The regulatory API reg_freq_to_chan_for_chlist() uses num_chans as the
bound to access the array chan_list. While an invalid index INVALID_CHANNEL
is prevented from accessing chan_list, values that are greater than
NUM_CHANNELS are not prevented from accessing the array.
Prevent access to the array when the index is greater than NUM_CHANNELS.
Since INVALID_CHANNEL is also greater than NUM_CHANNELS, the existing
condition that prevents access to the array when num_chans is
INVALID_CHANNEL can be removed.
Change-Id: I8b9793b84fc5b3a9406389fd3482673dd4667cfc
CRs-Fixed: 3196165
Use microsecond timestamping for the tx descriptors
for capturing the latencies more accurately. It will help
to determine the appropriate bucket.
Change-Id: I216e083a7c89b01b6f2f384c1c0a85ca323d3a3e
CRs-Fixed: 3165153
Bypass management Rx reorder if the number of MLO vdevs the
given pdev is zero.
CRs-Fixed: 3194267
Change-Id: Ia286999fdf99bb2b77008c31dfb3a52731cd5167
With MLO enabled, Rx monitor receives the sw_peer_id in the following
format in MPDU_START_TLV:
+---------------------------------------------------------------------+
| 15 | 14 | 13 | 12 | 11 | 10 | 9 | 8 | 7 | 6 | 5 | 4 | 3 | 2 | 1 | 0 |
+---------------------------------------------------------------------+
| CHIP ID | ML | PEER ID |
+---------------------------------------------------------------------+
For MLO station, sw_peer_id received in MPDU_START_TLV corresponds
to MLD peer. But for Rx monitor framework, link peer makes relevance.
Therefore, fetch and update the sw_peer_id with link peer_id in the
received ppdu in Rx monitor.
Change-Id: I8ffc32856fd364edbcec2d98c3f9e1f0643d40d1
CRs-Fixed: 3190668
Extend the peer queue flush command with following attributes
1. Enable to flush per TID peer queue
2. Enable to configure when to flush the peer/TID queue
Change-Id: I1c8a9ca82cdfdd67eaf69b73a8a7076ca9c02dab
CRs-Fixed: 3130761
Set num_max_active_vdevs to WMI init command and
extract it from WMI ready event.
Change-Id: I6d6377c03214fcab74c675839f136301ebf27834
CRs-Fixed: 3183010
As part of MLO stats feature, add support to send ll_stats
and get_station stats request to firmware for all the MLO vdevs.
Change-Id: Ic50a6294f868f0604d4cfc0d4b657b6085da2fb4
CRs-Fixed: 3181020
Modify CE4 flag in host configuration to support enablement of
copy completion interrupt for Lithium and Beryllium based
targets.
Change-Id: I71ccbaec9707dad59bcb2ee98919285a22eef351
CRs-Fixed: 3195349
Add the variable sta_sap_scc_on_indoor_channel in
psoc and pdev reg objmgr, to access the INI value of
sta_sap_scc_on_indoor_chan
Change-Id: Ie0c41d75d0f617274743f5196e65dab93414b7eb
CRs-Fixed: 3191788
In VoIP cases, no or minimal packet drop is expected during
roaming. Currently, data packets are dropped in host if they
are received after roam sync indication from firmware and before
peer map indication is received due to invalid peer.
This change allows rx UDP data packets during roaming before
peer map indication is received from FW if peer's auth_status
received in roam_sync_indication from FW is authorised.
Change-Id: Ic518b7cd0aef48a2b0f8c923a2e20838b07f3d1f
CRs-Fixed: 3168603
add tags straight from ppdu_info to headrom for mpdu
tags are positioned per msdu_index within mpdu headroom.
Change-Id: I4863a7929ffd2e0a5d7f91f31ab6059f2b882de4
CRs-Fixed: 3183692
Replace qdf timers with HR timer for pre CAC timeout and process
the preCAC completion work in workqueuecontext.
