The IOVA for the buffers that are attached to the direct
link receive copy engine need to be contiguous for optimal
memory mapping on ADSP.
Fix is to add multi pages prealloc support for HIF layer
and add a pool entry to multi pages prealloc for direct
link receive copy engine.
Change-Id: Ieb253bd3c1b6550e4c1c63cd587993891ac817f2
CRs-Fixed: 3497689
Log the AP link_id as part of the connect information log which helps
in debugging.
Change-Id: I5244e6c1d2cec905005515078839c0feda721a29
CRs-Fixed: 3508511
Upon the completion of roam sync indication event, the
roam sync complete cmd is first sent to the firmware.
This is followed by the netdev queues enable. This
approach adds latency in the first downstream tx frame,
and the FT roaming KPI is not met due to this latency.
To fix this, if roam sync handling is successful, enable
the queues and then send roam sync complete to the firmware.
It's observed that the current change improves roaming
latency by 6 msec.
Change-Id: Ib05ca01892b6d8b62badcafb64bb8c16270ee9f8
CRs-Fixed: 3483751
After ML STA link state changes, the connection table will
be updated. So defer the force SCC a few second to use
updated table to do force SCC.
Change-Id: I453b1a8d2707bb635f7c1d8d5b397811bf3b4f7b
CRs-Fixed: 3502411
Currently if AP+AP SCC on 5 GHz, next AP prefer on 2.4 GHz,
If SBS mode support, next AP can prefer also on SBS channels
and other MCC/SCC channels.
Change-Id: I3e44fd381527cae24aeff14f9158621f14a278b3
CRs-Fixed: 3499268
Currently roam pcl is updated after roam, connection and CSA.
Pdev current channel list is updated for many other scenarios,
but roam pcl is not modified accordingly.
To address this issue, update roam pcl every time after computing
pdev current channel list.
Change-Id: Ie86d5592255419de8f4040a3206d275ef926cf03
CRs-Fixed: 3481026
Currently host disables 6 GHz band when set_fcc_channel cmd
is received hence existing STA/P2P client 6 GHz connections
get disabled.
Add changes to keep existing STA/P2P client 6 GHz connections
intact and also keep SAP up on 6 GHz if STA+SAP SCC on 6 GHz
channel was present before receiving this command. Disable all
other 6 GHz operations.
Change-Id: I95c094e1b344d430a288b6f613842efe6e576e25
CRs-Fixed: 3495006
Currently the CDP handles like soc/hif/hal_soc are not
available in any of the the DP component contexts.
Modules like FISA etc need these handles.
Store the CDP handles like soc/hif_handle/hal_soc in
the DP component context.
Change-Id: Iaee4d0e15cbaa7d871dc9e246d177ad310b4c6ba
CRs-Fixed: 3505720
When SAP is on 5 GHz or 6 GHz channel other than the user
configured frequency due to concurrency and concurrent STA has
disconnected, then SAP doesn't move back to the user configured
frequency when MCC to SCC Switch mode is 4
(QDF_MCC_TO_SCC_SWITCH_WITH_FAVORITE_CHANNEL). SAP has to move
to the user configured frequency irrespective of the current
operating band.
Move the SAP to user configured frequency when SAP is standalone
after STA disconnection.
Change-Id: I3fb78cbd949c0e2ec7c393208569956dbcdd067a
CRs-Fixed: 3509288
Chan width 1 in vht operation means >=80M, while chan width 1 in eht op or
he op means 40M, need be handled in different way.
Change-Id: Ib65dca6245c33f07df351b6ac7eed86f91b372c7
CRs-Fixed: 3506080
Currently the txrx_ops are not saved in the
non-converged datapath.
Store all the txrx_ops in DP component, when
they are registered with the vdev. This can be
used in the DP component to get and use these
ops.
Change-Id: I30bfc69a32e0a9f0ab16f902a94a8c28ef4af8a6
CRs-Fixed: 3505719
On reception of reassoc req from user space to the
connected AP, wlan driver proceeds with reassociation
without clearing the keys. This leads to encrypted EAPOL
frames in LFR2. To avoid this, trigger disconnect & delete
keys before proceeding with reassociation.
