| Commit message (Collapse) | Author | Age |
|---|
| ... | |
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Aggregate QOS request is not implemented in qcacld. This is causing
host driver to wait and defer other FW messages whenever this message
was sent to firmware, thus causing a hang.
ROME firmware supports ADD_TS per AC (WDA_ADD_TS_REQ), so for
WDA_AGGR_QOS_REQ, multiple ADD_TS requests will be send and a dummy
response is sent back to LIM along the line of existing implementation
for WDA_ADD_TS_REQ.
CRs-Fixed: 582083
Change-Id: I0b842a6535a3083c90fec59b2444b3c24a787e38
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Update fw common interface files per CL 787026
Change-Id: I695b4e142c10d46e18ff2fd5328253458c0dfd6e
CRs-Fixed: 582536
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
When beacon tx offload is enabled, the firmware is going to add
the p2p ie and noa ie in the beacon. Host has to just send the
WMI_P2P_GO_SET_BEACON_IE command to firmware for updating the
p2p ie.
Change-Id: Iab1c4a5d1f3d66c4ba0c3198a9135246c08e1402
CRs-Fixed: 582525
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
This change provide the capability to offload the beacon
transmission into firmware. This requires SU49 or above to
work. The beacon is sent to the target by host once. The
firmware takes care of sending the beacons without host
intervention. The host can change the template anytime as
and when needed.
Change-Id: Ia5b4929ffec5291f4fce3e8374d238bc5a097f1a
CRs-fixed: 580855
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Spending too much time in processing Rx frames
in tasklet makes system unresponsive and hang,
the reason for the hang is that the CPU is not
able to scale up due to the higher priority
task (wlan_tasklet) which is hogging the CPU.
To come out of this condition, process 1000
rx packets in a single tasklet run and schedule
the tasklet to process the rest. This would
allow the process which takes care of CPU scaling
to run and increase the CPU frequency.
Change-Id: I4d1e302ccb52bdf268879ead9b2493f8c4bda99b
CRs-Fixed: 566643
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Update fw interface common files as per CL 786393
Change-Id: I1ccf74fec5c6b25db5600b837410300b7090df04
CRs-Fixed: 582395
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
ASPM is disabled till firmware and otp are downloaded but is not enabled
after that. This prevents the target to dynamically manage power when
the link becomes less active over time. Re-enable ASPM after the
firmware/OTP are downloaded and once the target is ready.
Change-Id: Ie3f1a1a655053ce54e21d338715b84e13eda83f6
CRs-Fixed: 582045
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
BD channel index is distributed in different layers for various platforms
(Discrete/Integrated)which in turn reflect with changes in TL layer.
To avoid changing TL to support multiple platforms, the logic mapping is
moved down to wireless device interfaces "(WDA in case of integrated
solution & WMA in case of discrete solution)in the WLAN driver.
Change-Id: Ib6896a31fd0c517969170f688c3313fc642952cb
CRs-fixed: 557229
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Update f/w interface common files as per CL 784827
Change-Id: Ib95b5189a6d37eed822648d266edfc089c240689
CRs-Fixed: 581440
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
FTM requires iwpriv command to be used for FTM start and stop operations.
Do this during driver load and unload to avoid using the iwpriv commands.
In the case of FTM, MAC is not required and so do not start it. Also,
skip sending suspend target WMI command as the FTM mode firmware doesn't
support it.
Change-Id: Icfadd3a22eb9d6af919c88a2885f121c1f7a823d
CRs-Fixed: 581303
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
WMA has to work around the fact that umac sequence of WMA
commands is different during reassoc. It does not send SET_CHANNEL.
Hence WMA has to use ADD_BSS command to perform VDEV_START and
defer the bss and sta configurations for later. Hence it saves
values of 11bCoexist, beaconPeriod, etc. uses them later
after association.
Root cause is that these parameters are saved only
when reassocReq flag is true, but they are used unconditionally
later even for first time association. Hence 11bCoexist
is set to 0 in firmware and causes this failure.
Fix is to move the parameter saving code to before check
for add_bss->reassocReq. Save them unconditionally.
