Add SAR unsolicited timer to the driver. This timer starts
on every data tx (if not already running) and stops on
command QCA_NL80211_VENDOR_SUBCMD_SET_SAR_LIMITS.
If this timer expires, a vendor event
QCA_NL80211_VENDOR_SUBCMD_GET_SAR_LIMITS_EVENT is sent to the
user space to issue the QCA_NL80211_VENDOR_SUBCMD_SET_SAR_LIMITS
command.
Change-Id: Ic30e5ac606d2b6b0cbc9209174adcd5213eac08e
CRs-Fixed: 2615510
Add SAR safety timer to the driver. This timer starts
when first QCA_NL80211_VENDOR_SUBCMD_SET_SAR_LIMITS
vendor command is recived. This timer gets
reset on every sar set vendor command.
When this timer expires, configure sar safety index to
the FW and restart the timer.
Change-Id: I0536ab88baf962e1bc2aef39478380534e0e1401
CRs-Fixed: 2615442
As part of new requirement add ini configurations
gSarsafetyTimeout, gSarSafetyUnsolicitedTimeout,
gSarSafetyReqResTimeout, gSarSafetyReqResRetry,
gSarSafetyIndex, gSarSafetySleepIndex,
gEnableSarSafety, gConfigSarSafetySleepIndex.
Change-Id: I18df23896b52aa1bbf8e071e27703d4e9e1738f1
CRs-Fixed: 2612843
As per the logic implemented in the FW, it will send a dummy event for
link layer stats in case the buffer allocation fails. In this case,
there can occur a scenario in host wherein we are not properly freeing
the previously allocated memory.
In the event handler wma_unified_link_radio_stats_event_handler, if the
event is dummy, we are updating the value of the num_radios inside
wma_handle->link_layer_results. This num_radios is later used to free
any allocated memory. On setting it to zero, we are preventing that
memfree to occur; leading to a memleak.
To fix this, in case there is a dummy event, first free any possible
previous results before setting the num_radios to be zero.
Change-Id: Ibe4ddf1a137a33130100483cefb004b7dae4ce02
CRs-Fixed: 2610754
The INI param gDot11Mode is used to indicate the support that the driver
should advertise. Currently the behavior of the driver is out of sync
with the mode supported over the air and the mode provided to the
framework.
In one particular case, the driver sends VHT supported as
true to the framework even though the INI gDot11Mode is set to 11n only.
This results in the frames to be transmitted with 11n capabilities while
the mode advertised is different.
To rectify this, check for the dot11 mode before setting the
corresponding supported bit in the wiphy. The value being advertised
should be the intersection of gDot11Mode INI and the driver and
firmware capability.
Change-Id: Ic2f1558c72fc801df9d7ca6b3a0182619b93cf0d
CRs-Fixed: 2612620
Add support to customize DSCP-to-UP map table and send the
customized map values to FW to update its corresponding
map table.
Change-Id: Ibe9704a90468c898dd2e60fdf83a271152f654ce
CRs-Fixed: 2616247
Roaming logs are logging to kmsg, so unused diag log
sending API can be removed.
Remove unused roaming diag logs.
Change-Id: Idd5cfd2ad1f63394cb6b1872512bcc17307a78e6
CRs-Fixed: 2617079
Add following changes to refine csr_dump_connection_stats:
1. Remove unused argument and rename the function to reduce the
number of characters used.
2. Add self mac address information.
3. Rename Auth type to AKM. Print the AKM string instead of
existing diag auth type enum as it doesn't include new WPA3 and
FILS related AKM.
4. Print channel and channel bandwidth on the same line.
New format of log:
csr_connect_info: 3692: +---------CONNECTION INFO START------------+
csr_connect_info: 3692: VDEV-ID: 0 self_mac:18:23:73:36:23:0f
csr_connect_info: 3692: ssid: FILS_AP
csr_connect_info: 3692: bssid: c8:b3:73:36:a3:0f
csr_connect_info: 3692: rssi: -25 dBm
csr_connect_info: 3693: channel: 6 channel_bw: BW_20MHZ
csr_connect_info: 3695: dot11Mode: DOT11_MODE_11N
csr_connect_info: 3700: AKM: FILS-SHA384
csr_connect_info: 3701: Encry-type: ENC_MODE_OPEN
csr_connect_info: 3702: channel frequency: 5745
csr_connect_info: 3703: Qos enable: 1 | Associated: yes
csr_connect_info: 3704: +---------CONNECTION INFO END------------+
Change-Id: Ia87357dd607bfa0d9256c8129bf8e4448ae93104
CRs-Fixed: 2614465
Wext handlers are not getting deregistered during the
interface down of monitor mode. So. After the idle shutdown
also the user can issue ioctls which can lead to system instability.
Fix is to remove the wext handler as part of adapter deinit.
change-ID: I49ab4b0cfa0125b253bc097f180226f2095d5df1
CRs-Fixed: 2615138
Currently there are unnecessary/duplicate logging. Identify and
reduce those logs that are not necessary.
Change-Id: I5ec0a7e06cdade73f0e99369ad89b0b677a9a96b
CRs-Fixed: 2615407
Currently the driver sets rmf capability and encrypts
management frames only if peer advertises MFP required and the
sta is also MFP capable. The SAP marks the vdev as PMF enabled
if SAP is MFP capable. This will result in back to back connection
failures or STA performing SA query procedures after disconnecting
and connecting to the AP again.
Check for the AP MFP capability and STA MFP capability to enable
PMF.
Change-Id: Ie851f8dc6d08c46d5eec36c3eff34d322d37fac3
CRs-Fixed: 2612342
Enhance logs in ACS module to remove duplicate
prints and keep only useful logs to debug.
Change-Id: I2cf98fb8b4d7a8391617c9ce1baaa5ff6ad42114
CRs-Fixed: 2614609
In case of successful wow enable ack from FW for system
suspend call cdp wow ack response process, so that dp
layer will take care of required operations before system
goes to suspend.
Change-Id: Ibc5bf8d62aa055c9c8c7242c00abc3b696bb0051
CRs-Fixed: 2608980
Currently, If STA doesn't send beacon report response to
AP, AP will not be able to know the reason why sta did not
send beacon report as no reason code is defined in
IEEE802.11-2016 standard.
Fix is to add Optional IE in Beacon Report Response which will be
populated based on err code and beacon_report_error_vsie ini.
Change-Id: I91ab0cdfadbe903bdc3085a94b38938a35b693e0
CRs-Fixed: 2614385
Currently, the driver schedules the message to set
power save enable/disable to FW as part of the
set power mgmt command received from the userspace.
This command is then scheduled and is sent via
the scheduler thread.
Generally, userspace sets the power save disable
when it starts the DHCP process and enables the
power save back when the DHCP is completed.
DHCP packets are sent through the Datapath and the
PS enable/disable command would be sent via the
control path explained above.
the Race could happen that the scheduler was busy
with some other task and the PS disable command
was in the queue and the DHCP process began.
This would result in DHCP packets going to the peer
with PS enabled.
The fix is to set the power save enable/disable directly
instead of using the scheduler path.
Change-Id: I0f2aed37f875c283f318fb44bcc40d0ab401413a
CRs-Fixed: 2611480
