Enable force wake recipe feature DEVICE_FORCE_WAKE_ENABLE
and disable the generic shadow register write feature
GENERIC_SHADOW_REGISTER_ACCESS_ENABLE.
Force wake recipe will be used to write to the REO remap
control registers by waking up the UMAC instead of using
shadow register writes.
Assert soc wake reg and poll on the scratch reg to check
if UMAC is awake.
Enable HIF_REG_WINDOW_SUPPORT to enable windowed reg
read/write in HIF layer.
Change-Id: Ib696e27e19a07c0084c097b95b7780b56e643c8b
CRs-Fixed: 2850590
In cm_allocate_and_copy_ies_and_keys instead of target's pointer
source's pointers are reset thus making the following code void
Fix this by resetting the target pointers which needs to be allocated
with new memory.
Change-Id: I4417492c43eb27833d365ec4aa644e07ed4eca51
CRs-fixed: 2884187
In current code logic, in case of channel utilization value is 0, the
congestion value treated as “not present” and give it a medium score.
This will cause a very clean channel get a medium score.
To fix the issue, need calculated congestion score from 1-100 and
not 0-100.
Change-Id: Iff030cf4a682d6e585d6f812a522cb8b6683cc44
CRs-Fixed: 2812978
During interrupt mask initialization bitmap of rxdma2host
destination ring mask is filled in host mode.
To avoid spurious interrupts reset this interrupt mask
when NSS offload is enabled.
Change-Id: I513a78639898c6f2bd9ea5418def03d64d7232d0
To check if a scan entry is a non-Tx VAP of our connected BSS, the current
implementation of scm_age_out_entries() searches through all scan entries
to find out our connected BSS. This check takes O(n) time where n = number
of scan entries. Running this check in the main loop of
scm_age_out_entries() results in an overall time complexity of O(square(n))
Time complexity can be improved if we do one iteration over scan entries
to find out our connected BSS, and save it in a local variable, and then
use that local variable in the main loop of scm_age_out_entries().
With this approach, we just do two iterations over scan entries
instead of nested iterations.
CRs-Fixed: 2877884
Change-Id: I51be764f4aa6f65ba9068d46eaa42791ac0f8d97
For some targets, FW expects WMI_REQUEST_UNIFIED_LL_GET_STA command
always over QMI irrespective of WOW state. FW will crash if this request
is sent over WMI on those targets. Currently there is no direct way for
the host to identify this requirement by FW. So FW may crash when this
WMI_REQUEST_UNIFIED_LL_GET_STA stats request is sent over WMI.
To resolve this, add service bit support for LL_GET_STA request over QMI.
With this change the INI parameter send_ll_and_get_station_stats_over_qmi
becomes obsolete and corresponding code will be removed.
Change-Id: I43f2827a0b1cae6e75f1224f452f8277d53736b2
CRs-Fixed: 2847483
Remove #ifdef WLAN_FEATURE_INTERFACE_MGR from connection mgr, as
interface manager .h files take care of it.
Change-Id: I5b1e07beecd93c98429318372ed72fdf414712a4
CRs-fixed: 2878656
When hif_force_wake_request is called from non interrupt context
use msleep to avoid mdelay since it is busy wait function.
Change-Id: Icda50e9d18b64369128a4f669ca1259931e5194a
CRs-Fixed: 2883178
For STA mode with STA CAC enable in auth/assoc state of assoc sm,
a northbound disconnect request would have to wait until CAC and
assoc is completed and connect request is removed from serialization
queue. During which OSIF timeout is seen as CAC can exceed the
serialization timeout of 60s.
Add change to handle EV_DOWN in disconnection progress state which
brings down the connection immediately on receiving a disconnect
request.
Change-Id: I301e49164c69df99e4617dd6fbc012e449963521
CRs-Fixed: 2882876
INI parameters parsing can fail because of fail to read INI file or
failed to parse at least one parameter. In some scenarios, driver should
assert if parsing is failed for atleast one INI parameter.
To differentiate these two failures, return different error status
values.
Change-Id: I187d2b2e8f55af47a14561302d8b62e84de7da5a
CRs-Fixed: 2878966
FW advertises WMI_SERVICE_OCV_SUPPORT bit in service ready event if
firmware supports Operating Channel Validation feature.
