset sign bit(13th bit) if hosts supports
320 MHz bandwidth. Firmware will identify max range of [-160,80] with
sign bit.
Change-Id: I33ea8e9185f8d25b81cad7a18dcca6d2d6f42669
CRs-Fixed: 3270525
Set dfs object parameter if host and fw supports 320 MHz bandwidth.
i.e. dfs->dfs_is_bangradar_320_enabled. Set to true if the host and
firmware support the bandwidth is 320 MHz. If there is no support, set
false.
Change-Id: I210823c448450f89dc65937201b4839168369d18
CRs-Fixed: 3270525
Currently, there is no ap reg rules from the regdomain for
some countries. But function reg_append_mas_chan_list_for_6g_lpi
and reg_append_mas_chan_list_for_6g_vlp get the master chan per
the ap reg rules number. It casues it can't get the master chan
list for 6 GHz even it has client reg rules.
To resolve this issue, using num_of_6g_client_reg_rules instead of
num_of_6g_ap_reg_rules.
Change-Id: If2dc6510edc61531cb1a3b8002e43a1bd69311e0
CRs-Fixed: 3278250
Add API to extract MLD AP MSD capabilities from MLO IE
common info field if present. This will be shared with FW
via peer assoc cmd.
Change-Id: I0ebcd5408a40f3314932d4a2a7e586c208af2ee5
CRs-Fixed: 3271118
Currently, the reg_get_client_power_for_connecting_ap() API
populates is_psd_power flag within the function and uses it
as a check to further populate EIRP power. The is_psd_power
flag is derived from current channel list chan flag which
returns true if corresponding channel supports PSD power.
Normally, all 6 GHz channels support PSD, so this flag is
usually set to 1. But, AP can transmit EIRP power in TPE IE
for 6 GHz channels, thus for MCC specific cases, derive this
flag based on tx_power interpretation field in TPE IE for
accurate value. WIN Host can still use reg_is_6g_psd_power()
to retrieve the flag in the caller APIs.
Hence, derive is_psd_power flag from TPE IE interpretation
value beforehand and pass it as an argument to
reg_get_client_power_for_connecting_ap() API.
Change-Id: Iabbcbd003f441151643a087ad4908bcdaed753a5
CRs-Fixed: 3268118
As 11be 2.1 spec 35.3.3.3 Fields and elements not carried in a per-STA
profile, an AP affiliated with an AP MLD shall not include a Timestamp
field, a Beacon Interval field. Change ML probe rsp per-STA profile decode
logic accordingly. Copy Timestamp from the starting of the probe response
frame.
Change-Id: I0fe5682c170dc3dcd6e5a93c68473cd4cb6999b8
CRs-Fixed: 3280227
Add rxops reg_display_super_chan_list, its core and dispatcher functions to
display super channel list for all modes.
Change-Id: I2b094d407d318ad84ac79ea3a4bc1127606c0fa2
CRs-Fixed: 3212642
This change adds protection, if partner link peer creation failure
leads to MLO peer free. It also returns failure.
Change-Id: I4f2097c3a2942cecf01f77c7e4899595bbce0dff
CRs-Fixed: 3277770
Currently, short SSID is not added to the hints in scan request
if BSSID hint is present. But there can be cases where both need
to be advertised and expected to be present in the probe request.
So, send short SSID hints also to firmware in scan request.
Change-Id: Ie66e6677185c106817f4d24c2f07730c53ae243a
CRs-Fixed: 3259555
Current code directly passes the result of wlan_vdev_get_psoc(vdev) to
wlan_objmgr_get_peer_by_mac, without null pointer checking, which can
result in null pointer dereferencing.
To fix this issue, add checking of the psoc pointer before passing,
and add verification inside wlan_objmgr_get_peer_by_mac.
Change-Id: Iaf0e2535c3a584db788ddc621745c66c296fe08f
CRs-Fixed: 3277651
For T2LM capable MLO STAs the AID is allocated from the second pool.
There is an inconsistency in allocating these AID pools.
Please check the example below.
Let's say the max_aid in ML AID manager is 12 and the start_aid is 3.
As per the current implementation,
aid_end1 = (ml_aid_mgr->max_aid - start_aid)/AID_NUM_BUCKET
= ( 12 - 3)/3 = 3
Only 3 AIDs are available in pool_1. But, all these 3 AID are reserved due
to start_aid = 3. Hence, we will not be able to allocate any AID from
pool_1.
With the fix,
pool_1_max_aid = (ml_aid_mgr->max_aid - start_aid)/AID_NUM_BUCKET
= 3
aid_end1 = pool_1_max_aid + start_aid
= 6
So for pool 1, AID 4,5 and 6 are available now.
