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
Currently, sizeof(qdf_dma_addr_t) is 4 bytes, but IPA driver
and WLAN FW always use 8 bytes, we should follow up with them,
otherwise, the numbers of rx complete ring entries will not be
same in both sides. Meanwhile, for tx complete ring, during the
first time initialization, we fill the buffer address with 4
bytes, while GSI FW read 8 bytes address content from tx complete
ring, which should cause the GSI FW get invalid address on the half
way and cause the smmu fault issue. So fill 8 bytes length buffer
address to the tx complete ring can avoid such issue.
Change-Id: I61021627b634d4b1525727cc944ee410ed3154db
In dp_rx_tm_suspend, suspend_event should be reset before waiting
for it. otherwise, the suspend function will be break.
Change-Id: I9132f45b02386c54fb8e27f0f7c2c71205450c9d
CRs-Fixed: 2606227
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.
Also remove the ops which are used to store
and retrieve such handles, that no more needed.
- register_peer
- clear_peer
- change_peer_state
- peer_get_ref_by_addr
- peer_release_ref
- peer_state_update
- get_vdev_by_sta_id
- register_ocb_peer
- get_peer_state
- get_vdev_for_peer
- update_ibss_add_peer_num_of_vdev
- remove_peers_for_vdev
- remove_peers_for_vdev_no_lock
- copy_mac_addr_raw
- add_last_real_peer
- is_vdev_restore_last_peer
- update_last_real_peer
- peer_detach_force_delete
Change-Id: I5134ca882f5fcf0e5941317e48712717eddb58fe
CRs-Fixed: 2541708
Currently cdp ops are given pdev/vdev 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/vdev_id which will be used to get the
respective handles and hence avoiding unwanted
access of the handles if it has been deleted.
Also deleting few ops which stores and retrieves
such handles and adding ops to set/get pdev params.
- txrx_set_safemode
- txrx_set_drop_unenc
- txrx_get_pdev_param
- txrx_set_pdev_param
- wdi_event_sub
- wdi_event_unsub
Change-Id: Ib3511cbdead75b30e899dbf00f936ffdae316b71
CRs-Fixed: 2541657
Today we enqueue nbuf list to rx_thread via nbuf frags_list.
If deliver_head happens to have frags_list then it is delivered alone.
Remaining list of skbs are delivered using frags_list of next skb.
It does not takecare to update QDF_NBUF_CB_RX_NUM_ELEMENTS_IN_LIST of
the new head_skb with remaining num skbs, make sure to update it.
Also add nbuf_queued to take care of this accounting.
Change-Id: I5557b5c56576c46fc8564a87a27d167b8785bb95
CRs-Fixed: 2559875
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
If there are still RX packets pending in the napi gro_hash list, do RX GRO
flush after flushing the RX thread pending packets when detach the vdev to
avoid net_device might be invalid later which has been stored already in
the pending skb.
Change-Id: Iff0a5cf49c87b91b8cae3dbcb541d1cf0bc480a6
CRs-Fixed: 2598149
In case of nbuf map failure as part of htc_send_pkt for
tx credit flow disabled endpoints, htc pkt is not put
back in the freelist as return status is not checked
resulting in memory leak.
Fix is to put the htc pkt back in the freelist based on
status returned by htc_send_pkt.
Change-Id: I9173fa3684bc887a360b92d48b689c53a5ad5dbe
CRs-Fixed: 2593729
Function ol_txrx_hl_tdls_flag_reset will call function
ol_txrx_get_vdev_from_vdev_id, it will return NULL vdev
because ol_txrx_hl_tdls_flag_reset is called before
TAILQ_INSERT_TAIL(vdev).
So, move it after TAILQ_INSERT_TAIL(vdev).
Redefinition of 'pld_pcie_register_driver' when
CONFIG_PLD_PCIE_CNSS not defined. For Hastings WHUNT,
HIF_PCI is defined but CONFIG_PLD_PCIE_CNSS is not defined.
