This is to enhance statics for runtime put/get, which is
to detect if there is mismatch for usage_count.
Change-Id: I24cddb9d10e4cb675c8375cbd0f589c7718bd680
CRs-Fixed: 2647972
As per the REO logic, if a packet is received with
SN <= current SN, then it will be treated as a 2K jump
error or OOR and the packet in consideration will be
dropped by REO. For NON-BA case this case will be
treated as 2k-jump error.
For the NON-BA case, the packets with SN <= current SN
should not be dropped.
The current REO configuration, sets the BA window size
to 2 for NON-BA case, which in turn enables 2k-jump
detection.
For configuring the REO to not drop packets with previous
SN, we need not set the BA window size to 2, thereby
disabling the 2k-jump check for NON-BA case.
A. SN = 1, 2, 3, 4, 2096, …
(good case, as long as the SN is within 2K range)
B. SN = 3, 5, 2
(3, 5 are good packet, 2 is bad packet and will
be dropped to ‘2K error’ with error code = 2K error.)
(note that this is for non-BA session, for BA session,
we detect SN as either
2K error or OOR error based on SN and window size).
After this change, we will treat this as a good packet.
C. SN = 1, 1, 1, 1, with duplicate detect enabled – these are duplicate
packets and will be dropped to ‘DD queue’ with
error code = DD
(No change)
D. SN = 1, 1, 1, 1, with duplicate detect disabled – packets will be
dropped to ‘2K error’ with error code = 2K error.
After this change, we will treat this as a good packet.
Skip the setting of BA window size to 2 for aggregated
packets in NON BA case.
CRs-Fixed: 2580605
Change-Id: I19d5eced7c8730a9c3820fd6fc69923d2a98263a
Add debugging infrastructure to record every event posted to reo
command ring. The infrastructure maintains the record of the last
64 events posted to the ring.
Change-Id: Id56fc352050eb664a64b0abb767f3b4a6b4c3aa3
CRs-Fixed: 2552822
When txrx_stats 28 is issued we post messages to reo command ring
to receive reo queue stats on the reo status ring.
This leads to acces accessing reo_cmd_ring head pointer shadow
register when runtime pm is suspended. Perform
a hif_pm_runtime_get to resume the apps before accessing
the shadow register.
Change-Id: Ie6df817fc114bd4c8ba7783df1fe3bd770deec40
CRs-Fixed: 2544174
Currently after runtime resume all SW2TCL data and reo cmd
srng rings hp and tp value are flushed. In case of IPA
offload case SW2TCL3 righ hp value will be updated by IPA
and not by host. In case of runtime pm enable host is
setting the value to zero as part of runtime resume which
results in incorrect hp value of SW2TCL3. As part of this
change set flush event for rings which are accessed by host
during link down state and after runtime resume flush the
rings for which flush event is set.
Change-Id: I5c9afa708277cf3a6e6d5ef99447bc21f88cfdcf
CRs-Fixed: 2514621
Disable PN error handling in REO Queue descriptor setup.
This is different from pn check enable.
When PN error handling is enabled, if a PN check fails
REO queue descriptor sets pn_error_detected,
All the subsequent packets will be exceptioned with REO error code 13
till SW resets the pn_error_detected.
SW is not doing any special handling when PN error handling is enabled
Hence disabling flag in setup time instead of resetting/updating
reo queue descriptor as update is an uncessary overhead.
Change-Id: I21e40b6f685e50630fcc4413f6cd27e5ed9ff66d
When the scheduler thread initiates the WMA_SET_BSSKEY_REQ
we send CMD_UPDATE_RX_REO_QUEUE to REO srng. This is done by
posting a descriptor to the reo command ring and then we
update the HP so that the HW can consume the descriptor.
Avoid accessing HP shadow address when we are in runtime
suspend state. Perform a hif_pm_runtime_get to resume the
link and access the shadow register and once done initiate a
hif_pm_runtime_put to allow device to go into runtime
suspend.
Change-Id: I24c3e046a5769f03a0f1969360cccdbe55b81d45
CRs-Fixed: 2495720
Add code to replace usage of void pointers from
HAL layer and instead use appropriate opaque pointers
Change-Id: Id950bd9130a99014305738937aed736cf0144aca
CRs-Fixed: 2487250
Add code to remove void pointer usage for hal_srng
and use opaque pointer dp_hal_ring_t instead.
Change-Id: I6907f7376d7fe3c9180b8795bd96f49fead2ec64
CRs-Fixed: 2484404
Add code to remove void pointer usage for hal_soc
and introduce opaque pointer to be used intead of void
from dp layer into hal layer
Change-Id: Ia38571174c6ed79558d0f0c9cd1a0f4afaa66483
CRs-Fixed: 2480857
To avoid race conditions where BA window size can change, always
flush entire REO descriptor from HW cache irrespective of current
BA window size.
Change-Id: I608996722e7dc2dc6acfd145b8c190b58ce09822
CRs-Fixed: 2251811
Some of the macro names defined in qca8074v1, are defined with
a slightly different name in qca8074v2, and few macros have the
same name in both headers but are defined with different values.
Fixed the same.
Change-Id: I5e948baf5326d1d8fdfa2bd7ee8aa072c710d17c
Set 'need_status' while sending REO queue update command
for deleting REO queues to get status from HW, which will
trigger callback to cleanup and free the queues.
Also fix issue in initializing REO command descriptor, causing
command number to be cleared.
Change-Id: Ie30fc80f76d835e239dd0b3d3e78c7f32bfa09a4
Upon receiving 2k jump exception, send delba
and track delba tx status and retries.
Change-Id: Ida35256233869dfa390c40030c9296b9c48481ce
Crs-fixed: 2239856
Clean up datapath component prints by correcting trace levels for
regularly occurring prints and removing newlines from converged
print APIs since qdf_trace_msh appends them by default.
Change-Id: Ie8fe319fcb737720f7400a165e134b6a18bd15b5
CRs-Fixed: 2243843
macros and fw cmn headers included in hal_internal.h is moved
to a new hal_hw_headers.h to avoid these getting included in hif
layer. Include the new hal_hw_headers.h file
Change-Id: I3390ae9e71ca561d94a265e1fdf4d9354e7b2474
When delete peer, host will try to flush all frames in REO queue
by sending a cmd through reo ring, but sometimes encounted failure to
send this cmd, dump ring status if ecountered above failure.
Change-Id: I3ea4e96e5999f85398b531ddf4f350e91e798d70
CRs-Fixed: 2167419
HAL REO flush cache command should use block resource only if cache
blocking is requested.
Change-Id: Ie52634cdb9908e4dfa12495dfc3d87a9d5d54c0b
CRs-Fixed: 2057946
Use normal allocation instead of alloc_consistent for REO descriptors to
allow freeing from interrupt context.
Also fixed an issue in ba_window_size setting in REO queues.
Change-Id: I91b06e04c0343eb7fe8580d8655c6bc5e33cfe06
Fixes for issues seen while sending multiple REO commands:
- Fix bug in error return from hal_reo API.
- REO status for updates done as part of peer deletion are sometimes
received after peer is deleted, causing crash due to invalid access
in REO command status handler. Fixed by using temporary copies for
rx_tid structures.
- Zero out REO command descriptors before setting up new commands.
Change-Id: I63409eb74fd91a21e0264fb2b3a62efb3a91bfe6