Change-Id: Ida1ee9874f9cd32a684ceee728e3323306137c64
CRs-Fixed: 2880794
QCA_NL80211_VENDOR_SUBCMD_GET_STA_INFO to get finer details on roaming
behavior, TSF out of sync count, and the latest TX rate, Rate Index used
for the transmission.
Change-Id: I34d112a263a587603f555069c0244f61dc44ec73
CRs-Fixed: 2876411
For WIN, a specific raw event handler is used for handling
WMI_CTRL_PATH_STATS_EVENTID received from FW.
As a part of control path stats infrastructure, event handler
for WMI_CTRL_PATH_STATS_EVENTID is different and this results
in wifistats functionality break for WIN side.
Make changes to use WIN specific event handler.
Change-Id: I2ac677c3ca55cd783734d1137c39b8ee1d5dcbf5
CRs-Fixed: 2878466
During wifi load and unload, multiple memory allocations and deallocations
are done in the path of dfs_deinit_precac_list, and at a particular
instance there is a crash seen due to an invalid paging request.
The QDF_TRACE_LEVEL_ERROR logs are not enabled by default. These prints
can be useful in debugging the issue when it occurs the next time.
Hence, the dfs_err prints are enabled by default.
When a crash occurs with the dfs_precac list functionality, debugging the
crash with dumps in T32 is difficult. During wifi unload, the api
wlan_dfs_pdev_obj_destroy_notification is called. On first step, the pdev
component is detached making the pdev component as NULL and after that
dfs component is detached. So when crash happens at this point, pdev
component is already NULL and analysing dumps with T32 is useless at this
stage.
Therefore, modify the order of dfs detach in
wlan_dfs_pdev_obj_destroy_notification. First detach the dfs component and
then detach the pdev component.
CRs-Fixed: 2872097
Change-Id: I157c6e6272bf4dd3676588b7ee6889fdb2efa5dc
Make CM serialization disconnect req as blocking or non blocking based
on feature macro
Change-Id: I71d664e28cba35d44c0c1fac2bd6a38c4bc4c9df
CRs-Fixed: 2879850
During multiple BSSID scan ie parse, there is memory allocation
on new_ie variable of size 1024 which may create buffer overflow
in util_gen_new_ie() if ie length is greater than 1024.
As part of fix, allocate memory of size ie length in new_ie.
And also add check before copying to pos variable in
util_gen_new_ie().
Change-Id: I55e0819817b5a616684067170bf28a314a145fc2
CRs-Fixed: 2867353
Add fields to the validate bss data struct to pass more information
to the interface manager when validating a candidate.
Change-Id: I82536585517da87ec0883c6010461d7ee929bd18
CRs-fixed: 2836749
After NOL expiry of the radar hit channel, the scan channel list
of the FW is updated to include 'NOL expired' channel as well
so that FW can initiate subsequent scan on this channel.
Only those channels which are not marked as NOL in regulatory
are updated in the scan channel list.
However, in the current flow of events, the scan channel list is
first updated via dfs_nol_update() API and then channel is
unmarked as NOL in utils_dfs_reg_update_nol_chan_for_freq() API.
Hence the NOL expired channel is not updated to FW and subsequent
scanning does not happen on this channel.
Ensured that the channel is cleared of NOL flag before updating
the scan channel list.
CRs-Fixed: 2877337
Change-Id: I7331bcc12c1dd6a26407c0753f28481cee1ce628
The host will send TPC power to FW via the new WMI cmd
WMI_VDEV_SET_TPC_POWER_CMDID going forward instead of
WMI_VDEV_PARAM_TX_PWRLIMIT to accommodate 6GHz channels.
But, to support backward compatibility for New Host+old FW
the TPC power will be sent to FW via legacy WMI cmds
WMI_VDEV_SET_PARAM_CMDID (during initial connection) or
WMI_VDEV_PARAM_TX_PWRLIMIT (power value change during RRM
req/bcn processing from connected AP) with the help of a
service bit WMI_SERVICE_EXT_TPC_REG_SUPPORT.
Change-Id: Iea1ecb62841c539c13a804ae9c475831b2d4c89c
CRs-Fixed: 2874250
The host will send TPC power to FW via the new WMI cmd
WMI_VDEV_SET_TPC_POWER_CMDID going forward instead of
WMI_VDEV_PARAM_TX_PWRLIMIT to accommodate 6GHz channels.