So, replace CONFIG_PLD_PCIE_CNSS by HIF_PCI and
CONFIG_PLD_PCIE_FW_SIM.
Change-Id: I05e58516efe6a9a445168a27df9e0cfc9de2e50c
CRs-Fixed: 2589802
Currently txrx_ops.tx.tx is given vdev handle
as its argument, which is directly accessed
by the caller. 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 handle and hence avoiding unwanted
access of the handle if it has been deleted.
Change-Id: Ie158a2ba805f52658cd6fb12c969fc0d677c1815
CRs-Fixed: 2541631
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
For vht bw, there is a transformation of rx_status->vht_flag_values2
in qdf_nbuf_update_radiotap_vht_flags(), so strictly copy
rx_status->vht_flag_values2 from rx_desc, do not transform twice which
leads to wrong vht bw.
For rssi, only show first mpdu rssi info in a ppdu, firmware reports
-86 rssi_comb for mpdus other than the first one, calculate
ANTSIGNAL by adding with noise floor will show a wrong rssi for those
mpdus.
Change-Id: I776d85cea2cf0f48beeed351ed0643699d805d54
CRs-Fixed: 2588768
For cdp_hl_fc_set_td_limit() API, use channel frequency instead of
channel number.
Change-Id: Ic053be33ac7c7cbd2543b0208a1184b3f20a4346
CRs-fixed: 2565734
In case of TSO, same skb results in multiple tx_desc after
segmentation. To avoid multiple free of skb "skb->users"
count is used. Currently, race condition is observed while
incrementing "skb->users" count in ol_tx_ll_fast and before
that, tx_completion is received for all the segments which
are sent before. In this case, "skb->users" count will be 1
so, skb is removed from the skb debug node which leads double
free of skb when tx completion is received for the last segment.
To fix this, increment the "skb->users" count before sending
first segment and decrement only iwhen the last segment.
Change-Id: I26000fcda8715b737cac803ec8e64c0fb640dfe4
CRs-Fixed: 2582617
Add credit tracking for HL Data Path in DPTRACE.
Also update proto_bitmap to dynamically enable/disable
credit tracking.
Change-Id: Idd9216b634586f3a81c2194ea4bd2122a472154a
CRs-Fixed: 2574053
Protect pktlog funcitonality under given mutex to avoid possible
race condition between pktlog_release/pktlog_open and pktlogmod_exit.
Also call pktlogmod_exit before calling wdi_event_detach to avoid
accessing freed memory in pktlog release.
Change-Id: I452af523338788447f5c2764e883165237083f7d
CRs-Fixed: 2087777
Local peer_id is being cleaned up across DP, HDD and PS/WMA.
So, any references to local peer_id/sta_id will be replaced
by peer mac address and all interactions between the layers
will be based on peer mac address.
Cleanup local peer id from ol_txrx APIs.
Change-Id: Ibc893fdf1f091f9fc423fca71eb96ee405373d99
CRs-Fixed: 2529036
Local peer_id is being cleaned up across DP, HDD and PS/WMA.
So, any references to local peer_id/sta_id will be replaced
by peer mac address and all interactions between the layers
will be based on peer mac address.
Cleanup unusued ol_txrx APIs which are based on local peer id.
Change-Id: Ia6949796bf6109884ec066544c39183f934f7d72
CRs-Fixed: 2529014
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
The query for an interface opmode returns the opmode
of a particular interface without checking if that
particular virtual interface handle is valid or not.
Fix this by checking the virtual interface handle
before trying to get the operation mode for that
particular interface.
CRs-Fixed: 2572981
Change-Id: I05b5f88455d6374adb41a037d3c5dda905de1497
In Monitor Mode, RSSI is always -96 for packets in TCPDUMP.
Populate the correct RSSI value in Monitor Mode.
Change-Id: Iba9fe7091043d27828f4742058ed3d389cde7e26
CRs-Fixed: 2568129
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.
