Priority Queue gets paused due to descriptor deficiency which
needs to be unpaused after distributing descriptors from a
source pool to deficient pool.
Unpause the Priority Queue of deficient pool when avail->desc
count is more than start_th.
Change-Id: Ia91fdd0ac39bf8a4ac42dafa42d346993bf6a38e
CRs-Fixed: 2318031
Fix the locking order of pdev->peer_ref_mutex and pdev->local_peer_ids.lock
spin_locks in ol_txrx_disp_peer_stats(), this was resulting in deadlock
with lock order in ol_txrx_peer_get_ref_by_local_id().
Change-Id: I51d39daed7f8fb04c30fd6a73e2f9ca388543620
CRs-Fixed: 2315808
Logging macros already print function name by default,
there is no need to print the function name again as part of the log.
Hence, remove the function name or
wherever required use '"%s...", __func__'
Change-Id: Ib6dc6cc03ef148146b1c626271d5ac3b7345c4bb
CRs-Fixed: 2312593
For peer map, in ol_rx_peer_map_handler->ol_txrx_peer_find_add_id
->ol_txrx_peer_find_hash_find_get_ref, peer ref count debug id is
PEER_DEBUG_ID_OL_PEER_MAP.
After LFR3 complete, host need delete old peer in
wma_roam_synch_event_handler ->wma_delete_bss->wma_remove_peer->
ol_txrx_peer_detach_force_delete->ol_txrx_peer_release_ref,
but peer ref count debug id passed is PEER_DEBUG_ID_OL_INTERNAL,
so peer ref count is failed to be decreased, the peer is failed
to be detached and released, new peer of roaming target AP can't
be created and attached.
Change-Id: I18572a8a8a4429c10a4862bb188b43aefb830a14
CRs-Fixed: 2311751
In the scheduler_post_message, src_id is now added to
know the source module of the msg. the present scheduler
doesn't know about the same which is scheduler_post_msg
Replace the scheduler_post_msg with scheduler_post_message
Change-Id: Ia79d1e8ceda97f956eadcef50524b8307fcc247b
CRs-Fixed: 2305993
In function ol_txrx_offld_flush, check for pdev being
NULL, before de-referencing it.
Change-Id: I8999d894c74b6ef10cef91345370d859f5eee8ec
CRs-Fixed: 2299780
Fix and add debug_ids iappropriately account for function holding the peer
reference. This will help to further narrow down the peer leak.
Also add signatures in ref_release functions to catch error path.
Change-Id: I7f091ca8270ba724e24550e80208739d63d57b22
CRs-Fixed: 2293477
1) This reverts commit
Iab0862eda2392bd516c8ba0b913441b8e0d4c493
2) Reverted changes have been taken care through
Idd7617782e71ee187eef7fcb3523c05b49f82094
CRs-Fixed: 2300054
Change-Id: Icc47ded9a585e356b7eae1ad53ffea6668510308
Do not required to print function and line number
for data path statistics function which is invoked
by ioctl.
Change data path statistics to info print so
that it can be printed in dmesg as well.
Change-Id: I4b5ea4202255ace71dbb6f9a4bbff6f93e496425
CRs-Fixed: 2278885
APPS watchdog bark timeout happens due to continuous excessive
logging of "Not able to allocate context" in ol_txrx_offld_flush
and in ol_rx_data_process. Remove these extra logs.
Change-Id: Idf96ca584d77c68ad771a150604d1b9ca65ba3eb
CRs-Fixed: 2289492
Currently for fragment packets received legacy rx buffer replenish
is done which does not take in to account number of msdu count but
depends on fill count. Variable fill count is not protected and the
value becomes incorrect if AP sends both the fragmented and normal
packets at the same time. This leads to rx buffer replenish failure
and eventually data stall. To fix this do rx inorder replenish for
fragmented packets similar to normal ones if rx full reorder offload
is enabled.
Change-Id: I144c10be7b45268300b040e07ecb909a1ca113cc
CRs-Fixed: 2289797
Separate out HL and LL Rx Data Path in different files
to compile out features cleanly
Change-Id: Ifc0e2d7e740142f0115d1fcd53db2936f7a30979
CRs-Fixed: 2287351
Separate out HL and LL Tx Data Path in different files
to compile out features cleanly.
Change-Id: I9033a5bc83bbc95ff00be3f46dc6eda65eab9e71
CRs-Fixed: 2287346
Initialise spinlock: tx_mutex in pdev_attach instead of
pdev_post_attach as it may be accessed
in ol_tx_flow_ct_unpause_os_q() just after FW download.
Change-Id: I9de2eacd573420d23bc3a71a52dfb27bc1cfdab5
CRs-Fixed: 2285503
Add per module logging macros without function/line info
to avoid adding function/line info where it is not required.
Change-Id: Id808fef623b82cec38fc31071ef384f3ff0c92da
CRs-Fixed: 2278874
Add setter and getter function of timestamp for management frames
for the peer based on peer's mac address.
