Change the qdf_timer_stop to qdf_timer_sync_cancel (synchronized
timer stop), so it can wait till previous handler is completed
and then stop/delete the timer, to avoid race condition.
Also, call this qdf_timer_sync_cancel in dp_vdev_detach_wifi3,
to be symmetry with timer start in dp_vdev_attach_wifi3.
Change-Id: I3e975554b5492ccfc684ffe8b021c62ef39bf4d7
CRs-Fixed: 2644432
In tx capture feature, the last retry packet has wrong rate.
This is because the peer is invalid at the time rate TLV is processed
for retries.
Change-Id: I0eb9a8c80f8ea0dde60fe819c7e667293630a379
CRs-Fixed: 2633669
In case of decrypt de-auth need to be sent in case of
invalid. In valid peer case no action needed
Change-Id: I355093734c20a113fc1df0306b426eeb4f92ac88
pdev_id is being initialized with 0. Since 0 is valid pdev_id, though
pdev is not present for that id, it is being accessed.
Initialized pdev_id to 0xFF by default. Added checks on API to
detect valid pdev_id value corresponding to lmac_id
Change-Id: I2b2a38783615494ccc08e265702815f7e562214b
Assign soc to dp_pdev soon after allocation, since the
soc pointer is required during deinit if pdev attach fails.
Change-Id: I87789deb82efe9f554b3310a904dd340c5eb9f7a
CRs-fixed: 2638378
Remove debug dump call to dp_rx_desc_dump() as cookie rx_descriptor is
invalid.
Change-Id: I106ebc2f872e43079abd6e6e493c90022fd09c3b
CRs-Fixed: 2638059
dp_rx_process stack frame has grown to exceed the
stack frame size of 4096. dp_rx_deliver_to_stack_no_peer
is a big function which should not be inline. Calling it
in other function increases the stack size consumed by the
caller function a lot.
Since dp_rx_deliver_to_stack_no_peer is not called very
frequently from dp_rx_process, changing its type to non-inline
function does not hit the core rx datapath much. Hence
change dp_rx_deliver_to_stack_no_peer to a non-inline
function.
Change-Id: Ib042f74c1f5a9cbe5fd947a24f004bb2fecf1fb1
CRs-Fixed: 2636365
Fix the issue on block ACK on response to BAR frame.
Parse BAR frame received from monitor destination ring and
generate block ack frame from BAR frame and previous block
ACK frame.
Change-Id: Ifc242b3850630aa25827b003a1e2078a9228358a
IPA pipes are not disconnected successfully when the last
client disconnects and despite that smmu mappings given
to IPA for the buffers are released.
Assert instead of releasing the smmu buffer mappings as
IPA has stalled and is in a non-recoverable state.
Change-Id: I9a9a35a4019726421d7435b62defe8da4907d4e0
CRs-Fixed: 2627889
In case the bus is in low power mode, the register writes (followed by a
memory barrier) may take a long time (~4ms). This can cause the caller
to block till the PCIe write is completed. Thus, even though PCI
writes are posted, it can still block the caller.
Hence, in case the bus is in low power mode (not in M0), or not in high
throughput scenarios, queue the register write in a workqueue. The
register write will happen in the delayed work context. In other cases,
i.e ,when the bus is not in low power mode or in high thoughput
scenarios, do the register writes in caller context.
Change-Id: Idf218e4581545bc6ac67b91d0f70d495387ca90e
CRs-Fixed: 2602029
Host rx return_buffer_manager should always be 4 or 6. Add check for
invalid return_buffer_manager value in ring descriptor.
Change-Id: I509dd58ddd89e6a0ce1bffa509dcfabbd0fbc975
CRs-Fixed: 2632372
Send peer_id in peer_add_wds_entry API of ol_if_ops,
so that the dp peer doesn't need to be explicitly
retrieved in corresponding offload API.
Change-Id: Ia32d31946de43e9faa1970f4e74f2810f4125995
CRs-fixed: 2629475
This change fixes issue with EAPOL frames in Tx capture
EAPOL frames completions are received with ACK completion.
For ACK completion, ba_size, ba_bitmap, ba_seqno are not updated earlier.
it was leading to dropping the frames in driver.
This fix updates ba_size, ba_bitmap, ba_seqno for QOS TID frames in ACK
completion handler
Change-Id: I58308a13c1b6addb95f424aa7ad45fe64e93b532
CRs-Fixed: 2620890
If RX packets reaped from REO2SW ring hit rx_reap_loop_pkt_limit,
REO2SW ring reaping will break and stop, but for scattered msdu case,
all related buffer should be received one time for further processing,
otherwise dp_rx_sg_create can not handle correctly.
