Lithium HW can parse each MSDU when packets are enqueued to transmit
classifier and each MSDU can get enqueued to different MSDU-Queue.
By default, there are 2 queues per TID – UDP and NON-UDP. Based on
the flow parsing, packets will be queued to either of them.
Enterprise customers uses RAW mode frames which transmit classifier
would not be able to identify L4 protocol for UDP or NON-UDP queue
for encrypted frames.
Customer can also have different application based packets which needs
to be enqueued to differently prioritized queues. They can choose to
tag each MSDU based on application DPI and derive the flow priority
as below –
• UDP
• NON-UDP
• HI-PRIO
• LOW-PRIO
Customer needs API support from host which will provide an ast-index
based on a particular flow id given peer mac address. Based on the mac
address, host will find the associated peer and return the ast_index &
cache set number corresponding to the flow id from peer flow based ast
entry table. Customer will then program this ast-index and ast override
in transmit classifier CMD ring to enqueue MSDU to a specific
flow-queuein TQM.
Host API:
Input parameter – uint8_t *peer mac address, uint8_t flow id
Output – uint16_t ast_index
Change-Id: I79557c31532514ad5ec0c1848827a3e5cd91cdf2
long length msdu is received and looks this msdu is spread across
multiple nbufs, there is no corresbonding logic for this case.
qdf_set_pkt_len will invoke pskb_expand_head to renew skb->head
buffer, but the rx_tlv_hdr is still pointed to original skb->data
buffer, invalid accessing will happen.
As a WAR, drop this msdu related nbufs after dp_rx_sg_create is done.
Change-Id: Iceb09fd04e4d768325018a8ddd4261ab4f75991a
CRs-Fixed: 2597927
Crash scenario:
(1) frag data A is dropped and related RX desc A is replenished
and reused, but pdev->free_list_tail is still pointed to RX desc A.
(2) frag data B/C is coming, defrag fails then pdev->free_list_head
will point to B-->C RX desc, but pdev->free_list_tail still point to A.
(3) for defrag failing case, host only will replenish 1 RX buffer for
current case, RX desc B is replenished, while C will be free back to
RX desc pool.
(4) dp_rx_add_desc_list_to_free_list will set RX desc A-->next =
free_list, free_list point to C insted.
(5) when step (1) RX desc A replenished buffer indicated to host by
REO2Dst ring, RX desc A -->nbuf actually is pointed to another RX
desc, invalid skb accessing will happen.
Solution:
a. reset tail pointer in dp_rx_add_desc_list_to_free_list at last.
b. reset tail pointer same as head in dp_rx_add_to_free_desc_list
if head->next is NULL.
c. set correct rx_bufs number for replenish when dp_rx_defrag fails.
Change-Id: Ib297baea3605a09dd7d85d1f5ceb95db48a2e1f1
CRs-Fixed: 2603676
Use lmac_id to get the rx_desc_pool while freeing
the rx_desc_pool for rx_desc_buf.
Change-Id: I549e7f54b267d2c82dceb391a766297dfe18ceb7
CRs-Fixed: 2608209
Incorrect size is given while copying pdev stats.
Fix this by properly passing pdev stats struct
while calculating size of pdev stats.
Change-Id: I4b4be2542d092c5c984252a62e2406cb1e83bcd8
CRs-Fixed: 2603674
Add ast-type as argument in wds peer create command.
This helps to identify the ast entry from the time when ast add is done.
Change-Id: Ia0c073cf8f2573608e6a84410dffee6edf9b064b
Provide an API to fetch required individual peer stats
from the data path peer stats instead of copying the
whole peer stats structure.
Change-Id: I1c9619bc7d9bb0e3e3df6d38cc5f4570869e31fe
As part of Dynamic mode change feature, rxdma_mon_status_ring
has been moved from dp_pdev to dp_psoc. Adapt CFR to this change.
Dynamic mode change feature change -
Ib017d49243405b62fc34099c01a2b898b25341d0
Change-Id: I43ee4e08ad70614d042be3e08c0add66e7e011f9
CRs-Fixed: 2606001
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: I4a41211464a30e018e8b73b8e9b91c7a509327b4
CRs-Fixed: 2541711
If the vdev is marked as delete in progress, then the
packets received on that particular vdev should not be
submitted to the network stack.
Drop the packets if the vdev is marked to be deleted.
CRs-Fixed: 2599128
Change-Id: I0996669bea368b6c0afc10e1d929bc2f314ae6fe
Peer id field is set to HTT_INVALID_PEER when failed to
populate cdp rx indication structure.Set num_user to 0 in
such cases.This avoids out of bound access during rx stats
update.
