The datatype of average and round up tx rate was uint32_t which truncates
the upper byte of the average and round up tx rate value.
changing the datatype of average and round up tx rate from uint32_t to
uint64_t.
Change-Id: I3809ba7cfc73ec0b4609981e70d57112ac958f33
CRs-Fixed: 2409751
Add a CDP function which processes the peer delete response
and deletes the static ast entry for the peer
Change-Id: Id646979ed07bd82609ca08e7ef3201382d07b1de
CRs-fixed: 2385115
In dp_peer_add_ast() call the CP callback inside ast lock
to ensure the ast add and delete commands are called in a
sequence
Change-Id: If76bf8dca34555e56074e74792f3283b6b5a28d3
CRs-fixed: 2424637
On low memory system, skb allocation failures are observed in
dp_tx_ipa_uc_attach during driver loading phase. Per kernel
memory team analysis, failures are caused by large amount of
atomic allocations. Currently when allocating IPA TX skbs,
srng spinlock is first grabbed and then skb buffers are being
requested, which makes skb allocation happen in atomic context.
Fix is to make skb allocations with srng unlocked since it is
safe for race conditions during driver loading time. This ensures
skb allocations in process context.
Change-Id: I1624276c087c8247d672fb7cea5daded07ab93a3
CRs-Fixed: 2426489
Count of MSDUs Tx with AMPDU aggregation and
MSDUs without AMPDU aggregation is maintained. This
count is obtained from per ppdu stats.
Change-Id: I7c38846534bbe5f2fd0a9b3602a9d453e1c1f449
currently host will dump TLV with info level if HW indicate RX
data with invalid sw_peer_id, this will introduce excessive logging
issue, host panic.
degrade the log level to debug level.
Change-Id: Ia24f411bdfd9ec8e315874340d3bf046c245a72b
CRs-Fixed: 2420640
Add support to measure interrupt latencies, store them
in different "bins" and display the constructed histogram
through user command.
Change-Id: I268be6b39745d550ac7db5d321c2d1fdb2d17d41
CRs-Fixed: 2365812
When the status is FLOW_POOL_ACTIVE_PAUSED, all queues are expected to
be paused. Still some on the fly packet can make it to the driver. In
such a case, the function dp_tx_desc_alloc takes a default action of
WLAN_WAKE_NON_PRIORITY_QUEUE. This is incorrect, since it will wake all
non priority queues, when they should be paused.
Make default action as WLAN_NETIF_ACTION_TYPE_NONE. If this is the
action to be taken, donot call the pause_cb from dp_tx_desc_alloc.
Rate limit log levels in case dp_tx_desc_alloc fails.
Change-Id: I1ef3018e90576d2c3aaa0d10d56e9b155681271b
CRs-Fixed: 2421813
In some cases, the msdu_len as retrieved fromm msdu_start TLV is
greater than RX_BUFFER_SIZE. In such cases, when qdf_nbuf_set_pktlen
is called, it will expand the skb to accommodate the bigger packet.
This makes the rx_tlv_hdr var invalid, since it continues to point to
the older skb->data. Access of this rx_tlv_hdr will cause exception.
Drop the packet if msdu_len > RX_BUFFER_SIZE. This is not expected while
reaping WBM RX Release ring.
Change-Id: I60890e4d3ee0afa451884d4df458eb50be280280
CRs-Fixed: 2426245
Filter out data packets from mon destination
ring in m_copy mode as we already receive
first 100 bytes of data in mon status ring
Change-Id: I09321aa42b850d7b3cdc0bda4c7c51f79a6f850e
IPA driver adds is_txr_rn_db_pcie_addr and
is_evt_rn_db_pcie_addr in ipa_wdi_pipe_setup_info and
ipa_wdi_pipe_setup_info_smmu structure to check if doorbell
address is DDR address or PCIe memory mapped address. Thus
set the addr flag accordingly for IPA transfer and event
rings.
Change-Id: Ia55d14535db3818439e3884cfb61c3a1d81b86fb
CRs-Fixed: 2422162
At the time of dp vdev registration, register tx completion
callback so that tx completions will be called for packets
whose tx completion is requested.