Change-Id: I9a6e39d9565da864ad5eade8138702488d834c1f
CRs-Fixed: 576251
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Firmware now translates auth and encryption types from rsn_ie and
wpa_ie of beacons into WMI types for comparison in roaming.
Modified wma routines to send same WMI auth and encryption types
to firmware. Added separate WAPI_PSK, WPA_PSK and RSNA_PSK types.
Change-Id: I13d3b4d0e489ce9ad374c98a163d6a959011c6cb
CRs-Fixed: 578195
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
below are the valid pwrgating value's
0 to disable power gating mode and
1 enable power gating mode.
WMI_PDEV_PARAM_POWER_GATING_SLEEP is used from host to
enable/disable power gating mode in fw.
Change-Id: I66bae8cdf57f73b7242919c4f23e4e439368641e
CRs-fixed: 573601
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Session ID is not extracted properly in eWNI_SME_SCAN_ABORT_IND msg
handling. Due to this, incorrect session id was passed to WMA and
was causing abort scan request to the firmware with incorrect scan id.
Also fixes incorrect logging level in hdd_conf_arp_offload().
Change-Id: I297ce2ad34a51427688cce7545223d16e9433414
CRs-Fixed: 580904
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
This patch will merge the following CL's 782830,784178
784337,782678,782707,782709
Change-Id: I8bd0df2dca63de64ed90564c7eaeaf6ab6818142
CRs-Fixed: 580850
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Removed rssi_thresh and other commands sent to roam offload
engine before sending command to stop roam scan.
Change-Id: Ie7cf7bd10f400faf6c710191c5c92a204cda3a7f
CRs-Fixed: 571099
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
This change is for chip level reset (SSR phase 1). When target crashes,
host driver will collect the RAM dump by dumping them to a dedicated
memory location.
Change-Id: I458b50913057ee80158d82d228396b27c5b7e7db
CRs-Fixed: 562329
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Update fwcomon interface files as per CL 780680 and 781419
Change-Id: I3fed688e365d81d588e0dac877727abb879e4f23
CRs-Fixed: 578887
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Some of the scanning and roaming parameters like rssi
thresholds and dwell times are now configurable through
WCNSS_qcom_cfg.ini. Roaming behaviour can be tuned using
these parameters.
The parameters are subject to min and max limits imposed
by LFR2.0 for Pronto.
CRs-Fixed: 576697
Change-Id: I7b2e04e6f3bf8fe292016a506e4a0dc60dd81bfa
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
New entry "gEnableWoW" added in .ini file to control WOW enable
and disable status. By default, WOW is enabled (magic and pattern
byte match are enabled) on this entry.
While configuring wow patterns to the firmware, first preference
is given to the user configured WOW patterns. If the user doesn't
configure any patterns, default wow patterns will be configured
based on interface current mode (STA/AP) while going to suspend.
If the user configured some wow pattern, then default wow patterns
are skipped and only user configured wow patterns will send to
the firmware.
If interface is in AP mode (SAP/P2P GO) and no user wow patterns
available, WMA would configure the below default patterns on that
interface,
+ Unicast IP, EAPOL-like and ARP packet pattern
+ ARP packet pattern
+ mDNS/SSDP/LLMNR pattern
+ DHCP broadcast pattern
If interface is in STA mode (STA/P2P STA) and no user wow pattern
available for that interface, WMA would configure below default
patterns,
+ Unicast packet pattern
+ mDNS/SSDP/LLMNR pattern
Change-Id: I09b7093d2417021919c63207eea31101c4da386b
CRs-Fixed: 540571
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Fix the issue of DUT SAP with hidden ssid enabled is not
sending the probe response frames to the reference STA.
SSID and its length are updated in SAP configuration for all
ignore_broadcast_ssid cases in wlan_hdd_cfg80211_start_bss.
So that the sessionentry has the SAP's SSID and the probe
response will be sent to SSID specific probe requests.
Change-Id: Icd562c021164df9f8a1273e2083cca5dd767e3ae
CRs-Fixed: 557250
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
PCIE SOC Wake needs to be set so that target will go sleep or low
power mode. But before loading the firmware target clocks won't be
set properly. In this case host cannot wakeup target properly.