Since preCAC expiry is processed in H/W interrupt context
(while using HR timers), no other timer cancellation is allowed
('qdf_hrtimer_cancel' function will wait for the handler finish)
But if we process the preCAC timer completion work in H/W context,
other timer cancellation will take time. To fix this, process the
preCAC completion work in workqueue context to allow other timer
cancellation.
CRs-Fixed: 3185213
Change-Id: Iaa24055e0d485f613e16d6110d22002fecf10821
Add mapping of wmi_vdev_param_set_traffic_config
to WMI_VDEV_PARAM_VDEV_TRAFFIC_CONFIG.
Change-Id: I2147cad3fb5726551de64edcfc41bee2707150a6
CRs-Fixed: 3191504
MLO peer attach fails if MLO peer with same MLD address gets
created. While freeing MLO peer, releasing of link peer ref
is missed
Fixed some of the corner cases handling with ml peer id, AID
Change-Id: Ia3ff13a3840083bfe389b39086a6c5aa1f709fca
CRs-Fixed: 3185069
Currently 6 GHz channel flags are not set for channels
present in the list from PNO scan request but are set for
all other 6 GHz channels which are added when either of the
two below conditions are satisfied
1. Userspace sets NL80211_SCAN_FLAG_COLOCATED_6GHZ flag
in PNO scan request.
2. At least one 6 GHz channel is present in the PNO scan req.
Add flags for channels in the PNO scan request based on the
scan_mode_6g ini configured similar to active scan request.
Add changes to remove the RNR flag for the channel if the
corresponding short SSID entry is present in the RNR db
Cache since if colocated SSID is different userspace doesn't
include corresponding 2.4/5 GHz channel information in PNO
request.
Change-Id: I22c4f7ecf65d4ab09dfbbe126aefe02f9145de42
CRs-Fixed: 3119113
Disable obss scan for all vdevs under each pdev
when try to suspend, not just only disable the first one.
Change-Id: I52e46f43727acf004165f39a262b018dc174cc7f
CRs-Fixed: 3190706
In QCN9224, when 11v DMS feature is enabled use
FW based DMS support instead of SW based WAR.
Change-Id: Icc5b9740f71bd00862e06950e4fe49c663d513b4
CRs-Fixed: 3182547
Currently, in case where AP advertizes BA window size
which is different than our max supported BA window size,
the HW gets configured for BA window size advertized by
the AP, whereas the ADDBA response indicates the window
size of the STA, which is less than the AP.
This mismatch leads to the BA bitmap to be incorrect.
Fix this by exposing an API from datapath for the update
of the final BA window size which is populated in the
ADDBA response.
Change-Id: I81a039f47a775990b426e9aca1daaaa9bdb6e88f
CRs-Fixed: 3187941
Count the MLO vdevs in a pdev. WLAN_VDEV_FEXT2_MLO feature
flag in vdev object is set for all MLO vdevs. Simple solution is
to increment/decrement the count on every set/clear of
WLAN_VDEV_FEXT2_MLO feature flag. Since this feature flag will be
set/clear at multiple places for a given vdev, this approach
will lead to wrong MLO vdev count. To fix this we need to
increment/decrement the count on first set/first clear of
WLAN_VDEV_FEXT2_MLO flag. Add a lock also to prevent the race
conditions.
CRs-Fixed: 3106235
Change-Id: Ice7edde04553088fbb7c9b769508d441ccd6e4bf
This is a WAR to match the TX_DESC_POOL size
to maximum number of VDEVs allowed.
Change-Id: I646a67ef2b611bea1ca5a6e2bf781a9454d409ed
CRs-Fixed: 3168359
Add init sequence handling to support AFC device deployment type
Set AFC device deployment type to regulatory private obj
Change-Id: Ib041d2010552412aaaa41395915d948aa2ea2bba
CRs-Fixed: 3190288
Add AFC regulatory target interface to support AFC device
deployment indoor or outdoor type in init sequence between
FW and host
Change-Id: I3caed77e943fdcaa947cb7e514c88b68604f55a3
CRs-Fixed: 3190262
Add AFC device deployment type support in regulatory pdev
private object and add API to return the deployment type
Change-Id: I45e44766b25b21ce3024d73f4ed7a4daf3ecb026
CRs-Fixed: 3190240
Create data structure to store the deployment type(location) of AP for
AFC. Append this structure to the end of AFC start request.