Change-Id: Ic3402f41a0962111901271fc8669974dfea2eb1c
CRs-Fixed: 2416554
Not typecasting 64bit value to 16bit value before assigning is causing
corruption of nearby structure members. Typecasting this value fixes
this corruption.
Change-Id: Iddaf697eb90241f9948aeaafe7985b8c2ea63048
peer attach failed due to memory allocation failure
which triggered soc restart, as part of soc restart
peer detach tried to free AST hash table and entries
which was not allocated.
For peer hash table and AST hash table free only if
the memory is allocated
CRs-Fixed: 2425963
Change-Id: Ib7a624c91f544f1a8da2b96b4d342a13b9f3b162
RX_BUFFER_SIZE macro got introduced by mistake during rebase. The macro
is present in hal_rx.h. Removing the extra macro.
CRs-Fixed: 2382076
Change-Id: Ia66079d6d4543b4e3fd99e9f0c0376353a92aa9c
Vlan tagged STP pkts are received in host from
wbm exception path with wifi parse error code.
This is delivered to stack. Stats added for wifi
parse error and vlan tagged stp pkt count.
Change-Id: I25af5ab403ee7a46b4abe2f60328314d61587b80
CRs-fixed: 2375575
Currently rx reorder queue setup command is sent from soft irq
context. There exists race condition between resume command sent
from worker thread context and this command sent from soft irq
context. In parallel execution on different cpu cores if the wmi
suspend flag is not sent to false as part of resume command at the
time of sending the rx reorder queue setup wmi command, the command
will not be sent. So serialize the rx reorder queue setup command
by posting to scheduler thread.
Change-Id: Ib4d0fb99df142be39d2d35e7f405130c0747efac
CRs-Fixed: 2410175
Populate LTF size value from TLV for HE and modify
HE_LTF enum value to match with Firmware values
Change-Id: I8e5cce31cbae4aa8722050b06d89c933b74aa752
This rx_desc flag was removed from the function by a previous change
with id Iaa345d04b8d48814f88ed6e2237fc67696f6a20c. This flag is
needed to keep track of whether the nbuf associated with a rx_desc has
been unmapped or not. Set it to 0 when a mapped nbuf is associated with
the rx_desc.
Change-Id: I24e103c5698a556f478a3f7917076d31b5490661
CRs-Fixed: 2393056
- Check for rx_desc.magic before using a rx_desc retrieved from the reo
destination ring via rx_buf_cookie (dp_rx_process). Add stats.
- Check if we end up re-using a rx_desc with rx_desc.in_use=1 in the
replenish path (dp_rx_buffers_replenish).
- Drop RX packet and assert in case MSDU Done failures are seen in
in the TLV during dp_rx_process. Add stats.
- Disable host2rxdma ring mask for MCL. MCL is not expected to be
receiving these interrupts.
Change-Id: Iaa345d04b8d48814f88ed6e2237fc67696f6a20c
CRs-Fixed: 2382076
In some cases non aggregate packets with active BA session land in
2k_jump_error owing to sequence number mismatch.
Fix those cases by sending delba and the next addba is expected to reset
the sequence number
Change-Id: Ib95edf8744761abddb80cf28bf405d6fa52181c8
In case of target recovery clean up the AST entries on
peer before peer cleanup
Change-Id: Ib4f119ec259f0de03b0513a8f0dc8ef4b431f24a
CRs-fixed: 2419890
TX NSS field is not supported from WMB TX completion ring, remove
the NSS print in host TX stats in case introducing some confusion.
Change-Id: I594eb5cd7b2bf88de14f45ff6ad8e3248eb36617
CRs-Fixed: 2350789
Support WLAN IPA WDI 3.0 with SMMU-S1 enabled. Key design
considerations are captured below.
- ring base addresses are passed as struct sg_table and IPA
driver will do the IOVA mapping to IPA domain.
- ring doorbell addresses are passed as physical address and
IPA driver will do the IOVA mapping to IPA domain.
- TX/RX buffers need to be mapped to IPA domain by WLAN driver
via ipa_wdi_create|release_smmu_mapping API.
- IPA doorbell addresses returned from ipa_wdi_conn_pipes are
DDR addresses and WLAN driver needs to do the IOVA mapping
to WLAN domain.