Change-Id: Iab0862eda2392bd516c8ba0b913441b8e0d4c493
CRs-Fixed: 2264667
At DP peer creation update UMAC peer objmgr handle to the
DP peer structure. Also update the DP peer handle to the UMAC
peer objmgr handle.
Change-Id: I9983fe9626eae9d2ad6e3b2a6fde9183f5bf6356
CRs-Fixed: 2259804
As part of DP CP interaction UMAC pdev objmgr handle is
passed at dp pdev attach. But in legacy DP config handle
is passed which leads to compilation issue. In this change
handle compilation issue by type casing handles.
Change-Id: I40986c8fd07d007aced91621f338f31332be6782
CRs-Fixed: 2260753
The txrx stats req has been freed in the ol_txrx_pdev_detach by checking
req_list of pdev.
Remove the txrx stats req free in the ol_txrx_fw_stats_desc_pool_deinit
to avoid the double free.
Change-Id: Idb2d5517e90ee873e7fd62d58c48a4f793266bac
CRs-Fixed: 2272696
Add CDP API to set RX MIC key in datapath.
Required to perform RX demic of fragmented packets.
Change-Id: I3c05da62ff20f8551cdc5780ce2f7b8d23c73a37
CRs-Fixed: 2217884
While processing the WMA_ROAM_OFFLOAD_SYNCH_IND event, The
driver takes nearly 150 to 200 msec time to just write fatal,
error and info messages to the kmsg log buffer. Also to get
tx power supplicant runs another thread while roaming. These
cause more latency(300ms) to send WMA_ROAM_OFFLOAD_SYNCH_CNF
message to fw. To decrease roam latency, the driver should
write the only necessary error message to kmsg buffer.
The fix is to change irrelevant fatal, error and info message
type to debug add a check in __wlan_hdd_cfg80211_get_txpower
to reject the request to get Tx power while roaming in progress
otherwise driver will process this and will increase the roaming
latency.
Change-Id: I6c31637f21991b5b29c1e2dd2c2935ae7d4fc3d0
CRs-Fixed: 2251824
The kernel address is used as cookie to keep track
of stats request. This address can be disclosed to
target leading to a security vulnerability.
Implement a FW stats descriptor pool, and use a
descriptor ID to keep track of stats requests,
instead of the kernel address, to prevent
kernel address leak.
Change-Id: Ib49150da899c0b9314f614868a90867f4aa92d3d
CRs-Fixed: 2246110
Return max tx groups supported in driver for all HIF types.
Earlier max tx group was being returned for SDIO only.
Change-Id: Id0431b03b89945386dd0f5aa0fe01b6d66fadde2
CRs-fixed: 2268757
1) Genoa FW by default enables HI_ACS_FLAGS_SDIO_REDUCE_TX_COMPL_SET.
When this flag is enabled, credits are reported through
HTT_T2H_MSG_TYPE_TX_CREDIT_UPDATE_IND and not through
HTT_T2H_MSG_TYPE_TX_COMPL_IND.
However when TSF and PTP features are enabled we need to get TX
Completions from FW.
Since credits can also be updated through TX Completions
we need to disable updation of credits through TX Completion for Genoa.
2) Enable flag : cfg_ctx->tx_free_at_download to free ol tx descriptors at
download.
Change-Id: I176dc8391ded9fc57f8be2b465effd8ae84eda49
CRs-fixed: 2268757
Map all module level log APIs to per-level log APIs
to compile out specific log level if required.
Change-Id: Icfcc28fb592ee99704d8fb23bc3cb45b6e4c9f53
CRs-Fixed: 2266718
When HL Flow Control is enabled, check for per vdev limits for making
a decision on whether frames should be discarded.
Change-Id: Ifde5f763d253eb644f97afeab9e4e1365c23e672
CRs-fixed: 2236321
1) Add ol txrx function to set TX descriptor limits for a
vdev.
2) Set the TX descriptor limits for STA and AP mode.
Change-Id: Ie9b1e52c3aff05db99ba3748a94792cc8116cdca
CRs-fixed: 2236321
1) Restart high priority netdev queue when tx descriptors get freed
and the vdev->tx_desc_count < vdev->tx_desc_limit.
2) Restart normal priority netdev queues when tx descriptors get freed
and the vdev->tx_desc_count < vdev->queue_start_th.
Change-Id: I99cc159289bc31558294e02cd9213a18472563f6
CRs-fixed: 2236321
Add function: ol_txrx_set_vdev_os_queue_status() to update
OS queue stop/start status in TXRX module.
Also call cdp_hl_fc_set_os_queue_status() from 'flow control
resume timer' to update OS queue status in TXRX module.
Change-Id: I03260985a48084bc523a3814c93c0e6b213e1970
CRs-fixed: 2236321
1) Populate legacy flow control ops 'cdp_lflowctl_ops'
for HL netdev flow control.
2) Also call cdp_hl_fc_register() to register HL flow control
callback.
