During idle shutdown(T1 Thread), Driver sends the PDEV
suspend command to FW. If ACK is not received, driver initiates
the work for self recovery. As a part of self recovery(T2 thread)
driver allocates the memory and sends the force fw hang cmd and
wait for the event.
T1 thread continues and stop all the modules and check for mem leaks.
Since t2 thread is not completed and free the memory, the crash
happens as a part of mem leak check.
Fix this race condition by cancelling and flushing the recovery work
in cds close.
CRs-Fixed: 2730021
Change-Id: Ieffff621e0838b341531c5b7feda1277ee4f70fd
Register STA tx completion callback with dp as part of
vdev registration to update successful ack stats.
Change-Id: I27f36336f0e37d43cc2cdff204c0217667a48b13
CRs-Fixed: 2682020
Currently once the gEnableSelfRecovery is disabled it is decided to panic
once the firmware complete firmware dump is collected, this has a
side-effect when the normal SSR is triggered for testing purposes.
This reverts commit Ia184a265d0f76e43d1d5edc4bcc0849a05726dac.
Change-Id: Icec14a2afb902a275ffe0bf202ec75d8e58d9f58
CRs-Fixed: 2725430
Presently, in the host initiated recovery the handler checks whether
the support for recovery is enabled via the ini or triggers the panic.
Whenever the panic is triggered the irq's will be disabled,
this results in some targets not able to collect the important
MHI, PCI information on what caused the issue to root cause the issue.
So, for different targets check the ini at different places. In this
case, move PCI based check to the start.
Change-Id: Ia184a265d0f76e43d1d5edc4bcc0849a05726dac
CRs-Fixed: 2720999
Replace wlan_reg_is_24ghz_ch with wlan_reg_is_24ghz_ch_freq.
Replace channel with frequency in cds_get_vendor_reg_flags.
Change-Id: I72ffc5550e45be1edf460f9619fead8889cf2967
CRs-Fixed: 2714723
Presently, in the host initiated recovery the handler checks whether
the support for recovery is enabled via the ini or triggers the panic.
Whenever the panic is triggered the irq's will be disabled,
this results in some targets not able to collect the important
MHI, PCI information on what caused the issue to root cause the issue.
So, for different targets check the ini at different places.
Change-Id: I0d0e322610abd9f8ceaba4fe9688cdd099c381e5
CRs-Fixed: 2715603
Currently there is a single INI item tcp_udp_checksumoffload to enable
or disable this feature for all the adapters. In some cases, we want to
be able to enable this feature selectively for NAN mode.
Add nan_tcp_udp_checksumoffload ini item to enable or disable checksum
offload feature specifically for NAN mode. If
nan_tcp_udp_checksumoffload is 'true' and tcp_udp_checksumoffload is
false, hardware checksum offload will be enabled only for the NAN
adapter and not for other adapters.
CRs-Fixed: 2698342
Change-Id: I56319095e6229d77ddb5c790d0af9c6ba0267611
While coping version strings in cds_hang_event_notifier_call, specify
exact size of QWLAN_VERSIONSTR/QDF_HANG_EVENT_VERSION to avoid buffer
overflow.
Change-Id: Ie2203600a2c92bb78ba076b609375ca219773b19
CRs-Fixed: 2702069
In cds_trigger_recovery_handler, Host driver first set flag recovering
before cds_force_assert_target. For Rome, ol_target_failure run when
FW crash, Self-Recovery stop when detect flag recoverying be set
Add CDS_DRIVER_STATE_ASSERTING_TARGET to record Host asserting in
cds_force_assert_target. Skip recovering check in ol_target_failure
if detect this flag.