Change-Id: Ia96966b542e68511acbf46de32448f0d95d31c69
CRs-Fixed: 3276564
Handle the following error scenarios in management Rx REO module.
1. Pdev corresponding to a HW link is deleted.
2. Pdev corresponding to a HW link not yet initialized completely.
CRs-Fixed: 3272812
Change-Id: I34fbb19e24649371ff0ddb591e2ec14bfdfee58a
Currently, for all link peers of ML peer, is_primary set as
true, since all link peers are attached to same PSOC.
But, FW and DP needs only one peer to be set primary.
So far, FW and DP are considering last peer created or last peer
assoc received as primary peer.
This method causing issue since the order is not guaranteed
between FW and DP layer.
So, added a change to set is_primary to one of the link peers
only
Change-Id: I1c1aa87056baf86091fefc780180b5fc6a16af0d
CRs-Fixed: 3274360
Multiple instances of the misspelling "recevied" are present, so
replace them with the correct spelling "received."
Change-Id: Ibd7f4a136541be8aa98f5bee4db7f75af0271f53
CRs-Fixed: 3273526
Add a flag to indicate assoc req is used for this roaming,
don't treat assoc req as reassoc req.
Change-Id: I58f357b5581ca71e97841f1d7e0f3daad5d60970
CRs-Fixed: 3273877
CH_WIDTH_320MHZ definition is in 11BE define currently and it's
causing duplicate code in 11BE supported/unsupported cases.
Definition need not be in 11BE define as it doesn't cause any
impact on memory.
Change-Id: I996aad941bef33eef6527c4c34c5b9efd75d9b84
CRs-Fixed: 3246577
During dynamic mac address change process, some link vdev mld
address will be changed successfully and some link vdev mld
address will be failed to be changed from target in failure case.
In vdev deleting, if no result to find mlo dev ctx by mld mac
address, try to use vdev->mlo_dev_ctx to detach vdev from mlo
dev ctx. This will avoid memory leak in above failure case of
dynamic mac address change.
Change-Id: I11304c92f9258e2390cfe2d03f29ada0db80e6af
CRs-Fixed: 3271092
For fixed field of per-sta profile for mlo, only assoc rsp is parsed,
need include reassoc rsp.
Change-Id: I8be3ab2d2f7719bda0190b59c6d24c07d12f21e8
CRs-Fixed: 3273819
Prensently for initial scan and roam passive scan dwell time,
the same ini CFG_PASSIVE_MAX_CHANNEL_TIME is used. This will
cause initial scan to have the roam passive scan value.
Decouple passive scan and roam passive scan ini parameter by
adding new ini param for roaming
Change-Id: I7bffb24a1e45548ed57492abc35e80958e9955ee
CRs-Fixed: 3272817
Currently host sends many vdev or pdev set param commands
separately to firmware.
Combine all vdev or pdev set params together where ever
possible, send to WMI.Based on service capability
WMI_SERVICE_COMBINED_SET_PARAM_SUPPORT, host decides whether
to send all set params to firmware at once to reduce number of
transactions or need to follow legacy way i.e send params
separately one after the other.
Change-Id: I0a36dd66a7ccb27f2ecd437cf7b9d08667e27472
CRs-Fixed: 3151418
Avoid code duplication by using common definitions
i.e use enum wlan_ml_linfo_subelementid.
Change-Id: Ia09c1a42207461878d023e8f4534f2d26fb2f81b
CRs-Fixed: 3227859
In current code, deauth is skipped for PMF clients
but ML peer state is updated for PMF clients also
This fix skips ML Peer state update for PMF clients
Change-Id: I41c870a5ff4bb658f378b65c729947fad324e807
CRs-Fixed: 3261671
Currently the INVALID_CHANNEL assumes the value 0xBAD. When a channel list
array is indexed, there is an array out of bound error when the channel
list array is indexed using INVALID_CHANNEL or when the index is greater
than equal to NUM_CHANNELS. So a check for an invalid channel enum is
likely to be, (ch_enum == INVALID_CHANNEL) && (ch_enum >= NUM_CHANNEL).
Redefine INVALID_CHANNEL enum constant with NUM_CHANNELS constant so that,
the validity checks for channel enum is compressed as,
(ch_enum >= INVALID_CHANNEL).
Change-Id: I5be588fe6011ac85031325f56eff208f2aa7a1e5
CRs-Fixed: 3238386
Based on the new requirement, add support to get requested
feature set info from different feature components.
Change-Id: I75c5a3062312b1124d21d1ae429a7c5a18d9f2d0
CRs-Fixed: 3262867