3) Increase number of netdev queues to 5
Change-Id: I6bdb015b3325e54386f2aad093c89f6f782c24ed
CRs-fixed: 2236321
During hard_start_xmit, set a flag in SKB CB to indicate DHCP ACK
packet, and a callback gets called on TX completion.
Change-Id: I4e11b2c35b4747c7af35a387b404e3ccb1e8e8e4
CRs-Fixed: 2253557
Implement API to set dp trace params debugfs which
sets proto bitmap, num of records and verbosity
in the same order. API expects exactly 3 params to be
passed in the same order.
CRs-Fixed: 2184740
Change-Id: Ia802243bbd2e9d985a58ccf913d9cfbe3d71c8a8
IPA SMMU mapping for RX buffers is needed only when IPA offload
and IPA pipes are enabled. Currently in STA only case where IPA
is not enabled SMMU map/unmap is done for RX buffers. So enable
SMMU mapping only when IPA pipes are enabled.
Change-Id: I88db2cc8606bdf4586644a7ffccd0415f85c8241
CRs-Fixed: 2213795
Currently, in ol_txrx_is_peer_eligible_for_deletion(), invalid
dereferencing of peer_id_to_obj_map[0xFFFF] to get peer_ref while
processing VDEV stop response handler may occur.
Revert the changes introduced by
Change-Id: Icf252612081a41f94db6df4684348f2962b2da9d and
Change-Id: I743e2e2c83c3e07e5d5ec4fde7fc3b098766ca96
Change-Id: I7aa104f69a5665f0e08314fb0a273e077f562939
CRs-Fixed: 2261088
Fix overrunning callee's array of size 19 by evaluating argument tid
not to pass the maximum number.
Change-Id: I993339f4b9aea51e9566d213c9828825c5f2bf66
CRs-Fixed: 2232744
In case the current selected txq group, does not have enough credits,
try to borrow credits from the other txq group.
Change-Id: I86fbe990853d90598f6e09b13f7061e4ba1a78ae
CRs-fixed: 2246206
1) When a group is created i.e. the first vdev is added to it,
assign all the credits to it.
2) When the second group is created, transfer some minimum credits
to it.
3) When a group is deleted, transfer its credits to the other group.
Change-Id: I0c5532033718b250ab0633b4da4e219c0315cac9
CRs-fixed: 2246206
1) Add function: ol_tx_update_grp_frm_count() to maintain count of frames
per group.
2) Call ol_tx_update_grp_frm_count() from ol_tx_enqueue(),
ol_tx_dequeue() and ol_tx_queue_free().
Change-Id: If1b07ea5bbdcbc6ad6d0c91e6b2060c4264b9472
CRs-fixed: 2246206
When DUT P2P Go/SAP deauth ref STA, in race condition, scheduler
thread may try to clear peer data and drop pending rx packets
after peer freed in peer unmap handler in soft irq context,
use after free issue will happen.
Error log:
BUG: spinlock bad magic on CPU#1, scheduler_threa/28550
Unable to handle kernel paging request at virtual address
6b6b6b6b6b715b
Stackframe:
do_raw_spin_lock+0x34/0x154
_raw_spin_lock_bh+0x24/0x30
ol_txrx_clear_peer_internal+0x68/0xb0 [wlan]
ol_txrx_clear_peer+0x78/0xa0 [wlan]
hdd_softap_deregister_sta+0xd0/0x200 [wlan]
hdd_hostapd_sap_event_cb+0xca8/0x20b8 [wlan]
Change-Id: Ib8d133528f5ff22125218861206d241f96eaf0da
CRs-Fixed: 2247334
Separate out QCA_LL_LEGACY_TX_FLOW_CONTROL
and QCA_LL_TX_FLOW_CONTROL_V2 flow control implementation
in different files to compile out features cleanly.
Change-Id: I5d6ddf9ea61b409b25d242852ed1f0102e94ad88
CRs-Fixed: 2228902
ol_tx_update_connectivity_stats() in tx completion
path updates connectivity stats referenced from tx_desc.
In cases when vdev has gone down and tx completion are received
leads to NULL vdev access. So, add check before accessing vdev.
Change-Id: I402d740ab3ecd923aa1b632bd0c59447599c17df
CRs-Fixed: 2225053
ol_tx_update_arp_stats() in tx completion path updates the per vdev
arp statistics. vdev is referenced from tx_desc.
In cases when vdev has gone down and tx completion are received leads
to NULL vdev access. So clear reference to vdev inside tx_descs when
vdev goes down.
Change-Id: Ic8c854b42ece41489f71e1374e5e72580308e9fe
CRs-Fixed: 2215312
During peer deletion, ol_txrx_is_peer_eligible_for_deletion() is
called to check if peer is eligible for deletion. Inside function,
vdev is dereferenced to extract pdev but due to race conditon peer
may get freed from the list and this may lead to NULL pointer
derefencing of vdev.
Avoid dereferencing of vdev and pass pdev itself as an argument to
ol_txrx_is_peer_eligible_for_deletion()
Change-Id: I743e2e2c83c3e07e5d5ec4fde7fc3b098766ca96
CRs-Fixed: 2252243
