Currently, in monitor mode, beacon frames are showing an incorrect phy
rate (0). This is because the values in the RX descriptor for
interpreting the legacy rates have changed in ihelium.
Make changes in host to interpret the new values into data rates correctly.
Change-Id: Ibf845666943fd718b657c0128758dabfb64a38ef
CRs-Fixed: 2082456
* Log EAPOL, ARP, DHCP, MGMT, ROAM Events, ICMP, Data packets,
TX-Completions by default.
* Set default verbosity as low (existing is high)
* Throttle ICMP and Data packets if threshold exceeds.
* Enable live mode by default. Added ini entry to disable
it and configure high throughput thresh - gDptraceConfig
* Log only SA and DA for proto packets (and not 32 bytes)
* Format prints to fit in 1 line.
CRs-Fixed: 2051134
Change-Id: Ia94732e49023950f458a7402650d4b8f5e1d5563
Pointer head_buf returned from call to function
htt_rx_offload_paddr_msdu_pop_ll may be NULL and will be
dereferenced
Change-Id: I6e1db520d525a42d59ebfb675325e37c4097ad71
CRs-Fixed: 2085696
RX hash lock is not unlocked in case hash_table is
NULL and context is returned immediately without
unlocking the rx hash lock.
Unlock the rx hash lock before returning from the context.
Change-Id: I4ec2c272ac92cf5c0e93c64e34ce96cafe5cc11c
CRs-Fixed: 2104297
When DUT connects to specific AP, it receives the frame whose
SA is self MAC and MIC failure flag is set at the same time.
This will cause disconnection with the AP because MIC failure.
To fix this unexpected disconnection issue, do not check MIC
failure if discard flag is set by fw.
Change-Id: I9029da9d329679bf0fd61be234690c764cceb57c
CRs-Fixed: 2090575
In the case of low memory allocation failures, the system should be able to
continue functioning after the connect_service failure.
Change-Id: Icd054d541aa95dbf4099426cf73f3dfe6c2c4565
CRs-Fixed: 2097391
During target attach procedure, there are a number of messages sent
to the FW. Some of these sends did not check return codes. Add
return code checks and treat the codes accordingly.
Change-Id: Ib515e040b6c7e2ecc23d2d422bb359e83342fcfa
CRs-Fixed: 2064904
When replenishing Rx buffers driver updates the address of the
buffer and the index of rx buffer in rx ring to the firmware.
Make sure alloc index write is reflected in the memory before
FW polls the remote ring write index as compiler can reorder
the instructions based on optimizations.
Disable compile time reorder by adding a write memory barrier.
Change-Id: Icb37598e6c36b4a128b1bcd91bfb9d643bf083e2
CRs-Fixed: 2084210
The target_delta and bus_delta combined with target_tx_credit
to fulfil the credit function of host.
A global lock is needed to ensure reenterable.
Change-Id: I71c006ccfe05b442706283490cd7bba4cc8a6f5f
CRs-Fixed: 964530
When HTT attached, driver allocates DMA buffer to record each RX ring
entry, with element size - target_paddr_t, which is bus size, and may be
different from kernel's physical address size.
While in htt_rx_detach, size of qdf_dma_addr_t used for buffer free.
In case kernel built as 64-bit, and bus size as default 32-bit,
it will crash when free.
Use same element size to avoid it.
CRs-Fixed: 2087592
Change-Id: Iabe95f92ea82b736213a8b81453163229b2a3a3b
warning: implicit conversion from enumeration type
'qdf_nbuf_l4_rx_cksum_result_t' to different enumeration
type 'qdf_nbuf_l4_rx_cksum_type_t
Change-Id: If2b7ee179fa260c14cde450cc72272c41a9f4adb
CRs-Fixed: 2055487
Rx hash deinit sets hash_table to NULL, at the same time
there can be active tasklet context accessing the rx hash list.
Prevent rx hashlist access after rx hash deinit and set the netbuf
to NULL once netbuff is popped out of the list.
Change-Id: I4e30dd69ece33c3cc768842274d5307c0bf29a37
CRs-Fixed: 2049121
The function is not present as inline in the non debug build, which will
cause a compilation error.