Change-Id: I290e2f21bd0bca6dc6d76de43de4f8dd801d7ff5
DP init/deinit flow to have dp_soc_init/dp_pdev_init
APIs where it takes care of initializing rings,
descriptors and DP data structures that was earlier
done as part of dp_soc_attach/dp_pdev_attach itself
This is to maintain symmetry across wifi load/unload
wifi load to call:
1. dp_soc_attach
2. dp_soc_init
3. dp_pdev_attach
4. dp_pdev_init
wifi unload to call:
1. dp_pdev_deinit
2. dp_pdev_detach
3. dp_soc_deinit
4. dp_soc_detach
Change-Id: I5889e0eafdb2fa6319c440725590e7a3014d8be0
CRs-Fixed: 2674081
Presently whenever the control path is allocating the skb it is
passing the filename rather than than the function name required
by the qdf nbuf framework which is not much useful since when the
absolute path of the filename is considered and that can be truncated
leaving with only linenumber to understand the issue.
Pass the function name so it would help in root-causing the issue.
Change-Id: I998a88a3d22354500f3d01d09c91f0367755a179
CRs-Fixed: 2659490
In case wait for ready event timeout from fw, need to
suspend fw firstly, otherwise fw is not aware host
will free resources.
Change-Id: Ifb7ab7fcf0cf468d59dc462dfdb27851fa4e596e
CRs-Fixed: 2669433
Update the hang data for the cds module, as part of the hang
data following information is filled:
1) Hang Event version
2) Wlan Driver version
3) Reason for recovery
Change-Id: Ie5a008d1f9512f8e83a1d099fe333b0d2ce90abb
CRs-Fixed: 2650848
The QDF framework for triggering self recovery is undergoing changes to
include support for multiple psoc driver. As the MCL host is still
multiple psoc agnostic, create dummy wrappers over the
cds_trigger_recovery so as to maintain compatibility with modified
qdf_trigger_self_recovery
Change-Id: I42163879c3e786a2b9ab2f9f5785be5531754e77
CRs-Fixed: 2617709
Handle send delba indication from DP. It will help DP to
do aggregation tid management.
Change-Id: I18455107ef8d042644efc8ce549a104612eec05c
CRs-Fixed: 2637485
When suspend/resume timeout, it will trigger recovery,
current as it direcly trigger assert to fw, which will
take more time to dump ramdump, so DPM WD timer expired.
To avoid DPM WD timer expired, trigger ramdump in a
separate workqueue, so it will not block DPM WD timer.
Change-Id: I1a9349a05a37544329946270065037cd90172fc7
CRs-Fixed: 2645163
Enable IPA SMMU S1 translation feature flag for
Lahaina target.
asm/dma-iommu.h file is removed in latest 5.4 kernel,
resulting in include failure. Remove as it is unused
for version 5.4 and above.
Change-Id: Ie2147f2fc2debad4d700fae5414d1d3becb28f13
CRs-Fixed: 2638580
Currently many layers store the datapath handles i.e.
pdev and vdev handles and use them for passing as an
argument to the CDP APIs. This gives way to possible
race conditions, where the datapath handle provided by
the outer layer might be a stale pointer, which in turn
leads to unexpected results.
Remove the storage and usage of all the datapath handles
from the layers currently using them.
- cds_context
- vdev objmgr
- pdev objmgr
Change-Id: Icf24bd3792af203d1a06e153f0badd6e7405559f
CRs-Fixed: 2602199
Use newly defined QDF API in CMN qdf_is_driver_unloading() to check
if driver unload is in progress in DSC.
Change-Id: I2958c6cc3ee6f2ac4302e2dbbf1c68d986fea175
CRs-Fixed: 2605706
Modify the max tx power value from 30db to 24db
to pass certification test cases.
Change-Id: Idf9c3ca10d3216f0f4b1cf78f87586c6bd3c2c39
CRs-Fixed: 2601405
There might be race condition when doing RX GRO flush between dp_rx_thread
and vdev detach thread.
Only doing RX GRO flush in the dp_rx_thread context, and use the events to
sync with dp_rx_thread and vdev detach thread.