Change-Id: Ib287b5d43ae69a7696a28f0e59347c7376d581c3
CRs-Fixed: 3505422
Currently, eMLSR concurrency API considers both the links to
decide whether to allow or disallow the command while
force-enabling links. But one link might have got disabled to
handle some concurrency and might not be present in the
policy mgr table. Driver has to consider the links that are
present in policy mgr to allow/disallow link-enable command
when an eMLSR connection is present along with other concurrency.
Also, user might try to force-enable the disabled link. Allow
that only if eMLSR concurrency is not present.
Change-Id: I0a68059b7d16629016b3c40dbc0345523d46ef21
CRs-Fixed: 3500106
Add support to parse t2lm ie from association response frame
and store in mlo dev context. Once all links are up
the t2lm configuration will be applied.
CRs-Fixed: 3449605
Change-Id: Icd1f3c9bbdfbe1c2a8bfe119d92d6e686e21a7fd
Currently, to switch links from link A to link B, Host sends
two back-to-back WMI commands to FW, by setting link B force
active and link A force inactive respectively with force
mode MLO_LINK_FORCE_MODE_ACTIVE/MLO_LINK_FORCE_MODE_INACTIVE
but if association is in EMLSR mode, FW considers the first
WMI to set link B active (when link A is already set active)
to be a trigger to enter EMLSR mode and sends OMN frame as
part of it. To avoid this, Host has to send a single WMI
with force mode MLO_LINK_FORCE_MODE_ACTIVE_INACTIVE and
include both active and inactive vdev bitmaps within the
same command to clearly indicate the link switch.
Change-Id: Iae58fedc800c891625d737daf238fa476382f4a1
CRs-Fixed: 3487826
Currently, data rate for management tx packets is filled only as
1 or 6 Mbps which is wrong.
Sniffer expect data rate multiple of 2. Hence fill the data rate with
multiple of 2 for tx management packets which received from firmware
and converting into mbps.
Change-Id: I5dd6515804cd3751925af0992a30a7441ebf9bbb
CRs-Fixed: 3444177
Currently, mlo_roam_get_link_freq_from_mac_addr() returns 0 for
non-11BE platforms. Valid freq is present in roam_sync indication
and return the same.
Change-Id: I7156adadae9303257bf96e91e4cb8b5d11706104
CRs-Fixed: 3502610
Currently if hash-based routing fails, packets will be sent
with default routing which will result in packet being queued
to REO2SW1 with tlv reo_dest_indication = 1. Host will create fisa
flow with napi = 0 for REO2SW1.
When next packet received with correct tlv reo_dest_indication value
on different REO with flow_index_invalid = 1 and if tuple matches
with earlier created flow, it can result in REO mismatch issue.
To fix the issue if rx_hash is enabled add check for
tlv reo_dest_indication value.
Change-Id: Ib0895ce666cc788a6df8ae3483e0777e84937ef8
CRs-Fixed: 3500655
Currently OSIF calls the CDP API for DP suspend/resume
handler directly.
Move these calls to DP component so that DP component
can handle the suspend/resume related operations for
internal DP sub-modules (which are outside CDP).
Change-Id: I080809904f619260c707cacbcdbf158250320081
CRs-Fixed: 3502507
Current hw mode is SBS low share. STA 5180, SAP 1 2412,
SAP 2 5745, but SAP 1 is 2 GHz only, can't move to STA 5180,
SAP 2 is SBS with STA, policy_mgr_are_2_freq_on_same_mac
return false for 5745 and 5180 and finally function
policy_mgr_is_restart_sap_required return false, no force
SCC on SAP 2.
Add mcc connection count check for SAP2, if SAP 2 channel
is different from all of existing 2 or more connections, then
try to force SCC on SAP 2.
Change-Id: I7029fd3ec0b8e0ae4543edd9667daa1f87802a7a
CRs-Fixed: 3497431
In wlan_cm_get_associated_ch_width, if vdev is NULL, which means
reference is not gotten. do not call wlan_objmgr_vdev_release_ref.
Otherwise there is QDF ASSERT.
Change-Id: Id1cc2be3869ad555ebeb1c91a4b5f75d499e70ac
CRs-Fixed: 3503877
When roaming from OWE SLO to OWE MLO, host considers it as SLO roaming,
although FW sends both assoc link and partner link information to host.
Currently host adds partner link to sta_ctx->wlan_connected_links in
mlo_roam_update_connected_links, since it already receives partner
link information from FW. Then RSO fails to translate to stop state in
cm_fw_roam_complete because partner link is not up.