(1) make sure all buffers for scattered msdu is received then allow
break when rx_reap_loop_pkt_limit hit.
(2) refine skb unmap location in case msdu_scatter_wait_break logic is
hit, then may double unmap for same skb(not for current issue).
Change-Id: I85d385ee9c3b1a5ed56ae5e5b68636d04968553f
CRs-Fixed: 2632082
The rx descriptor obtained using the cookie
can be NULL if the cookie is invalid. Hence
dereferencing the rx descriptor without any
validation can cause invalid address access.
Fix this by validating the rx descriptor
which has been obtained using the cookie from
the hal ring descriptor.
Change-Id: Ib584f0d8175b581d15b0e1c67d2f6ed9119ecbfc
CRs-Fixed: 2629254
Move dp monitor inline APIs to header file from dp_rx_mon_dest.c
to dp_rx_mon.h to use it in component dev
CRs-Fixed: 2626054
Change-Id: I3a31dacd7d7e817f144708f9bda5536808080864
In current monitor status ring implementation,
on pdev_attach, (srng->num_entires – 1) entries
(to keep one entry slot between hp and tp)
are replenished and last entry is not replenished to HW.
With qcn9000 monitor mode HW enhancements, status and destination ring
can be made lock-stepped.
for qcn9000 lock step is achieved by making monitor status ring
follow the monitor destination reap for a PPDU
However in existing flow during attach monitor status replenish logic
do not fill last entry but is filled up during first subsequent reap.
for first ppdu, i.e. after reaping destination ring,
when status ring is reaped, as first entry (hp = srng->num_entires – 1)
in status ring is NULL, so lock-stepping is not achieved.
To address this issue for qcn9000 as well as HK:
a. Replenish last entry in monitor status ring during attach
b. Modify src srng peek API to peek it from hp+1 entry
c. Introduce new HAL API get cur desc and move next
CRs-Fixed: 2626049
Change-Id: I7da6cd6006243cbca2151da22a31a6f5ed0015f3
Peer key security type is set in cdp_set_pn_check,
this API is called in key install step, but only
update Ucast field even if in Mcast key install.
This may be set wrongly if Ucast and Mcast security
modes are different. Add a new API to set
security modes of differ key types separately.
This change should be reviewd in associate with
review #3036742
Change-Id: Ib9723416fbedbe3dce61224928d51ef203cf7068
CRs-Fixed: 2623114
Macro ATH_11AC_TXCOMPACT is defined, so CE4 sending completion will
not free the related htt_htc packets. Currently it depends on
htt_soc->htt_htc_pkt_misclist to free them once this kind of htt_htc
is no longer sitting in CE4 source ring which means CE sending
completion is done already. misclist will free the old/tail of
htt_htc packets if there is > 256 htt_htc packet is pending in list,
but CE4 source ring has 2048 entries, for extreme case, if CE4 sending
is blocking and > 256 htt_htc packets is pending in CE4,
misclist will free the tail htt_htc packet, but later CE4 sending
completion is coming, invalid access to htt_htc packets will happen.
Enlarge misclist trim threshold > 2048 to ensure that when free
htt_htc packet in misclist, the htt_htc packets has done sending
completion already.
Change-Id: I2d86f68485b1d45a5ea4edb6b952f746700df54f
CRs-Fixed: 2609903
Move RCC specific debug statistics to Data path
to avoid allocating nbuf for the PPDUs for which
MAC has not sent a Freeze TLV. Instead just update the
debug statistics at an earlier point irrespective of
MAC sending Freeze to PHY, to gather below information:
Status of number of PPDUs being captured by PHY
Status of MAC sending freeze TLV for the PPDUs
Update freeze reason for the number of PPDUs received
Add support to clear debug statistics at user's will.
Change-Id: I8def3ce1f3ecc0b169030704db3fb1fb1c0a20ee
CRs-Fixed: 2609604
Intercept monitor destination path for NDPA frames,
use it to generate noack frames for tx capture.
Change-Id: Ia54196bd84729a7863954b6deec7e02b52f23556
CRs-Fixed: 2605672
Currently when checking for duplicate peer, only the vdev
id on which peer is going to be created is checked and not
on other vdev ids. So correct it by iterating over all
vdev ids.
Change-Id: Ie05ecc125d3285a522021b630fc7244501b6a797
CRs-Fixed: 2629146
Handle minidump logging using dynamic
configurablity options. Data structures
to be logged in minidump can be configured
using internal INI file.