Change-Id: I2dbf92c2c11e950ca790dd7429c8bb235c239d0c
CRs-Fixed: 2601826
Create a mon thread to process the packets for pktcapture
mode only if packet capture mode is enabled.
Change-Id: I64e3687d08acd0d27019ed9e11b5b92b13fcd878
CRs-Fixed: 2528545
Update packet capture mode param in the HDD context
after reading the INI parameter.
Change-Id: I3ea88e376f1ad8ebc96715d08395331470c23d4a
CRs-Fixed: 2528536
Currently cdp ops are given pdev/vdev/peer
handle as its arguments which is directly
accessed in those APIs. This can cause a
race-condition in access of the respective
handles if it has been deleted in parallel.
Hence as a part of cdp convergence, pass only
the pdev_id or vdev_id or peer mac address,
which will be used to get the respective handles,
and hence avoiding the unwanted access of the
handles if it has been deleted.
- txrx_data_tx_cb_set
- txrx_mgmt_tx_cb_set
- txrx_mgmt_send_ext
- txrx_pdev_attach_target
- txrx_pdev_detach
- txrx_vdev_attach
- txrx_vdev_register
- txrx_vdev_detach
- txrx_peer_create
- txrx_peer_delete
- txrx_peer_delete_sync
- txrx_fw_stats_get
- txrx_get_vdev_from_vdev_id
- txrx_pdev_set_ctrl_pdev
- txrx_pdev_attach
- txrx_get_vdev_mac_addr
- txrx_pdev_post_attach
- txrx_pdev_pre_detach
- txrx_get_ctrl_pdev_from_vdev
- txrx_get_mon_vdev_from_pdev
Change-Id: Ib8b0b1edd63e7c3f8813b11ca26b50ec121fe255
CRs-Fixed: 2541459
Presently, during Subsystem Restart(SSR) the object references
are not released cleanly, invoke stop adapter during the SSR
to have common functionality between the netdevice going down
and SSR.
Change-Id: I2980379022f62ef27dea92868c8033c087544d50
CRs-Fixed: 2587443
Currently cdp_host_get_vdev_stats API provided
with vdev handle as its arguments, which is
directly accessed in that API. This can cause
a race-condition in access of the respective
handle if it has been deleted in parallel.
Hence as a part of cdp convergence, pass only
the vdev_id which will be used to get the
respective handles and hence avoiding unwanted
access of the handle if it has been deleted.
Change-Id: I4cc959b9244f5c22dea71194c86c583cd9c2a071
CRs-Fixed: 2583358
To avoid conditional compilation inside function body use QDF API
qdf_set_wake_up_idle in RX thread startup api.
Change-Id: Ibe78fb19893c6a942cfbdde2f33684ea402af47b
CRs-Fixed: 2580958
On some platforms with TQM, target don't know packet number of
successfully transmitted after more than one retransmission attempt,
and after TQM retry, host data path will be notified.
Adjust tx retries stats number from data path, and indicate to upper
layer when query by wlan_hdd_cfg80211_get_station() on these platforms,
other platforms keep tx retries from wmi vdev stats.
Change-Id: Ie4b9a9f3b9241a59452e3f21f68757f3638b1a19
CRs-Fixed: 2574368
During uplink traffic, affine RX thread to big cluster if
throughput is high and rx_packet counts above threshold.
Change-Id: I34034e79ee07c229966a279919acc11d1942b890
CRs-Fixed: 2557745
Currently control path handles of pdev/vdev/peer
are given as arguments to cdp ops, which are
directly accessed by dp layer. This can cause a
race-condition in access of the handle if it
has been deleted in parallel.
Hence as a part of cdp convergence, pass only
the psoc handle which can be accessed all over
to retrieve objmgr_pdev from psoc.
- rx_mic_error
- txrx_pdev_attach
- txrx_vdev_attach
- txrx_peer_create
Change-Id: Iaec8b913810f9035faa0010117cdff19bdded37f
CRs-Fixed: 2541431
Add logic to queue and bundle packets before
giving it to scheduler to ensure predictive HTC
bundling for improvement in TX throughput.