Fix issue by having an empty inline function in the header file.
Some counters need to be defined outside of DEBUG_RX_RING_BUFFER.
Change-Id: I755616aa55c2f38bca017137e91cc469de7fbdcf
CRs-Fixed: 2004927
Get tx/rx tsf time for each msdu, and then
time stamp the msdu with the corresponding
host time.
Change-Id: I7af80318b60c28d53e95fbe545e82da4191544b6
CRs-Fixed: 2057693
If do not define QCA_WIFI_3_0, it alloc ipa uc tx buffer by
qdf_nbuf_alloc, but free by qdf_mem_free_consistent, which cause
memory leak. Solution is freeing with corresponding method.
Change-Id: I407712e545acfe5aeef8c8afa58785e42556ca02
CRs-Fixed: 2061922
implicit conversion from enumeration type 'QDF_STATUS' to different enumeration type 'A_STATUS'
implicit conversion from enumeration type 'A_STATUS' to different enumeration type 'QDF_STATUS'
CRs-Fixed: 2063339
Change-Id: Ib96a49f40ea581127405560087354e7fa9bf46ec
ipa_uc_rx_ring_elem_t structure is platform specific and current
definition cannot be used on 32-bit ARM platform where IPA 3.5
is attached. Add IPA 3.5 and 32-bit ARM compatible structure to
avoid IPA crash when data traffic is running
Change-Id: Ia5c141c9405675a64f59ba4c09f1ffd911158c9b
CRs-Fixed: 2046905
Increase packet header bytes to 256, HL1.0 target offload features on
iHelium FW requires header bytes to be increased from 64 to 256bytes.
And round up the HTT_RX_BUF_SIZE to CACHE_LINE_SZ.
Change-Id: Iec45f5747956d0797411f76c2fec1368a13e7d6d
CRs-Fixed: 1039073
When commands to Firmware fail during probe due to Firmware being down, in
most cases QDF_BUGs are hit, not allowing re-probing to take place. A new
API is defined to check if the failure is due to Firmware being down or not.
If so, return error gracefully allowing re-probing to take place
Change-Id: Ib2f21a6f1ced1c16e9d2bdbda810b160ff63e2c6
CRs-fixed: 2027444
Fix replenishing logic by taking care of fragments as well
in monitor mode. Also drop the packets with mpdu_length_err
bit set in rx attention desc and don't send those packets to
upper layers.
Change-Id: Iaeeb0e40e4988ca5a2d4c5e7c98ed3f82dfca533
CRs-Fixed: 2037633
Add FEATURE_TSO flag in Kbuild to enable the
TSO feature for Lithium based chipsets
Change-Id: Ie26b2b9a191828b3ce833000a7abbb09dd492713
CRs-Fixed: 2042950
1. Add an extra parameter pdev id to
qdf_trace signatures to avoid compilation break.
2. This variable allows to collect records per pdev
for Lithium based chipsets.
3. Set this variable to a default value QDF_TRACE_DEFAULT_PDEV_ID
for older chipsets.
Change-Id: I2f4c2e3525ac0d1d53f1af3db7e43acadf6ab835
CRs-Fixed: 2049412
Handle error case if DATA2/DATA3 htt service connection
with fw fails.
DATA2/DATA3 connect service is done after FW ready indication.
Hence it's safe to handle the error scenarios.
Change-Id: If6f07f3c29cbc5f1486231a7897d1223ad1c24bb
CRs-Fixed: 2034843
Modify the legacy LRO implementation which is in HDD to
use the implementation in QDF instead. This is to avoid
code duplication and to unify the LRO implementations for
Napier and Helium.
Change-Id: Ie9d3bb6d3dcda5a7d1d5fda6e7cbdef92be30461
CRs-Fixed: 2042812
Do not typecast the physical address passed to hash list to
32 bits. The physical address maybe greater than 32 bits which
will lead to a mismatch during look-up.
Change-Id: I095e1e63c59da74bc8a2507273e7c6487d0b19e7
CRs-Fixed: 2044161
Limit the detection of non posted rx buffers from FW only to
64 bit DMA address system. For systems, where IPA has 32 bit
DMA address limitation, this debug feature should be disabled.