Change-Id: I99f12b3f98c4a9c0e15c3e5d611019e6b8d0909a
Sometimes ack ba status tlv comes twice which cause ppdu desc to be hold
as waiting for some more tlv as mpdu ack ba tlv is not matching with
completion common tlv mpdu count.
Change-Id: Id946ecd53018ee7abe93aa6c5f62e69d41b750e8
PDEV was being obtained using lmac_id by directly indexing the
pdev_list array. Instead, we need to use dp_get_pdev_for_lmac_id.
Change-Id: I1c4a0f3df5db59390e17666a5f712c5412e22bb1
CRs-Fixed: 2627909
Add doppler and spatial reuse in ppdu descriptor to be able to use it on
HE information for tx capture
Change-Id: Id87a692a89d00c64929624c22c2efc27bc873e43
We are seeing a invalid memory access crash
in the dp_get_pdev_for_mac_id call from
dp_rx_process_invalid_peer, due to invalid mac id passed,
probably due to some stack correction.
We should instead use dp_get_pdev_for_lmac_id from
dp_rx_process_invalid_peer, where for invalid
mac id, we assert.
Change-Id: I0737132b5bbdd2fcbdb714d4643a69184ae3821e
CRs-Fixed: 2618432
There is possibility that DUT receives non-aggr frames
with non-aggr SN during BA setup during BA session
setup. SSN updates when ADDBA req is processed, should
also update REO SSN one more time when ADDBA resp sends
This will let REO to be able to receive frames in BA
session with new SN.
Change-Id: I4791bcd6f1ede90615c883d31426fcdecd50a631
CRs-Fixed: 2623505
Add code to move enum _ol_ath_param_t out from common DP code
and move it into OL layer
Change-Id: I92fa5f52c581a18449ab42da41ff5f6f1a2bc595
Crs-Fixed: 2560866
Currently mesh mode uses channel numbering which is derived
from APIs that don't support 6GHz channels numbering due to the
overloading of 6GHz channels with 2.4GHz and 5GHz.
Add support to obtain the correct channel number (and auxiliary
information like band and frequency) through the new APIs that
support 6GHz.
Change-Id: Ib0b39ebae2a22bd6b2b5d17b9058c3c2100e0d59
CRs-Fixed: 2605229
Packets delivered to FISA via exception err path doesnot have TLVs.
FISA handling requires additional TLVs. dp rx core handling
skips TLVs, save TLV length info in nbuf->cb so that TLVs
are recovered back in FISA.
Change-Id: I53fab2e19abcbf82697ea6f53a4ddf3ea0dd0699
CRs-Fixed: 2620844
Currently if peer reference is reused valid flag is not set.
In peer delete function if valid flag is not set, function
returns with out decrementing the ref count and also not
releasing the local id. This leads to local id exhaustion if
back to back peer reuse happens which is common in case of back
to back roaming to same AP. As a result connection to AP fails
due to unavailability of local id. In this change set valid flag
if peer reference is reused.
Change-Id: Ibeccbc4d33aac9e4285b48a4b8abdd2276fd7ede
CRs-Fixed: 2624798
For Disassoc frame, fw send ppdu stats to host with peer_id
of the previous associated client. peer id is valid and peer
structure is NULL making framectrl as zero for disassoc frame.
with the fix frame ctrl are populated before checking for invalid peer.
Change-Id: I5960685a4c90f00a850f4f1fa2b3d883117fb443
Check the max value for max desc extracted from cfg
before using it for actual allocation of tx descs.
Change-Id: I9bd6c88f136870a42596afbf25328965fce959be
CRs-fixed: 2614329
Break up the 2MB descriptor bank memory allocations for WBM
idle link ring. Use multiple page allocation and populate the
WBM idle link descriptor ring with physical addresses of each
DMA page allocated in the descriptor bank.
This is to ensure that no requests for contiguous memory
allocations are made that might result in allocation failures.
For MCL set the page size to 4KB and leave it to max_alloc_size
cfg ini param for WIN specific code.
Change-Id: Iec30321044827c0174366cc02df25a42d38309e0
CRs-Fixed: 2565817
Host queries for reo hw stats by sending reo queue status command
on all the rx tids. After sending host waits to receive for all the
rx tids reo hw stats. If the hw stats reo status is not received for
all the rx tids with in the specified timeout value corresponding
peers's reference is released. In parallel if disconnect happens
from userspace, there is always a chance to access the freed rxtid
reference in rx hw stats callback. So in case of rx hw stats event
timeout ignore processing the reo command status rx stats callback.
Change-Id: I0aba346a25564a3c29751f1f5a1d4e9204b93d33
CRs-Fixed: 2619720
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: 2596297