- ipa_get_resource
- ipa_set_doorbell_paddr
- ipa_set_active
- ipa_register_op_cb
- ipa_get_stat
- ipa_tx_data_frame
- ipa_uc_get_share_stats
- ipa_uc_set_quota
- ipa_enable_autonomy
- ipa_disable_autonomy
- ipa_setup
- ipa_enable_pipes
- ipa_disable_pipes
- ipa_rx_intrabss_fwd
Change-Id: I678d7a7de7132417ff6051b0fd6da5d14426d21e
CRs-Fixed: 2540861
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 bus_ops
- bus_suspend
- bus_resume
CRs-Fixed: 2539749
Change-Id: I7f57c09fe4620f90c5cfab781635f6f0ada08ec5
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
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 l_flowctl_ops
- register_tx_flow_control
- set_vdev_tx_desc_limit
- set_vdev_os_queue_status
- deregister_tx_flow_control_cb
- flow_control_cb
- get_tx_resource
- ll_set_tx_pause_q_depth
- vdev_flush
- vdev_pause
- vdev_unpause
CRs-Fixed: 2539748
Change-Id: I8ad4a20914c654b8e8aaf629dda7f673fdb110e4
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 mob_stats_ops
- clear_stats
CRs-Fixed: 2539747
Change-Id: I5ee060b62d15380df4b8d15bb5709d700902aafd
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 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 throttle_ops
- throttle_init_period
- throttle_set_level
CRs-Fixed: 2539744
Change-Id: I0e1339ead92d66463d3b58ed9268c7194fd2bd27
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 delay_ops
- tx_delay
- tx_delay_hist
- tx_packet_count
- tx_set_compute_interval
CRs-Fixed: 2539742
Change-Id: I201d343f4ced3b861a65ddf78e10e66f786e5e91
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 flowctl_ops
- flow_pool_map_handler
- flow_pool_unmap_handler
- dump_flow_pool_info
- tx_desc_thresh_reached
CRs-Fixed: 2539738
Change-Id: I9b11e66d8c2644f16e0f2464c323798a4cba007a
In non concurrency scenario with TDLS Offchan link, FW is sending
two pauses namely PAUSE_TYPE_CHOP and PAUSE_TYPE_CHOP_TDLS_OFFCHAN
back to back when DUT dwells between TDLS Off channel and AP channel
respectively. HOST handles these events as vdev based event instead
of peer based event and remove all the tx queues from scheduler for
all the peer irrespective of the pause type. This will lead to data
stall in TDLS offchan scenarios, as at any point of time, host will
be receiving one of the two pauses and keep its peer queues out of
scheduler.
Changes are done to decouple the handling of pause/unpause events
for TDLS peers with Offchan enabled from vdev based pause/unpause.
TDLS peers will be paused/unpaused based on PAUSE_TYPE_CHOP_TDLS_OFFCHAN
pause type. Other pause/unpause events will be handled as usual.
Change-Id: Iccb46fd6d121d5df6d53633c9978ddc8e02f588f
CRs-Fixed: 2558612
Make tx resource hi/low threshold configurable through
config files and update hi threashold from 7.5% to 8%.
Change-Id: I2e52b7910caedf6e0af24242b79e1e9d42c6b85b
CRs-Fixed: 2548100
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
When a particular vdev is deleted, the corresponding rx
packets which have been queued to the rx thread are not
flushed. Hence when such packets are submitted to the
network stack, the dev for this skb will be invalid,
since we have already freed the adapter.
Flush out the packets in the rx thread queues, before
deleting the vdev.
CRs-Fixed: 2552140
Change-Id: Ia49af2c203c64077f7fd87524bb4caa4060e0044
Fix compilation issues related to IPA and DP when built for target
QCA6174. Here is the change to fix them.
Change-Id: I4441858450a78cdd98e3c5e25cdeee00c3a7db9e
CRs-Fixed: 2554619
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
