Add changes to move ioremap of PMM SCRATCH register outside
interrupt context, as ioremap in interrupt context in not valid.
Change-Id: I0fa645c55e10a5241011a1f4ffdccdca48c50d29
CRs-Fixed: 3371194
Currently the debug log for NULL peer, just gives
the REO error code info, which may not be relevant
if the error source is not REO.
Dump both REO and RXDMA push reason and error code
if NULL peer is encountered in WBM error handler.
Change-Id: I95dabb8c9de1bf95081ccb389805d3ef3376915a
CRs-Fixed: 3374313
Cleanup the tx desc clean up logic used in
direct switch and umac reset cases.
The excess code being removed tries to
get the next descriptor address in boundary
conditions is not useful, as the descriptor
extracted this way will be overwritten
immediately in the beginning of next iteration
of the loop.
Change-Id: Ibcd873719929b94147152ff48205774d3ed4f452
CRs-Fixed: 3371831
In MLO scenario where start response for partner ML AP vdevs are
processed simultaneously on different CPU cores, the order in which
the vdev sm lock is acquired may lead to deadlock.
To fix add change to dispatch the MLO_SYNC_COMPLETE under mlo_dev_ctx.
Change-Id: I86ef51fcc8d51277163fcc5a6afdee6d873f2e63
CRs-Fixed: 3368033
remove unused var and declarations in
dp/wifi3.0/monitor/2.0/dp_mon_filter_2.0.c
CRs-Fixed: 3284798
Change-Id: I602d0e1ce5a74b981f34f7d5d4a5cb83693f62db
remove unused var and declarations in
dp/wifi3.0/monitor/1.0/dp_rx_mon_status_1.0.c
CRs-Fixed: 3284798
Change-Id: Iea3f8e2213fef9d10fe20c90c2ac0017f7d17692
remove unused var and declarations in
dp/wifi3.0/monitor/1.0/dp_rx_mon_dest_1.0.c
CRs-Fixed: 3284798
Change-Id: Ibfeaca928fe43e6f0cc5b52a2114708810c09805
remove unused var and declaration in
dp/wifi3.0/monitor/1.0/dp_mon_filter_1.0.c
CRs-Fixed: 3284798
Change-Id: I289aac6ae32c22539c98d02673b9af9a48d46192
Added changes in wifi module to update PPE2TCL & REO2PPE ring
size as 8k to improve the perf for MLO case
Change-Id: I6ca241c21c44b4cda8869e5e2ae9a4a906d98fc8
CRs-Fixed: 3365361
As per kernel doc, this file does not have proper
comments and spelling mistake are present, due to
which WLAN Automated code review for kernel-doc
checker is giving error.
Fix is updating the comments based on Kernel-doc
guidelines and correct the spelling mistake.
Change-Id: I7f8b8505b9810de2678ecfa893be78c0451cf612
CRs-Fixed: 3368472
Add API to check state of all mlo vdev to avoid roaming and disconnect
race issue.
Change-Id: Idc9a9a57f433aeff185fe78826faa84e08c09ddb
CRs-Fixed: 3372132
Add support for mlo groups in mlo soc/link setup and teardown
sequence participating in mlo.
Change-Id: I2b431b0ee1e5816b92ad183b52ad2998eb5bb46c
CRs-Fixed: 3343170
Currently in dp_peer_find_id_to_obj_add while adding peer to soc
peer_id_to_obj_map list, if peer entry is already present for the
peer_id due to no umap event for older peer, driver calls QDF_ASSERT
and returns, As QDF_ASSERT is NOP it results in older entry still present
in the table. which results in peer leak issue as while doing peer unamp
driver gets peer from soc peer_id_to_obj_map table based on peer id,
Since older entry is still there older peer will be freed and new peer
unamp will not happen.
To fix the issue call qdf_assert_always if peer entry is already present
for peer id which will always assert and add additional logs to print
peer information.
Change-Id: Ic777e9c7881fcbc67f54002bd619a8e6d52de126
CRs-Fixed: 3369736
Non-assoc link might be superior in features compared to assoc
link and the per-STA profile info may carry corresponding IEs.
These IEs are extracted and added to IE list of link probe
response while generating it. So, the link probe response
generated from assoc link probe response might be of more size
than assoc link probe rsp. It's caller responsibility to allocate
buffer for the derived scan entry considering the copied IEs from
ML per STA profile.
