It has an effect on following case:
If STA firstly connected with 6ghz channel AP and Application create p2p group without config, wpa_supplicant also tried to select STA-associated channel(in this case, i.e. 6ghz ch).
Bug: 281733591
Test: atest android.net.wifi.cts.ConcurrencyTest#testPersistentGroupOperation
Change-Id: I0015fb9ad602f585a6b1572cf13967b0a895f0f2
The country code updating from scan results generic case can cover setup wizard case.
Bug: 279625751
Test: manually verified generic case
Change-Id: I91815abfb4ec9e6c2bc324c9bf3eb6ffee7d6cab
note: add smartfeature.gstf that only used for Hotword
pipeline in AoC which is sharing to chre usage too.
On WHI_PRO
CHRE PDM Gain(gain=158)
CHRE AEC Gain(gain=-42)
CHRE SRC AEC Timeout(timeout=0)
This table also contains hotword gain control as
(Undocked)
Non Eraser gain (gain=220)
Eraser Pre gain (gain=220)
(Docked)
Non Eraser gain (gain=356)
Eraser Pre gain (gain=170)
Eraser Post gain (gain=186)
Test: end-to-end verify by CHRE
using aoc_audio_cfg to query gain as below
adb shell aoc_audio_cfg hotword_get_non_eraser_gain
adb shell aoc_audio_cfg hotword_get_eraser_post_gain
adb shell aoc_audio_cfg hotword_get_eraser_pre_gain
Bug: 271078701
Change-Id: I4e3416af224742d939c0928c6623705a01924581
enable load kernel modules in parallel but load some
specific modules in sequential
Test: reboot more than 500 times without failure
Bug: 229794277
Change-Id: I41cb5a09fb88ff7faa5ec12f71209b35cd941948
Signed-off-by: Chungkai Mei <chungkai@google.com>
Upstream has fixed relationship between uclamp_min and margin which
means uclamp_min value between 80-100% of capacity of the CPU is a valid
value and will not cause upmigration.
To keep the current hints in powerhint.json the same, we must update
uclamp_min to be capacity_orig_of(medium_cpu) + 1.
This change is dependent on uclamp vs fits margin in GKI, vendor
modules. And an improvement in sched_pixel to ignore the 25% boost if
the only reason to request the frequency is uclamp_min.
To cater for the later - we multiply each uclamp_min value with 1.25
- except ones that look intentionally set around capacity boundaries.
Bug: 261695590
Test: PtsUiBench
Signed-off-by: Qais Yousef <qyousef@google.com>
Change-Id: I1e6889ebea9877cb23708722a73e97177e6be185
On this device, the default touch slop value of 8dp is too large. It corresponds to a physical distance of ~ 1.5mm.
This results in long presses getting triggered too easily.
Reduce the touch slop to 5 dp, which is about ~ 1mm, and closer to a reference device that doesn't exhibit the long-press issue.
Bug: 270835406
Change-Id: I1973e7831f527f2f654661bd4917f48992c74b61
Test: use a test app for long presses
Using the = operator makes it so no one else can assign the system
property anymore. Other build targets downstream do try to set these
properties explicitly to false. Since no one sets the property before,
we can use the ?= operator to apply the values, which still allows other
makefiles downstream the chance to the change the value.
Bug: 274950131
Test: m
Change-Id: I113da6efefccef562bb13cb73087e4b5c7eca453
Bug: 262056923
Bug: 274537630
Test: Screen is off when put a magnet close to hall sensor.
Change-Id: I950c1d5f6284f8ce133ca94ce98eec78ec9cfa93
Merged-In: I950c1d5f6284f8ce133ca94ce98eec78ec9cfa93