So till the driver is loaded PCIE wake will be set to awake.
Once the driver is loaded PCIE wake will be set to sleep mode.
With this max perf mode i.e keeping the target always awake
is disabled as well.
Change-Id: Idb5b88e8d5d68cf4c1e226a673cfac53572ceb01
CRs-fixed: 548245
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
The earlier commit "cld: Stop data transmission prior to disconnect
request" fixed system hang issue when the user is initiating
disconnect. But it has to be fixed when AP is initiating disconnect
and the station is pushing uplink traffic heavily. Once SME receives
deauth/disassoc indication, it is proceeding to roam or post SME
command to disconnect without stopping data transmission. Under
heavy uplink udp traffic, due to the latency in MC thread processing,
all this data is causing the AP to repeatedly exchanging deauth
frame with reason code 7 (Class 3 frame received from nonassociated
station) and causing system hang.
Change-Id: I6e0167cf393bda5775715d5ef5c11f7e07ad45e4
CRs-Fixed: 573721
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | | |
Revision 4.0.0.53
|
| | | | | | | |
|
| | | | | | | |
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
On receiving Disconnect command from CFG, change the connection state to
Disconnecting rather than changing it to NotConnected. Hence processing
of Connect command even before Disconnect has completed is avoided.
Change-Id: Id3a39e33f4a6fdeb052cddcbace2aa2959f15987
CRs-fixed: 550307
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Session ID and user priority is required to delete
traffic specification added in firmware for discrete solution.
Change-Id: Id6e4737fc0e6e8a1387c48ce78eb6219e80ce9ca
CRs Fixed : 533518
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
With the existing implementation, beacon report from STA reports
details pertaining to only one BSS per channel and this behaviour
is not meeting the VE cert test plan requirements. Hence support
is added to send multiple measurement reports corresponding to
multiple BSSs per channel as the STA identifies via scan.
Change-Id: I2e277157e2529134998ba87a208bbad848838cb4
CRs-fixed: 564929
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
During module load/unload stress, it is possible
for isr to access copy engine related memory which
could have freed much before in HTCStop() and
HTCDestroy() which are called from vos_close().
Fix this freed memory access, make sure to stop
the tasklet and disable the interrupt before freeing
the memory which will be accessed in the respective
handlers.
Change-Id: I382c342ed6b410936fa645043bedfee4864d8f3b
CRs-Fixed: 568737
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Update fwcommon interface files as per CL 779227 and 779401
Change-Id: Ie852769781e221a216e8ec4ea35e18f452e211ab
CRs-Fixed: 576475
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Update fwcommon interface files for WMI_P2P_NOA_EVENTID
Change-Id: I03355f1206312952150d2397f790e00a7c4d4bbf
CRs-Fixed: 575474
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Adding new ini variable to configure the maximum number of wow
filters required.
Also add and delete wow pattern functions are modified to use
the value configured through the ini.
Change-Id: Ib6d8cccd51c97d4fe5fd5348fe7b4132e37805b7
CRs-fixed: 576658
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Removed duplicate code in limProcessMlmFTReassocReq was causing peer to
be created twice which was causing firmware buffer for peer table overflow
and hence not properly update peer ID's for peer mac address.
This was bug only in qcacld-2.0 hence no fix is needed in qcacld-new
Change-Id: I1843ccb6730c7e26ae4a2ca5bd61d5bebc15de90
CRs-Fixed: 569771
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
During split scan in standalone or in concurrency cases the time
taken to complete a scan request takes more than 5 sec. This leads
to scan completion with time out reason that purge the scan results.
Hence increasing the maximum scan duration to 30 sec. that consider
all split scan cases and also matches with cfg80211's bss expiry
period.
Change-Id: I358dd53e4f36a1d9dbf733d4bb7c64ad05a48498
CRs-Fixed: 572008
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Merge the dbglog common file changes from
perforce.
Change-Id: I4568c4ad6b4031f8c4f907506add7104653658a6
CRs-fixed: 574325
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
This patch will add dbglog support for
SAP mode.