Change-Id: I112713deee1ec3cb9c37ba3571aff05fb816c76d
CRs-Fixed: 2596779
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_get_psoc_param
- wdi_event_sub
- wdi_event_unsub
Change-Id: Id089d9b6b4737d700d2436e2081291a3741affb5
CRs-Fixed: 2541658
Change set 2 of ctrl_ops APIs to replace pdev, vdev and peer
dp handles with pdev_id, vdev_id and peer mac address
along with dp soc handle
Change-Id: I3f180c9c360d564f0b229b447074ad23b7c0a737
Change set 1 of the ctrl_ops APIs to replace pdev, vdev and peer
dp handles with pdev_id, vdev_id and peer mac address
along with dp soc handle
Change-Id: Iedf6dea2215d3d04b2212c71c8ec8c7328c559e8
Since delete in progress for peer is set in roaming cases
explicitly, do not return from peer teardown after checking
this flag.
Change-Id: Icc514d9500075409d907546a26b2725b5334cbce
Add the datapath changes for supporting
the newly added IPCI bus type for QCA6750
CRs-Fixed: 2597326
Change-Id: I27e3e9f746eb383e85c7345150bda03320d041b9
Currently host writes to TX doorbell register after ipa connect
pipes to indicate number of free TX buffers. In case when IPA
clocks are off and host tries to access the register will leads
to invalid access. So access the tx doorbell register after ipa
enable pipes which makes sure IPA clocks are always on.
Change-Id: If11fdc9c408ebf7f22eef6a020b507bc39ee6a8e
CRs-Fixed: 2599029
multipass_rx_pkt_drop is peer level stats counter
used to count multipass rx packet dropped frame.
Accumulate this counter at vdev and pdev level.
It also initializes multipass_en flag to false at
vdev attach.
Change-Id: Idaa85a71c80eefb9359abb026402b71aa28ad6a2
CRs-Fixed: 2595551
1. Move all LMAC rings to SOC from pDEV
2. Dynamically obtain lmac->pdev mapping while handling LMAC interrupts
Change-Id: Ib017d49243405b62fc34099c01a2b898b25341d0
Add monitor mode support to capture packets over 6GHz frequencies by
getting capture frequency from pdev.
Change freq type to qdf_freq_t.
Change-Id: I7b6edc43e254dc98a3c2939c369874bec9d16ddd
CRs-Fixed: 2568970
the current HAL API is to read the WBM internal error
bit from the wbm release ring descriptor is always taking
HKv1 HW structure. But the wbm_internal_error bit
placement has changed from HKv2, for this reason we have
to use target specific HAL API.
Change-Id: I44789180754ca21ae59650b6d8620321a1f12569
CFR has specific modes where status TLVs from non-associated
clients will be used for correlation with CFR data. In the
common API used by STATS and CFR, add an exception for CFR to
collect the necessary PPDU information. STATS on the other hand,
does not collect information from non-associated clients.
Change-Id: I2987010ae84ff02a25bfce5cc77bbc3e5383560b
CRs-Fixed: 2598547
CFR information extracted from PPDU status TLVs is already stored in
HAL ppdu structure. Extract the same and copy it to CDP ppdu structure.
This is delivered to subscribers through a WDI event, similar to PPDU
stats.
Change-Id: I4315626c7f79f85b75b8d1b9e1e5caf8d65abed9
CRs-Fixed: 2593416
No need to protect with vdev list spin lock for dp_rx_vdev_detach and
which might wait event from dp rx thread when doing the RX GRO flash.
Do the same for dp_tx_vdev_detach.
Change-Id: I8f191ed62a174a9a88999c36d47ef18f39e8bd82
CRs-Fixed: 2600680
Add code to process lmac rings in polling.
Lmac rings can be switched between polling and interrupts
by using lmac_polled_mode flag in dp_soc.
Change-Id: I231e906930c9b42100a2df3e52acac97deba1c34
Created a new interrupt mask for msi chipsets so that both ahb
and pci chipsets can work together.
Change-Id: I78d1b9641e58f07d1582152c691c5ae29e6ddb03
Setting interrupt mode in dp_soc_init as setting it in
dp_soc_attach is resulting in interrupt mode not set on
subsequent wifi down/up. soc_attach is called only in case
of wifi unload/load and not in case of wifi down/up.
Adding interrupt mode setting in soc_init helps in setting
correct interrupt mode in case of wifi down/up.