To make RSO state translation work, host need add partner link to
sta_ctx->wlan_connected_links when partner link connection is
triggered, which is mlo_roam_prepare_and_send_link_connect_req.
Do not add partner link to sta_ctx->wlan_connected_links in
mlo_roam_update_connected_links.
When roaming from OWE MLO to OWE MLO, host triggers disconnection in
partner link, which should not enable roaming in connected STA.
Change-Id: I99938a0805d60287eb2714e5d271579ac2fb4aeb
CRs-Fixed: 3498822
Current SAP does not support MCC and for low latency SAP,
host driver needs to support MCC.
To support MCC with SAP, add new concurrency mode for low
latency low throughput SAP as PM_LL_LT_SAP_MODE.
CRs-Fixed: 3469425
Change-Id: Ifb4b2db3bf3c2b433324580fde32d1673df5ae67
During the conversion to the Bazel DDK build system the qcacld-3.0
driver was built without the FEATURE_WLAN_RA_FILTERING flag being
set. This resulted in two separate build errors due to incorrectly
testing using #if instead of #ifdef, so fix those tests.
Change-Id: I9265475c1d33918e8fac7fc138e0bfdc301e5eef
CRs-Fixed: 3497220
Currently the CDS layer calls the cdp functions
directly to initialize common DP.
Move these calls to DP component, so that DP
component can handle other DP related initialization
inside the DP component itself.
Change-Id: Ib257ac4f41d468706078cd34a8cbec004a9a622e
CRs-Fixed: 3499726
The number of perf CPU clusters can vary across
different targets. Currently only the second CPU
cluster is considered as the perf cluster, whereas
there can be more than 1 perf clusters.
Fix this perf cluster derivation for interrupt affinity.
Change-Id: Idecc98701e38bd2315846d6500068dcdd8f1309e
CRs-Fixed: 3498023
Remove tdls set state cnt check before sending
fw tdls state update. Fw state updates could be for
multiple vdev's set state cnt check works for only
single vdev.
Change-Id: Id154e623a03e65a99f27360c33b261209ac2e754
CRs-Fixed: 3483291
Currently, if active offload is disabled and ARP offload,
NS offload is enabled from ini host doesn't configure
NS and ARP patterns to fw. This results in frame drop when
device is in RTPM and NS, ARP frame is received as fw
doesn't send WOW wakeup to host.
To address this issue, configure ARP, NS pattern to fw if
active offload is disabled.
Change-Id: Iad54263f7a8a52f0a9ca3c56f9a2e3ccb19ea6c0
CRs-Fixed: 3468310
As per the requirement, add a new Ini CFG_PCIE_CONFIG to
enable or disable externally. This Ini is used to control
pcie generation speed and lane width parameters to max value.
If enabled as of now this Ini makes generation as 3 and lanes
as 2 for pcie.
Change-Id: I96c1a133fa0b22d3b35a229b2fd9ba2904bb3d8f
CRs-Fixed: 3481707
Current code does not enable DRV in unit
test mode. Fix this by setting drv flag
when unit test mode is detected.
Change-Id: I0996860ca6549235aae59459a710f10279a181d5
CRs-Fixed: 3500370
FW drops the management frames received on the inactive link,
so timeout happens on SCS action frames from userspace that are
sent on inactive link.
Set link agnostic flag in all management frame of subtype action
so that FW can send the packet OTA on the active link.
Change-Id: I4677617b52eca3f933aae2c4a7584750a95b3302
CRs-Fixed: 3486974
Populate the user-defined listen interval in association
request frame for STA.
Change-Id: I1733ed1d528fa99fc5556639ddfbd68ca97a40e0
CRs-Fixed: 3438369
The policy manager API to update hardware mode list is
invoked before the host can extract EMLSR capabilities
from the target via service ready extended event. Thus,
invoke the policy manager API again through SBS
frequency update API to set the EMLSR capabilities in
the host.
Change-Id: I4d71b983ddb4b09e2010d24f0af0c26e2077cbd7
CRs-Fixed: 3417254
When ANI_LITTLE_BIT_ENDIAN is not defined, a compiler error is
flagged in the definition of struct mlme_ht_param_info due to a
misplaced #endif, so relocate it.