Add a check to validate the available buffer while copying the
ML IE (as it's copied after deriving IEs from per STA profile) as
this is missing currently.
Change-Id: Ieafc9730ad098abb80fb1f3c14eb22b6b590ff20
CRs-Fixed: 3364159
Initialize Spectral Scan detector ID to Spectral mode table for
QCA5332. Detector ID 1 and 2 are not supported for QCA5332.
CRs-Fixed: 3366784
Change-Id: I7fe9b53819c649455a67957827e4f900e39257e0
Rx PN error stats for broadcast management frames is being tracked
as a ucast stats. Fix it to be tracked as part of mcast stats.
Change-Id: If76c512107728b792ed6d92d56036325592f0fd1
CRs-Fixed: 3361840
from Beryllium onwards a single Flow Search Table
will be used across multiple radios. to ensure this
a single FST attach is called for the first SOC's
pdev and FST detach is called only during last pdev
of last SOC. the same FST instance is saved in all
pdevs of all the SOCs
CRs-Fixed: 3366409
Change-Id: I42fbdc6f09fb902021877e100a2831a1a24bc975
In error conditions(packet truncation or flush detected)
packet buffers are not freed, changes are done to
iterate over all the packet buffers and free them.
Change-Id: I48ef0022c7a9c8c6fd067d78d6b46daf84855cc8
CRs-Fixed: 3363841
Currently, BSS beamformee score is calculated from vht_cap or
eht caps, so for 6 GHz HE capable BSS it will calculated to zero.
Due to this 5 GHz will always get selected compared to 6 GHz BSS with
all other capabilities and conditions same in both BSS as 5 GHz
will have more total candidate score.
Fix this issue by correct calculation of AP beamformer capability
in case of 6 GHz HE BSS.
Change-Id: I0f3285ce1c5c1aeeba624c5371ff0f884ae78c58
CRs-Fixed: 3351991
Use mac_addr pointer as constant in wlan_hdd_ml_sap_get_peer,
hdd_set_mld_address and use the same for consecutive APIs.
Change-Id: I3ab7b0f3c03e9175c62179029efb53fb9fe75277
CRs-Fixed: 3342115
Currently HTT_TX_FW2WBM_TX_STATUS_DROP and HTT_TX_FW2WBM_TX_STATUS_TTL
are mapped to HAL_TX_TQM_RR_REM_CMD_REM which results in the wrong Tx
status print in case HTT_TX_FW2WBM_TX_STATUS_TTL where No ack is received.
So to fix the status log issue map HTT_TX_FW2WBM_TX_STATUS_DROP to
HAL_TX_TQM_RR_REM_CMD_REM and HTT_TX_FW2WBM_TX_STATUS_TTL to
HAL_TX_TQM_RR_REM_CMD_TX to log correct Tx Fail Status.
Change-Id: If125313f6b831ce5ca0f7eec9e8e3f1a03229f16
CRs-Fixed: 3362679
When mlo_connect is posted, mlo_dev_ctx lock is acquired.
As part of which is CM is not in init state, disconnection is posted
before processing of connect request.
In disconnect, mlo_sta_vdev_list is fetched which internally takes the
same mlo_dev_ctx lock leading to deadlock
Add fix to avoid deadlock in back to back connect handling.
Change-Id: Ic88229a7709e11edc790fb66401bb0336d8a1e27
CRs-Fixed: 3365031
In case of ml_activate_pend_disconn_req_cb, mlo_disconnect is called
which validates disconnect request. If the disconnect was deferred due
to race in mlo connect notify event and disconnect received by mlo
manager, then ml_activate_pend_disconn_req_cb will end up in a deadloop
where it will always defer any disconnect due to above race.
Hence, add API with flag to validate disconnect request.
Change-Id: Ifa434a2f2f5e7a4d17d25349cb7ab86fa50ce043
CRs-Fixed: 3361387
Validate the result of spectral specific ucfg control operation
after processing the control request to get the spectral system
capability information.
Throw error if it is an invalid operation.
CRs-Fixed: 3308464
Change-Id: I11bbd3a9a9ffc8462e420946d54f06a11fad558e