Change-Id: I9e6ad94e42b7bc2bf84f832ece8dc8779e335905
There is excessive console logging in case of peer or vdev
null in mic error handling and wbm err processing in softirq
context which is delaying the interrupt to get enable and as a
result host is slow in reaping the wbm2sw3 ring. This leads to
wbm2sw3 ring full back pressure at FW and finally FW asserts.
As part of this change reduce reduce the error logging level from
err to info and for MSM info goes to the wifi logger logs and not
to the console.
Change-Id: I9b8be1feee5e807cecbe307b24a323e2b0ae631a
CRs-Fixed: 2596656
If two back to back HTT msg sending happened in short time,
the second HTT msg source SRNG HP writing has chance to fail,
this has been confirmed by HST HW.
For monitor mode, HTT msg for monitor status ring configuration
is the last msg for sending, if the 2nd HTT msg for monitor status ring
sending failed, HW won't provide anything into 2nd monitor status ring.
as a WAR, add some delay before 2nd HTT msg start sending, > 2us is
required per HST HW, delay 100 us for safe.
Change-Id: Id2a5a8798fde79267f9d6f9ba23d8ed54a400aa7
CRs-Fixed: 2597248
Currently in lithium FW is not always in datapath. As a result
can not provide all the wlan vendor extended stats. So as part
of this host provides some of the extended stats.
Change-Id: I876b6e905f7e25b7088f80f07e55e26a7a409241
CRs-Fixed: 2594797
There is a chance of leak of RX buffers if peer disconnects
while we are in middle for processing the list of MSDUs in a
AMSDU
Change-Id: I0081ec96da95ea570903dbd5d91c866c8c141667
Add a check to validate invalid_peer_head_msdu before accessing
to avoid NULL dereference.
Change-Id: I9218bdd1100b48a32240546f380b1437ae72c406
CRs-Fixed: 2585651
Channel Frequency Response(CFR) captures FFT bins corresponding to a PPDU
received and this is relayed to user space application. Since FFT bins
alone will not be meaningful, additional information about the PPDU such
as Peer addr, BW, NSS, etc are extracted from multiple PPDU status TLVs
and passed on to userspace along with CFR data.
Host RX monitor status ring is configured during CFR enable and the host
subscribes to relevant PPDU TLVs that provides all necessary information
for CFR data correlation.
Change-Id: Ia8080fe96c330db42d371634ba4363eb36c05402
CRs-Fixed: 2593416
The printk buffer is protected by a global raw spinlock for readers
and writers. This restricts the contexts that are allowed to
access the buffer. So large latencies exists when printing to
kmsg done with QDF trace levels WARN, INFO, ERROR. This results
in roaming KPI for maximum transition time exceeding the expected
100ms.
So change info/err prints to debug prints in roaming path
to reduce delays in roaming time.
Change-Id: I9cdf043da6e88d735ab39812e74ed24784ceede3
CRs-Fixed: 2594867
crash scenario:
a. dp_tx_vdev_detach will reset the vdev of TX desc to NULL
by dp_tx_desc_flush.
b. in the meantime, if TX completion is coming and when all TX desc
is recycled, the pool will be freed if pool status is invalid before.
c. invalid TX desc accessing will happen in (a).
add TX desc pool lock protection in dp_tx_desc_flush.
Change-Id: I65c570aa6a3044a478dbe51bbf396cf24612f675
CRs-Fixed: 2595755
REO CMD send failure floods the console with error logs.
Ratelimit the error log for REO CMD send failure
and increment stats to keep track of the number of
failures.
CRs-Fixed: 2593996
Change-Id: Id415a17b1ba1f4c044bf34eb31e81e2cea825f80
Currently wlan host will re-inject defrag data with RBM 6 to
REO, this data buffer will go to REO-->REO2SW4-->IPA-->FW2RXDMA,
fix below issue introduced by this RX buffer path.
a. FW assert due to FW2RXDMA DMA address not 4 bytes aligned.
b. host skb double allocation due to qdf_nbuf_linearize() for frag skb.
c. Invalid RBM 6 for fragment RX due to RX buffer resue.
Change-Id: I36d831fc14b6b9aa0cea32682823de348f7eecd3
CRs-Fixed: 2591453
In vdev to vdev roaming on HKv2 we can have dp peer on
different LMACs with same mac address. To get the right
peer on PDEV use id match while searching from hash table
Change-Id: I60b4d7018120c9bd6a56913822ce42f24598184a
