SystemProperties.get() is not particularly fast,
especially if a string is returned. Since ro.* values
are unable to be changed, there's no need to
continously re-query them. Cache the value at
static init time to trivially fix this.
Test: refactoring CL.
Change-Id: Iccb021d3cb2ba3a4a1d0048ddec6811bb7409eec
(cherry picked from commit aa67f684ff)
* commit 'a19056c35d16ddb5a6c1d3343729701b8939f1e1': (35 commits)
NetworkMonitor: send one DNS probe per web probe
NetworkMonitor metrics: add first validation information
APF: also drop any ICMPv6 RSs
ConnectivityServiceTest: fix testAvoidBadWifiSettings
Fix ConnectivityServiceTest testRequestBenchmark
Switch over to new "time.android.com" NTP pool.
Define API for metering network stats buckets.
Refactored NetworkStatsServiceTest to use Mockito instead of EasyMock.
Use @Ignore to explicitly disable a @Test method.
Fixed NetworkStatsServiceTest and converted it to JUnit4.
VPN network stat accounting changes.
ConnectivityThread: use lazy holder idiom
ConnectivityManager: use ConnectivityThread looper
ConnectivityManager: a simpler CallbackHandler
Indicate the NsdServiceInfo attributes are only filled in for a resolved service.
Add a null check for the OnStartTetheringCallback.
TokenBucket for rate-limiting and throttling
IpConnectivityMetrics reads buffer size in settings
CaptivePortalLogin: set mixed content policy to compatibility.
Add IP conn metrics to dumpsys and bug reports
...
This patch adds a TokenBucket class that can accumulate and distribute
tokens over time.
Test: comes with unit tests.
Bug: 32198637
(cherry picked from commit 998493f0ee)
Change-Id: I4012ae6d02f7004bceee9a2ba03914f2a2506031
Fixes a bug in MessagingLinearLayout where we would not
recompute the height of it after applying the computed
insets to the text views, leading to incorrect padding
and in rare cases a cut off message at the bottom.
Change-Id: If87a527555158e94e501832e9f49e380cc7da2bb
Test: runtest -x core/tests/coretests/src/com/android/internal/widget/MessagingLinearLayoutTest.java
Fixes: 31463075
(cherry picked from commit 4ac8f40d96)
Preloading EGL in Zygote was originally a memory footprint
optimization, but it turns out to be an important app startup time
optimization as well. Preloading EGL in Zygote is incompatible with
updatable graphics drivers, but we don't want to do it on-demand as
part of drawing the first frame either, since that increases
first-frame latency unacceptably.
This change removes Zygote preload, and instead loads EGL on a
low-priority background thread immediately after choosing which
graphics driver to use. This means it is usually done well before
drawing the first frame, without significantly disrupting other
activity launch work.
Test: observe systrace of Calculator launch on bullhead
Bug: 34404021
Change-Id: I6a0f6b90ade21848a10d51ddae62c936f70151b5
Merged-In: I887aa09bd35b088b16f53a89838a0c7c98f15761
Preloading EGL in Zygote was originally a memory footprint
optimization, but it turns out to be an important app startup time
optimization as well. Preloading EGL in Zygote is incompatible with
updatable graphics drivers, but we don't want to do it on-demand as
part of drawing the first frame either, since that increases
first-frame latency unacceptably.
This change removes Zygote preload, and instead loads EGL on a
low-priority background thread immediately after choosing which
graphics driver to use. This means it is usually done well before
drawing the first frame, without significantly disrupting other
activity launch work.
Test: observe systrace of Calculator launch on bullhead
Bug: 34404021
Change-Id: I887aa09bd35b088b16f53a89838a0c7c98f15761
It's not clear this actually makes much difference on modern
devices/drivers. With updatable graphics drivers, we'd have to be able
to unload the preloaded driver from processes that don't use the
system driver, which is additional complexity and risk.
On bullhead and sailfish, meminfo actually showed slightly more memory
available while sitting at launcher just after boot with this change
than previously. Looking at detailed stats, the differences appeared
to mostly within run-to-run variation, but there wasn't evidence of a
regression.
Bug: 33531483
Test: boot through lockscreen/launcher
Change-Id: I1892302c1750cdbeaf5b9979f8da4dc6bd7b3e75
Merged-In: I1892302c1750cdbeaf5b9979f8da4dc6bd7b3e75
On FBE devices, don't save the metrics to disk but compute them when the
password is first entered and only store them in RAM.
Merged-in: 5daf273b7e
Bug: 32793550
Change-Id: Icee7f615167761177b224b342970a36c7d90f6ba
Add file descriptors that should be ignored in the file descriptor
table check.
Use this to ignore the pipe file descriptors when starting an app
with invoke-with (wrapper).
Bug: 32607405
Test: m
Test: Device boots
Test: Apps start
Test: adb root && adb shell setprop wrap.com.android.calculator2 logwrapper && adb shell am start -S -n com.android.calculator2/.Calculator
Change-Id: I4a30dfc9382d3172cc815bd41b17a123799354f6
This refactors the whitelist to be a class, rather than just a static C array.
The whitelist can then be augmented dynamically when the package path is known
in the webview_zygote.
Test: m
Test: sailfish boots
Test: Enable Multi-process WebView in developer options, perform a search in GSA.
Bug: 21643067
Change-Id: Ia1f2535c7275b42b309631b4fe7859c30cbf7309
(cherry picked from commit 061ee3088a)
In some cases (e.g. Android Wear) SystemWindowInsets can be non-zero
due to overscan layout hints even when no SystemUI is present.
This change resepects the overscan flag on input method windows
allowing an IME to opt-out of the navigation bar guard and receive
full-height content view.
BUG: 32700226
Change-Id: Ic38f204a892bf34e8dae65990d5aa8c95af555d8
OEMs can overlay the default 9 dots by providing two drawables that represent
those dots:
* lockscreen_notselected: asset to display when a cell has not been selected.
* lockscreen_selected: asset to display when a cell has been selected.
BUG: 33755663
Change-Id: Ic595b01f5e1321696b7a3feb0ff73c1acccfb942