Change-Id: I19756fd3ca7a77f11b673d2bf3b0382f29732847
CRs-Fixed: 3177624
Declare the AFC deployment type enum and add deployment type member to
pdev_priv_obj.
Change-Id: Id11a0aac4fcc1b655436d9dbd876607178f27bd4
CRs-Fixed: 3186216
When AP moves from DFS to NON-DFS channel, dfs chan state is
not updated and have state DFS_CAC_STARTED. Similarly, if the VAP
operating in a DFS channel goes down, the dfs chan state needs
to change to DFS_CAC_REQUIRED but the dfs chan state remains in
DFS_CAC_STARTED.
During CAC STOP, send the DFS event CAC RESET to change the dfs
chan state.
Change-Id: I21831e2c79da67375f8e340964f899ebd169fd85
CRs-Fixed: 3160346
"-" is used to differentiate between state and substate.
User "_" in substate string convention to avoid confusion between
state and substate strings.
Change-Id: I9efb8fd7edd60f5d28c8f774fb17bda49e64ca24
CRs-Fixed: 3188646
In the case of target recovery avoid accessing of HW registers
during the VAP down.
Change-Id: I4061d75ec0c153710aaa99ab7414a96136ff22f5
CRs-Fixed: 3190730
For KIWI, STA AST index lookup search is enabled and remote
AP peer's AST index will be stored in STA Vdev if receive
AP peer htt peer map msg. but if STA TDLS connection happened,
host will receive another peer map msg for remote TDLS STA peer,
this remote TDLS STA peer's AST index will overwrite original AP's
AST index. if TDLS disconnected, STA vdev will still use remote
TDLS STA's AST index for TX, then TX to AP might fail.
Add is_tdls_peer flag in dp peer and configure this value by
cdp_peer_set_peer_as_tdls() from control path right after TDLS peer
created. if is_tdls_peer is true, do not store this TDLS peer's
AST index to STA Vdev, skip bss_peer flag setting as well.
Change-Id: I7b5df1caca6a0b5305a6e867cd92099b5f6a4890
CRs-Fixed: 3151035
Userspace sends the vendor command WIFI_TEST_CONFIGURATION by
setting the attribute WIFI_TEST_CONFIG_IGNORE_H2E_RSNXE when
driver needs to ignore the SAE H2E mismatch for 6 GHz connection.
This is needed for certification test support
Change-Id: I5c1838598478ab74d72d57a765452d7fd989b701
CRs-Fixed: 3072578
- Add NULL check of the pointer "afc_req" in the function
reg_get_partial_afc_req_info() before it is dereferencedi to avoid
NULL pointer derefernce.
- Add array size check in the reg_compute_chan_to_freq_for_chlist()
API to avoid the array out of bound access to "chan_list" array.
- Make use of the function reg_is_supp_pwr_mode_invalid() which
does the boundary check for the arrays chan_info->state_arr[] and
chan_info->chan_flags_arr[] to avoid out of bound array access.
- Add a check in the reg_freq_to_chan_for_chlist() API to avoid chan_list[]
array out of bound access.
Change-Id: Ifcade9e971a77f483d5c5138206eab47454f28ad
CRs-Fixed: 3184214
Update TLV length in peer map handler to correctly point to the next TLV
header.
Change-Id: Ie60c62040cb35b60a3bfbd302a176294903c94e6
CRs-Fixed: 3183057
In mlo_ser_set_link_req, req is again allocated and thus
the req->ctx.cb_arg doesn't points to the new memory.
so avod reallocating the req memory again in
mlo_ser_set_link_req and use the passed memory directly
as umac_cmd.
Change-Id: I0bdc9cf746b2a36c69b5ca0f773f06900a9783dd
CRs-Fixed: 3188736