Change-Id: I2398df58c9a9d002af7035f239eda37a86f5d00f
CRs-Fixed: 2422155
In Tx completion handler, nbuf was being accessed after it was
freed in vdev detach path. Fix this by moving the access to nbuf
after NULL check for vdev.
Change-Id: I08bab01bcc2f9d2e739a3897502d2d2b96bf9f43
CRs-Fixed: 2393459
Per the Linux Kernel coding style, as enforced by the kernel
checkpatch script, pointers should not be explicitly compared to
NULL. Therefore within dp replace any such comparisons with logical
operations performed on the pointer itself.
Change-Id: I61f3adab1208682d36235421f44a048e35df346d
CRs-Fixed: 2418258
Increment number of packets processed per pdev
per interrupt before updating rx histogram
stats counter.
Change-Id: I0d6c59a6a115df168f7755ce94faa90dcc753e13
CRs-Fixed: 2418384
Rx desc pools are reused across wifi up
and wifi down, only nbufs part of these
pools are free during down
Free Rx desc pools during WiFi Unload
CRs-Fixed: 2418941
Change-Id: I5cc3a82d99efd84f51f9eb048cf06da5a4495e55
Add change to skip monitor ring configuration
for smart and lite monitor through dp_set_monitor_mode
API instead will be handled to configure
cdp_txrx_set_pdev_param from caller as required
during restart, stop and up path.
Change-Id: If842399c64601049efdf9233a83ede5ce368802c
CRs-Fixed: 2421573
cdp_rx_indication_ppdu structure is used to indicate ppdu stats to upper
layer. Use the size of this struct instead of hal_rx_ppdu_info struct
while allocating SKB for notifying this event to upper layer.
hal_rx_ppdu_info struct is smaller sized structure and this can cause
invalid access.
Change-Id: I0db5a04f6e8ca8d8d38ddc83bcb7859292b45ba1
CRs-Fixed: 2419923
Compare REO ring descriptor paddr with rx_desc->nbuf->cb->paddr,
possible corruption if they differ.
Change-Id: I064df12495a9934fe2396305a7a4a8cdeacdfd4f
CRs-Fixed: 2398155
Do not print rate statistics on the console or kernel logs. Excessive
logging in kernel messages can lead to kernel panic.
Change DP_PRINT_STATS to not print in kernel logs.
Note that this is an interim solution. Ideally these prints from DP
module should be QDF_TRACE_LEVEL_DEBUG and logged to QXDM. Currently
DEBUG trace level is disabled for DP module.
Change-Id: I54af98eda0dd4ea53ae20f25ced6b03287d8b4e4
CRs-Fixed: 2414635
Compiler threw error 'defined but not used [-Werror=unused-function]'
when macro DISABLE_MON_CONFIG defined.
Wrap those functions by DISABLE_MON_CONFIG, which used only
within DISABLE_MON_CONFIG.
Change-Id: Idd92b9c481d580a6ca3df39963fa98e11d9d0753
CRs-Fixed: 2418584
For low mem config, use user configured max tx descriptors to limit the
allocation of software tx descriptors.
Use dp_txrx_pflow_update_pdev_params function to print stats and to update
pdev param.
Change-Id: I8fa6f0bb8841de68e8dc205ffcb0fde264f1b0e0
CRs-Fixed: 2414452
The rx descriptor may be invalid if RBM is incorrect.
Check the RBM value before fetching the rx descriptor.
Change-Id: I12cc99ceb31b8bbf0127591e1c275b13264eb80c
Several qdf_*_outline() APIs exist which are straight aliases to their
regular qdf_*() counterparts. Remove them, and update any callers to use
the regular APIs instead.
Change-Id: Ibe6e9cced521b93061e89690ab817ceafe49a469
CRs-Fixed: 2408267
Delay counters per TID have been implemented for following types:
1. Linux stack to hw enqueue delay
2. HW enqueue delay to tx completion delay
3. TX interframe delay
4. RX interframe delay
5. RX frame delay from ring reap to networking stack
Change-Id: I836596cbd878a43955c18b4981cb5b7b43d4df5e
In handling addba request, if we receive a request
for a tid that is in IN_PROGRESS state, we ignore
this new request.
Change-Id: I71e3f78cf43f25c184161669a359a4b703cab3cd
Crs-fixed: 2408140