Change-Id: I72df79a1dfae84c2cd6797d1d112de6ef8ca7ef6
CRs-Fixed: 3498279
SAP 1 is started on 5 GHz 5180, SAP 2 is started on 5 GHz 5745.
ML STA 2 GHz 2437 + 6 GHz 6275 is up. SAP 1 is moved to 2 GHz 2437,
SAP 2 is not moved to 2 GHz 2437 because the SAP2 PCL is PM_SCC_ON_5_CH_5G
which have no 2.4 GHz scc channel. STA+SAP+SAP SCC is allowed
in same mac, but MCC is not supported.
Fix by:
1. Change PCL to include SCC channel in 4th pcl table for SAP/GO.
2. Add API policy_mgr_modify_sap_go_4th_conc_disallow to filter
out the channels which not supported in same mac based on
current concurrency combinations.
Change-Id: Ieb482076d961234d3987b2daebab4c407ee0b7f5
CRs-Fixed: 3479297
When FEATURE_WLAN_MCC_TO_SCC_SWITCH is not enabled, a compilation
error is observed due to the fact that the stub implementation of
policy_mgr_change_sap_channel_with_csa() does not return QDF_STATUS,
so add an appropriate return statement.
Change-Id: I568289cc825adffa3fbcbab2419e093dc5c44cfa
CRs-Fixed: 3498278
Currently center frequency pass to API to check unsafe channel is
not correct set for bandwidth greater than 80 MHz. It lead to
unexpected shrink bandwidth.
For channel bandwidth greater than 80 MHz, center frequency should
set to mhz_freq_seg1 instead of mhz_freq_seg0.
Change-Id: I278590cd128b4760e35f4ea5921a6ed687dc1308
CRs-Fixed: 3466845
Logs without meaningful message or values serves no purpose in debugging.
Hence, add more context to the logs.
Change-Id: Ie913f7460356a1550b5252fd72ed5b454c67b3cf
CRs-Fixed: 3492709
Scenarios are:
1. Connect to AP in HE80/HE160
2. Set the bandwidth to 20 MHz (SET_MAX_BANDWIDTH 1)
3. Disconnect from AP
4. Reconnect to AP.
5. AP should be connected to the max bandwidth.
To fix this, host should update channel bandwidth to
the max supported bandwidth on disconnection.
Change-Id: I139d082ca4e4950d760f1da074dbcc021fb8a9e1
CRs-Fixed: 3491677
Currently, Fw did not set NACK reason code
to Host when Fw is not ready for WoW.
As part of the new change, Fw will send NACK
reason code along with WoW NACK to host.
Make changes in host to store and print
NACK reason code.
Change-Id: Ic978af9572c8de37df3fb926083aac5e6ddefda7
CRs-Fixed: 3481177
When user space use "iw dev" to get tx power info,
driver send WMI_REQUEST_STATS_CMDID to fw,
fw update tx power info with two events, usually we
parse the last event, and maybe get invalid value.
Filter invalid value for each event, and get valid
tx power value from the mac which the AP/STA mode work.
Change-Id: I2411bb87e03fe385759852b028819e19a055b7f3
CRs-Fixed: 3474361
When STA roams from multi link AP to single link AP, because MLO link
vdev is not disconnected yet, PCL is updated according to MLO station
connection and SLO station connection concurrency scenario.
After MLO link vdev is disconnected, PCL should be updated again
according to SLO station connection scenario.
Change-Id: I6a297c05a605fefe27c20a90d36e419819e6415d
CRs-Fixed: 3497055
Currently, packet capture packets are processed until the mon list
is empty. There can be situation that while mon list is not empty and
mon thread is running, mon interface is deleted and as part of it
PKT_CAPTURE_REGISTER_EVENT is cleared. After clearing the bit, interface
deletion will wait for completion of mon_register_event and flush the
remaining packets in mon list without processing them. In this case
as mon thread is already running and processing the packets from mon list
without check for PKT_CAPTURE_REGISTER_EVENT bit, it may lead to
timeout in deletion of mon interface in case of high throughput scenarios.
To avoid this situation, add a check to test PKT_CAPTURE_REGISTER_EVENT
each time before processing a packet from mon list.
Change-Id: I21e7adc0149c2330f6008d54db8576ca705f2b55
CRs-Fixed: 3475349
