If 11ac override and 2.4ghz channel bonding is enabled in INI, for
2.4 ghz channel the channel width is always set to HT40 even if
hostapd indicate channel width as HT20 due to OBSS scan.
Thus the channel width indicated by OBSS scan is ignored.
As a fix, even if 11ac override and 2.4ghz channel bonding is enabled
for 2.4 Ghz channel give preference to the width indicated by
hostapd.
Change-Id: I091c8672b664a4f24c976ef33816ab10f32a052e
CRs-Fixed: 2032597
Regulatory is currently checking if either no-HT_40+ or no-HT_40- flags
set. If one of them is set then it claims that 40Mhz BW for that channel
is not supported which creates issue for SAP where it considers channel
36 (5180Mhz) is only 20Mhz capable.
To fix the situation, check both no-HT40+ and no-HT40- flags before
claiming 40Mhz/80Mhz not supported.
CRs-Fixed: 2042420
Change-Id: I8c73ae8df21efa23f2967f17febe165cff55c43c
Firmware crash due to scan received on
IBSS vdev and non DBS mode.
Fix is to reject scan when in ibss mode.
Change-Id: I2ca126ded013beb658b0bb82d854b0a3acccba5a
CRs-Fixed: 2038271
Move the logs to appropriate log levels to reduce
the log spam in lim_update_overlap_sta_param
Change-Id: I2325a4c5e3b3429ff46bc329371711576398e919
CRs-Fixed: 2039869
In RRM module, caller is creating memory for beacon report. for
each beacon report, it further creates memory for bss descriptors.
caller sends a entire becon report message to callee. Both caller and
callee are releasing memory for beacon report but they don't release
memory for bss descriptos correctly.
To fix the situation, caller needs to release bss descriptos memory up
on failure to deliver the message to callee. callee needs to release
bss descriptors' memory up on successful reception and after processing
it.
CRs-Fixed: 2040435
Change-Id: Ia4e0a1cdc56c177e86683847973acfab0e9a96bf
While sending message from one module to another module,
bss descriptors' memory get assigned but never get released which
creates memory leaks.
Release bss descriptors' memory up on its usage.
CRs-Fixed: 2040435
Change-Id: I7d9ef4e9b65f4b63251490f4ef3671a350ebf81f
qcacld-2.0 to qcacld-3.0 propagation
MAC layer counters such as per channel signal quality,
per peer TX/RX statistics, are monitored by FW.
FW reports these counters periodically as host setting.
This change:
- Add one parameter to enable/disable mac counter
- Convert FW indication to NL80211 attributes and send the
same to user layer
- Convert user layer's NL80211 attributes to parameters for
link layer monitor and send the same to WMA
Change-Id: I17a81fcec96076e14ff9f33d0aeb0983a25de912
CRs-fixed: 1071059
Remove flooding log from hdd_ipa_nbuf_cb.
Also fix INVALID REASON log from hdd_ipa_uc_op_cb even when
hdd_ipa_uc_op_metering() is succeeded.
Change-Id: Ided529cca3d84cbc3dd6b33dd31fa9086489a176
CRs-Fixed: 1116111
vdev and pool could be NULL during processing the forwarding packets
processed during the driver deinit.
Add a code to return false if pool is NULL in ol_txrx_fwd_desc_thresh_
check(), then the caller will discard the packet without any further
processing.
Also return false when vdev is NULL.
Change-Id: I45afcbe0e8b953bd1be7b3f1f5315f35879edec5
CRs-Fixed: 1112619
Currently, Host configures ITO/TX_Wake_Threshold on each
suspend and resume along LI/DTIM setting. But in a case
where LI/DTIM is configured by user then host skip to
change LI/DTIM values. Hence host skip ITO/TX_Wake_Threshold
configuration too.
As a part of the fix, Decouple ITO/TX_Wake_Threshold from LI/DTIM
configuration.
Change-Id: Ie1c8b7821dcba89d888e18dd7726958a9381e038
CRs-Fixed: 2038986
Present wlansap_close API walks out without releasing the sapctx when
hal context is null which creates the memory leak.
Release sapctx in all cases when caller requests to close the
sap session.
CRs-Fixed: 2040045
Change-Id: I3c8f0a33c11de9d88d6b29f726f36a838f2428d2
Current debug code incorrectly releases parent skb multiple times
while traversing the skb_list. Fix the bug by releasing individual
skbs in a list.
Change-Id: I7047dfc07407275aa7f06cb22840023331ec186f
CRs-Fixed: 2040538
Add TSO segment debug code, in a featurized way, so that
when disabled it will tolerate redundant returns to the internal
pool and when enabled will check sanity on free, and record a
history of last 16 actions on the segment in the object.
Goes with the corresponding cmn component counterpart.
Change-Id: Ifab52fc2032da4d53b708b3c6716d5270c0690c1
CRs-Fixed: 2031485
Limit the detection of non posted rx buffers from FW only to
64 bit DMA address system. For systems, where IPA has 32 bit
DMA address limitation, this debug feature should be disabled.
CRs-Fixed: 2010583
Change-Id: I9a26d0a423acedb068d99f39929029182f7c6fa1
P2P-GO whenever goes offchannel, issues NOA. With this behavior, we
have issue whenever legacy STA connects to P2P-GO as legacy STA
does not understand NOA. To handle this case, changes are done in HDD
and SAP component and the revised behavior will be:
1. By default no change in NOA.
2. Whenever first legacy STA connects to GO, SAP informs HDD with
new parameter station type. HDD will communicate the same to
p2p protocol component. Once the connected station is authorized
p2p protocol component will disable NOA (i.e. P2P-GO will start
issuing Self-CTS whenever it goes off-channel).
3. Whenever last legacy STA disconnects from GO, p2p protocol
component enables NOA (i.e. P2P-GO will start issuing NOA whenever
it goes off-channel).
Change-Id: Iadab47c9a2aeb434371d4c262623d7196f3b1ab4
CRs-Fixed: 2035609
Fix memory leaks of TSO num segments by freeing remaining
TSO num segments in ol_tx_ll_fast and ol_tx_ll if somehow
allocated tx desc is NULL.
Change-Id: I577a95b784c776633cf685cd16b101755a48c521
CRs-Fixed: 2014587
TX BUF POOL vaddr storage alloc is not freed in
htt_tx_ipa_uc_attach when paddr storage allocation fails.
Fix is to free memory in failure case.
CRs-Fixed: 2034943
Change-Id: Ifb26dfdcff19b3ab66f11a89b9c185777eeec535
In csr_scan_copy_request_valid_channels_only(), the DFS channels
are ignored if ini parameter gSkipDfsChannelInP2pSearch is enabled.
If the supplicant issues a single channel scan for a DFS channel,
driver will do a full scan. As its a single channel scan which is a
DFS channel would be ignored and the num of channels to scan would
be zero.
And if the number of channels to scan are zero, driver
does a full scan which would impact the overall connection
time.
To address this, enable DFS channel for a single
channel scan.
Change-Id: I166b9674a6cb6bb9c180c1c8d68480440ea867be
CRs-Fixed: 2038974