CRs-Fixed: 2010583
Change-Id: I9a26d0a423acedb068d99f39929029182f7c6fa1
TX BUF POOL vaddr storage alloc is not freed in
htt_tx_ipa_uc_attach when paddr storage allocation fails.
Fix is to free memory in failure case.
CRs-Fixed: 2034943
Change-Id: Ifb26dfdcff19b3ab66f11a89b9c185777eeec535
In monitor mode, while filling the radiotap header, rate and bw
info for HT and VHT frames was not filled correctly due to which
packets in tcpdump are showing incorrect rates and bw for HT/VHT
frames.
Fill rate and bw info for HT/VHT frames correctly in monitor mode.
Change-Id: If15708101ac123fbdba141b04aad881d988e6b43
CRs-Fixed: 2034455
Currently, rx-hash calculation and hash look-up is performed only on
lower 32 bits. On some systems with a particular memory map, we have
observed that one could have buffers coming from memory map regions
that only differ at the higher (>32) bits.
For such systems, make sure that the hash entry storage and comparison
operations are done on the whole paddr range.
Change-Id: I279f673ee465a811d20606b260be084f984135d5
CRs-Fixed: 2024095
Add support for 32-bit host architecture where dma_addr_t
is 32 bits and bus addresses are of 64 bits on MAC and IPA.
Change-Id: Id4a4763ed388973297c63df290b578effdc5b284
CRs-Fixed: 2006548
As kernel was complaining an order-5 (2^5 * page-size(=8K)) allocation,
change the allocation of rx_buf_debug memory debug structure to be done
from BSS instead.
Change-Id: I78658f2f723d230cc6443c0897a6a4b31c753849
CRs-Fixed: 2011052
Enhance htt_rx_ring_fill_n by:
- Start timer in case the function is not able to find memory to fill rx
buffers.
- Handle case where one stream can starve the other stream by constantly
adding to the debt.
CRs-Fixed: 2001325
Change-Id: Idff33d9c28daaf7a00334134459a3633f5ccde7c
The offending code did not get the order of operations right.
It was: shift 16, cast 64,shift 16 and should have been:
cast, shift 16, shift 16.
CRs-Fixed: 2000844
Change-Id: I65cf4aaf0c20535070eb63b15d77b8c3c7dcb4e4
Add detection of non posted RX buffers from FW (rx_in_order_ind msgs)
by marking the higher(16) unused bits of the 64-bit physical addresses.
On reception, if the paddr received does not have the markings, assert,
otherwise clean the markings and proceed.
Duplicate rx-buffers will be detected in the FW.
CRs-Fixed: 2000844
Change-Id: I9c50cce182a12f3edd3c5b2606836d62f684f816
Add the following:
- Global cumulative counters for posted and received.
- CPU for each individual records
- Counters for refill timer starts
- iwpriv support (dumpStats 10) to dump the counters
CRs-Fixed: 2001331
Change-Id: I66a8a818ffa0b2522358bab404c92a3cdaa603f7
In case of parallel rx execution, fill_count was being updated
by multiple threads and this was causing fill_count to be incorrect
in parallel rx execution cases. This was causing rx ring to be under
filled and in time was causing rx starvation (data stall).
Fix this by:
- making sure that we always fill the right number of buffers based
on the msdu_count in the in_order_indication message.
- making sure that only one execution thread fills the ring
- making sure that the lock holder (filling the ring) does not
run too long by introducing a max fill limit, in which case the
filler will give up and the contender (the other rx execution
thread) takes the lock and fills the ring itself.
CRs-Fixed: 1103851
Change-Id: I59ee2903d51877afc112a1a76b7a5529f1c6a5c0
There was a typo in USR variant of the DEBUG_RX_RING_BUFFER
featured code, which caused a symbol not defined in the header
file, triggering an implicit function definition warning/error.
Fixed typo.
CRs-Fixed: 1104718
Change-Id: I6573904add1453d8c3765f030349a214b058e2c0
Enhance rx ring buffer history to include both post and recv
time-stamps. Add a new event for alloc-index updates.
Change-Id: I061afdcda5d2d41c016be9d3b12e5c4a1cda0c65
CRs-Fixed: 1098656
