Modify the existing radio_stat structure to add time spent in each
tranmission power level.
BUG: 27227497
Change-Id: Ifba9f22bc7a404161ae9124c787854ed3431cb8d
cherry-picked from:
https://partner-android-review.googlesource.com/#/c/543388
Given that there is an explicit enumerated value to indicate
a drop due to lack of buffers, we shouldn't recommend using
the OTHER value for out of buffers.
BUG=27352446
TEST=compile
Change-Id: Id4fdbf530ed9a99e2b9693acb649137d598abf85
For consistency with other packet_fate functions,
wifi_start_pkt_fate_monitoring() should name the first
parameter as |handle|.
BUG=27409524
TEST=compile
Change-Id: Ic6df01e7e80f79eaf5653374106ba6e6a7609e31
Initial check-in of ND offload HAL API support
BUG=27598285
Change-Id: I9f2652de2bb0e22a25fe7c0ded6d1f5e6500de6d
Signed-off-by: Jun Lim <jun.lim@broadcom.com>
Modify the existing wake accounting reasons HAL API so that the
individual wake accounting reason array buffers are allocated
and freed by framework.
BUG:26902794
Change-Id: Iea74e61890f785107b05f9b4a3a58449ba449f0d
disabled was not very clear as to the actual meaning of the event and
left the implication that the scan could be reenabled.
Change-Id: I3f53b13a958cd35e93fde7bf9690d84744ffe63a
When a connection attempt fails, we're often unable
to determine the cause from existing log files. This
is because our log files only report the view from user
space. From this vantage point, we can't tell if,
e.g., an outbound frame was ever sent over the air.
(Or, for inbound frames: if they were heard by our
radio.)
To give ourselves a broader view, we add the packet fate
monitoring API. This API will let us determine, e.g.,
if an outbound frame was transmitted by our radio.
BUG: 26807558
Change-Id: I197729117f9086ece8cf00313bb9d3e7c0197b97
The current list of events supported by the wifi_connectivity_event
ring buffer omits some important events. Add three new events, so
that we can log the events for driver/firmware that support them.
BUG: 26832275
Change-Id: I2a4e674e5f7b52cc630eac32191fdc0b75d60422
Add API for dumping state of WiFi driver. This API
will be used in combination with the (existing)
firmware memory dump API, to identify inconsistent
state between driver and firmware.
BUG: 26803847
Change-Id: Ie65f095ca6be4203f8aad0b541c1903dc6cce1f5
Changes related to NAN APIs and structures.
1) Remove NanHeader
2) Replace wifi_handle with wifi_interface_handle in NAN APIs
3) Add publish/subscribe id
4) Remove match_handle and name it as requestor_instance_id
5) Add NanPublishResponse/NanSubscribeResponse structures
6) Remove user_context
7) Add appropriate enums
8) Remove unused structures
9) Add new NAN status types for TCA, Transmit, and invalid
configuration of Band, DW/DB intervals, random factor
update and scan period.
10) Add new fields in Publish/Subscribe Stats.
11) Replace wifi_request_id with transaction_id to have a
common interface across all layers.
Bug: 26216681
Change-Id: Ife7cf7e7ed64d2f8db236f23919cd6412dc8b496
Correct the comments on exponential back off scan formula.
Rename "exponent" to "base" to match the formula.
Also make the meaning of AP (Access Point vs. Apps Processor)
explicit for struct wifi_scan_cmd_params.
Bug: 26236392
Change-Id: I5873e1c8667dd964d125e37e2f0eeddec79eda13
To enable the vendor implementation of the vibrator HAL, it is needed
to use hardware modules, a default one for AOSP and the vendor one.
The legacy implementation is no more needed so let's remove it.
This change is related to other changes in:
- hardware/libhardware
- frameworks/base
Change-Id: I844279f5535289f079d412fdc44c5cb3c9c1130c
Author: Vincent Becker <vincentx.becker@intel.com>
Signed-off-by: Vincent Becker <vincentx.becker@intel.com>
Signed-off-by: Shuo Gao <shuo.gao@intel.com>
Signed-off-by: Bruce Beare <bruce.j.beare@intel.com>
Signed-off-by: Jack Ren <jack.ren@intel.com>
Signed-off-by: David Wagner <david.wagner@intel.com>
Author-tracking-BZ: 49760
Reserved a bit for RSSI monitor feature in the feature bitmask
Change-Id: Ied470f65c26bed25f7ccc4faa3ac8fbac1e74009
Signed-off-by: Ashwin <ashwin.bhat@broadcom.com>
e28e2bb Introduce REPORT_EVENTS_NO_BATCH flag for ePNO
ddcab17 WifiLogger: Add new events and tags
1abb250 change wifi Logging header
bd14c1c add pointers to new roaming HAL functions
Change-Id: I4645ceb57de34109d417bc9b84ff6d2cb202a78d
When GSscan and ePNO are being executed together, we
don't want APs on additional channels from ePNO to be
cached. Making this requirement explicit by adding
REPORT_EVENTS_NO_BATCH flag on bucket.