Change-Id: Ia6e739ce995c52e486cbd57edb1847a00b5c3c58
CRs-fixed: 575528
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Pass the MAX and REG power values to the FW, so that FW uses the
appropriate power value for Tx packets.
Change-Id: I7f1ebd0ab70bc8f32b67a6305d3e1b2190055cec
CRs-Fixed: 574133
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Required to fix the beacon allocation for softAP in 11a mode.
The beacon should still be allocated if the tail == NULL or
tail_len == 0.
This gerrit is for qcacld-2.0
Change-Id: I2eed96d7fb6aaa94a70e752618bbf9c53498c933
CRs-fixed: 574425
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
This patch will merge the dbglog changes.
Change-Id: I8968e87484135759e15d4f08d4e34414de0ce046
CRs-fixed: 574325
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
In Rome Host supports the following powersave modes.
a) Legacy Powersave with Deep Sleep Mode disabled
b) QPower with Deep Sleep Mode disabled
c) Legacy Powersave with Deep Sleep Mode enabled
d) QPower with Deep Sleep Mode enabled.
Deep Sleep Mode is not yet supported.
QPower is based on dynamic txrxinactivity as well as
speculative PsPoll. So in QPower Ps Poll is always
configured.
Default value of QPower Ps Poll Count is 1 unless otherwise
specified in ini file. Default Tx Wake Threshold is 2 i.e
max tx packet before coming out powersave.
QPower works strictly on DTIM so configure Listen Interval
will be overridden.
The value of gEnablePowerSaveOffload will be passed onto WMA
as mac parameters while opening WMA.
WMA use this to enable QPower appropriately upon enabling
station mode powersave.
Change-Id: I91a5811944674e7e7b3a6bc22887267175b28dc5
CRs-Fixed: 570158
|
| | |\ \ \ \ \ |
|
| | |/ / / / /
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Remove function calls from AOSP code to utils.mk
& replace them with actual definition.
Change-Id: I8b75a96614d16af50efcc7da3c853d0614eb17cc
CRs-fixed: 566287
|
| | |\ \ \ \ \ |
|
| | |/| | | | |
| | |/ / / / |
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | | |
Revision 4.0.0.52
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Update fwcommon interface files as per wmi version 17 on f/w side.
Change-Id: I849e51af00e9a024a050e7bce0ca7d4c1b1e888a
CRs-Fixed: 574022
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
cfg80211 layer (kernel version 3.10 and above) issues disconnect on
connected interfaces during platform suspend if wow is not enabled
from the user space via nl80211.
CLD takes wow parameters (enable/disable, patterns,etc) via iwpriv.
So, cfg80211 layer never knows wow setting done via iwpriv command.
If cfg80211 layer issues disconnects, CLD can not put the chip
into wow mode. To stop cfg80211 issuing disconnect, this patch
exposes wow capability to cfg80211 and let the user to enable
wow via nl80211.
Change-Id: Ib25ea5867a7500351ee94fdc9d8c92f6ee0e4b33
CRs-Fixed: 540571
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Once WMA completes all configuration needed to keep the target
in suspend state, ready to suspend indication is posted to UMAC.
HDD would unblock cfg80211 suspend after getting this event.
And also takes care of changing hdd_suspend_wlan() func prototype
in if_pci.c to align recently changed one.
Change-Id: I0f2b9aac1dc3ee5ff2ce914b2cdf540fe5325b77
CRs-Fixed: 540571
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Suspend and resume configuration (say, all offload configuration) are
handled in SETSUSPENDMODE ioctl in HDD. Whenever HDD gets this ioctl,
it doesn't mean that platform is going to suspend/resume state. It's
just indication that display is turned on/off.
Discrete solution expects suspend/resume configuration to be happen only
when the platform goes actual suspend/resume. This patch moves suspend
and resume configuration from SETSUSPENDMODE to cfg80211 suspend/resume.
New event called ready to suspend is added in this patch. This event is
posted from WMA when it completes all configuration in target which are
needed for suspend. HDD blocks cfg80211 suspend until it gets ready to
suspend indication from WMA.
Change-Id: If642743912df23049ca8cf37bb78d6c1fb21db5f
CRs-Fixed: 540571
|