Change-Id: Ib00135ca59b4c5f0f05edc93dca115bf17174a36
CRs-Fixed: 2561671
Send vdev create to firmware when the mlme vdev create handler
is invoked as part of the vdev object create.
Change-Id: Ibba02f6b5885d7b3a62529ef51768f520395513e
CRs-Fixed: 2555943
Currently the cdp apis are given pdev/vdev/peer
handle as its arguments, which is directly
accessed in those APIs. This can cause a
race-condition in access of the respective
handles if it has been deleted in parallel.
Hence as a part of cdp convergence, pass only
the pdev/vdev id or peer mac address, which will be
used to get the respective handles, and hence
avoiding the unwanted access of the handles if
it has been deleted.
Converged misc_ops
- tx_non_std
- get_opmode
- get_tx_ack_stats
- set_ibss_vdev_heart_beat_timer
- hl_tdls_flag_reset
- set_wisa_mode
- txrx_post_data_stall_event
- update_mac_id
- pkt_log_init
- pkt_log_con_service
- get_num_rx_contexts
- set_wmm_param
- flush_rx_frames
- bad_peer_txctl_set_setting
- bad_peer_txctl_update_threshold
- txrx_data_stall_cb_register
- txrx_data_stall_cb_deregister
- vdev_set_driver_del_ack_enable
- get_intra_bss_fwd_pkts_count
- mark_first_wakeup_packet
- register_pktdump_cb
- unregister_pktdump_cb
- pdev_reset_driver_del_ack
- runtime_suspend
- runtime_resume
CRs-Fixed: 2539735
Change-Id: I2baad09fde29d439ae2099e01e4e68ed6b94196d
policy_mgr_is_hw_dbs_required_for_band is for any HW where
PHYA/MAC0 doesn't support the given band. We need to switch
to DBS mode to support the band. So far we have
overloaded policy_mgr_is_hw_dbs_2x2_capable to achieve above.
Need a cleaner API to work in a generic way
(for 2x2, 1x1 or any chain mask configuration).
Change-Id: I1cb28bece1242fc749a0a6a56fadca0502850c43
CRs-Fixed: 2565939
Currently the cdp apis are given pdev/vdev/peer
handle as its arguments, which is directly
accessed in those APIs. This can cause a
race-condition in access of the respective
handles if it has been deleted in parallel.
Hence as a part of cdp convergence, pass only
the pdev/vdev id or peer mac address, which will be
used to get the respective handles, and hence
avoiding the unwanted access of the handles if
it has been deleted.
Converged ocb_ops
- set_ocb_chan_info
- get_ocb_chan_info
CRs-Fixed: 2539745
Change-Id: I7de53c48ec6b7f0c09be7c9cf642d8cf4f8eae78
Currently MCL and WIN both have separate queue.h
files with the same declarations and definitions.
As part of cleanup single queue.h is created in cmn code,
use this newly created queue.h file instead of MCL
specific queue.h file.
Change-Id: I91619e5fff2040cb588cd0413cb42d5d7469f41f
CRs-fixed: 2505800
As a part of cdp convergence the pdev_id
is used to obtain the pdev handle on need
basis.
Instead of the pdev handle, use the pdev_id to
retrieve datapath pdev handle.
CRs-Fixed: 2539733
Change-Id: I887ea2f1a312207cbb642b33435f4e7111068c62
As a part of 802.11ax amendment, 6GHz band operation is added.
Since the 6 GHz channel numbers are overlapping with existing 2.4GHz
and 5GHz channel numbers, use frequency to identify unique channel
operation instead of channel number. Channel frequency is unique across
bands.
As part of above requirement add logic to process rx mgmt
packets based on the frequencies instead of channel numbers.
Change-Id: Ib063070738ecdb4f83379eafe50629778a490aae
CRs-fixed: 2522693