Change-Id: I61dcd608d3be8acb697508607dbe5077ebe836c0
As part of the Debug Framework development new events and tags
are required in G_SCAN. Additional generic tags are also being
added to support various events.
Change-Id: Ia5bc8dc1a5310818c3fcf3f9e166ffc5341f032c
1) add wifi_set_alert_handler for the alert case
2) add wifi_set_log_handler for the ring data
3) remove unnecessary wifi_request_id for set commands
4) change to use ring_name instead of ring_id.
5) add wifi_reset_alert_handler
add wifi_reset_log_handler
Change-Id: Ia9aaba050196af6684191cdcdfb0c1d21e917ed2
Signed-off-by: eccopark@broadcom.com <eccopark@broadcom.com>
Android framework should provide a buffer to the wifihal which needs
to be filled with scan_result. This commit will ensure that would
happen by passing a fixed max buffer size as part of each
wifi_cached_scan_results.
Change-Id: I005dd39f26be62694c46369e16d5f7f89f64021d
This change introduces wifi_set_passpoint_list and wifi_reset_passpoint_list
which implement the functionality of offloading passpoint match to
the chipset firmware.
Change-Id: I8a3730d4e23891b1dce47b8cf877e9b3c04ac279
1) add the request_id in wifi_get_logger_supported_feature_set
2) Change the wifi_get_firmware_memory_dump API
1) Normally the size of firmware dump is 700KB.
Thus, Driver cannot send the data in one time because of limitation of skb length.
In order to send such big data from kernel Driver,
we need to send data using event continuously until we send the memory dump completely.
In the event handler of hal, it will allocate the
big buffer to store the data.
Once event handler copy the data completely, event handler
will call the function callback(on_firmware_memory_dump).
So, upper layer has to free the buffer after handling the buffer.
Change-Id: If80b03ec8bd9349d952743e17a001c04096aac95
Signed-off-by: eccopark@broadcom.com <eccopark@broadcom.com>
1) define entry type for parsing the data easily.
2) the payload will be attached at end of the wifi_ring_buffer_entry
3) added the feature set
4) remove duplicate wifi_get_ring_data
5) add request_id in wifi_start_logging
6) change the pointer of buffer to double pointer of buffer
in wifi_get_ring_buffers_status and wifi_get_firmware_memory_dump
wifi_get_firmware_version and wifi_get_driver_version
to indicate the buffer supplied by Application layer
Change-Id: I5dc2811155e7efa058a0c3414036461de3e45d98
Signed-off-by: eccopark@broadcom.com <eccopark@broadcom.com>
Mainstream wpa_supplicant now has common driver_nl80211.h
Change-Id: I6d886daf39110969b2442c3943806364e63a5d04
Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
Added single shot scan, and maxScansToCache. Also fixed getScanResults()
to report scan results to listeners.
Bug: 18087596
Bug: 17694040
Bug: 18059273
Change-Id: Id623ab9ee2ab89c86a35082279331fa609016347
Mainstream wpa_supplicant now has common driver_nl80211.h
Change-Id: I6d886daf39110969b2442c3943806364e63a5d04
Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
When the wifi_hal.h is included as part of the c source
code files some of the functions are not exported. The
const variable as an index to an array is causing compiler
error. The changes are to address these issues.
Change-Id: Ia2dbd7a87315cee8d0246a8ac6539c9b4b22ffd4
Added single shot scan, and maxScansToCache. Also fixed getScanResults()
to report scan results to listeners.
Bug: 18087596
Bug: 17694040
Bug: 18059273
Change-Id: I8a6da5426561e22e72d1e99c5537205b893077dd
New stream types have been added, so make sure that
legacy implementation uses legacy stream definitions
Bug: 18067208
Change-Id: I8042ebe9ac3e7eda380b8c91650f2ac16b44bc34
* commit '69f48b86d78ac8968ccafc6287eed3079f1cb56e':
[AU]Buganizer ID: 17182487, sprout audio HAL should mute all audio output with property ro.audio.silent set to true
Added a new macro in feature set to indicate wifihal support for
link layer stats collection.
Bug: 17359758
Change-Id: I3f8ecce67430bd8270fb157a4e6cd246503a3d9f
Fixes a race condition whereby two requests attempt to change the audio stream
(and interrupt each other). For example, a user space app makes a call through
the audio API and before completion, a kernel function such as jack removal is
detected and is serviced. In this case, you have a nullified audio route that
needs to be re-routed.
In order to know a routing update is needed, a boolean member mForceRouting
has been introduced to enforce routing for the next attempt to set audio output
with a non-null device.
Change-Id: Ia0aa2f27fe63c39f0980dec2e1e79dbc26b81e56
Author: Vincent Becker <vincentx.becker@intel.com>
Signed-off-by: Vincent Becker <vincentx.becker@intel.com>
Signed-off-by: Zhi Jin <zhi.jin@intel.com>
Signed-off-by: Shuo Gao <shuo.gao@intel.com>
Signed-off-by: Bruce Beare <bruce.j.beare@intel.com>
Signed-off-by: Jack Ren <jack.ren@intel.com>
Signed-off-by: Robert Beatty <robert.m.beatty@intel.com>