But, to support backward compatibility for New Host+old FW
the TPC power will be sent to FW via legacy WMI cmds
WMI_VDEV_SET_PARAM_CMDID (during initial connection) or
WMI_VDEV_PARAM_TX_PWRLIMIT (power value change during RRM
req/bcn processing from connected AP) with the help of a
service bit WMI_SERVICE_EXT_TPC_REG_SUPPORT.
Change-Id: I22e11ff847cd384fac8ea0061e5df42cea0c8de8
CRs-Fixed: 2874245
The host will send TPC power to FW via the new WMI cmd
WMI_VDEV_SET_TPC_POWER_CMDID going forward instead of
WMI_VDEV_PARAM_TX_PWRLIMIT to accommodate 6GHz channels.
But, to support backward compatibility for New Host+old FW
the TPC power will be sent to FW via legacy WMI cmds
WMI_VDEV_SET_PARAM_CMDID (during initial connection) or
WMI_VDEV_PARAM_TX_PWRLIMIT (power value change during RRM
req/bcn processing from connected AP) with the help of a
service bit WMI_SERVICE_EXT_TPC_REG_SUPPORT.
Change-Id: I727fd5a055e4e400ebc174cfb504d60e3335c3c4
CRs-Fixed: 2872244
Use tsft timer time stamp in HTT STATS IND header for
management TLV of CBF record since pktlog parser use
tsf time stamp in the ptklog record as packet receive
and transmit time.
Change-Id: I9643df11c0f9aa158f7b94e4b2083531044bcf66
In client roaming scenarios, when the ast entry is
updated, make sure the ast entry is updated with
new vdev id.
Change-Id: I0bd5235eeb28909cabe48f577d4477789d7be9e8
Currently, reg rules are processed in the order of LPI, SP, VLP. The
order from halphy is SP, LPI, VLP. Update the processing order so
correct reg rules are passed north.
Change-Id: I5bb22a1dcd3f1c95d8f4498b03d6cc6eedcfe44c
CRs-fixed: 2874752
The current value of scan timeout value is 10 secs.
Earlier for ap platform the scan timeout wass set to default as 50 secs
and and also for ap platform the scan could be triggered for internally
triggered connect and the max scan timeout
of 50 secs should not result in supplicant timeouts.
So setting the max scan timeout value of 50 sec for ap platforms
CRs-Fixed: 2870782
Change-Id: Icd2841f1f5503623f15bdbdea3430ef6559b9a1b
This is a cleanup fix to remove duplicate functions and remove unused
debug logs.
Change-Id: I0ebd203f0d19ad52cf97dcfe32914df59d1ea102
CRs-Fixed: 2863865
Currently driver is using sap notify api of regulatory to
update channel list for disable channel list command
which is not the correct way as sap notify api is designed
to be used by sap releted operations.
To address above issue write a separate api for
disable channel list command.
Change-Id: I79ded748bd46ec4be431c1ff6ab64c6ac8bf46be
CRs-Fixed: 2875324
1) Map Chile to FCC2_6G (Hex Code 0x05) for 6G, and to FCC14_World for
2G/5G.
2) Map San Marino to ETSI1 for WIN and DSRC, and to ETSI13_World for MCC.
3) Map UAE to ETSI13_World for WIN and MCC.
4) Add hexcodes for 6G sub domains.
Change-Id: I1627f2e739ad7b862e684673c36de7a823dcb714
CRs-Fixed: 2857528
sequence number populated from tlv are incorrect for some frames.
Firmware populate correct sequence number for management and control
frame either in payload or tlv itself.
added field to get valid sequence number.
Change-Id: I21b1c34c0d66cb46c3a0baaaa231c952de065534
In case of handling connect scan response, if cm_req is not found for
scan id, cm_req will be NULL and is getting deref at
cm_send_connect_start_fail.
Add fix to avoid NULL pointer dereference
Change-Id: I6b30a4bc41d3b0568dd307c91a7a7202d92b9c87
CRs-Fixed: 2877829
added new WDI event index for tx capture, share same index cause
corruption when switching different feature without delay.
Change-Id: I9e7a849968544957491891629eb42974c8b65728
Add delayed SRNG register writes support for Tx Ring, also add
dedicated workqueue to do the delayed Tx SRNG register writes.
Change-Id: I8dd157d341f3035e988804eab50d1ca681ab789b
CRs-Fixed: 2868989
