The new API can be used to power up/down the SIM. This is equivalent
to inserting and removing the card.
Test: Telephony sanity tests
bug: 32224755
Change-Id: Ib82ab1eff613946e7c31c9b2131183a40d09af80
Previously, we were using provider name as the "SSID" for
Passpoint networks, and use it to compare against real SSIDs,
which obvious doesn't match. So fix it.
Also removed the Passpoint specific logics in WifiTracker for
tracking Passpoint networks, since there shouldn't any
difference in tracking Passpoint APs and normal APs.
Additional work is still needed for managing installed
Passpoint configurations in the "Saved networks" page,
and it is being tracked by b/34207710. The Passpoint configuration
is not an "Access Point" per se, it is a network profile.
Bug: 34455883
Test: Install a Globla Reach Passpoint profile on a bullhead,
verify device auto connects to a Global Reach AP and UI
(QS, QS+, and Settings) correctly reflects it.
Change-Id: I63e66f683f8cb96e664516b130f5e7bc02358a10
Merged-In: I63e66f683f8cb96e664516b130f5e7bc02358a10
The previous changes to these files, which inlined the behavior
of Assert.assertTrue() were done in a hurry to fix the build.
This change makes use of the existing Utils.assertTrue() method
that was created for this purpose.
Bug: 30188076
Test: make checkbuild
Change-Id: Iacd505f8c3f7cefe6de94d30b56e39600bb9b0af
ENABLE_CPUSETS is usually set in BoardConfig.mk to use CONFIG_CPUSETS
feature of Linux kernel.
This change turns the build time variable into runtime decision by
checking if a special file, "tasks", can be found under the mount
point, /dev/cpuset, for cpusets.
Bug: 34726944
Test: Tested on Sailfish
without cpuset mounted
with cpuset mounted, but without schedtune mounted
with cpuset mounted and with schedtune mounted
Change-Id: I1c044c3c9d4d15360a1605e7022da84d94df4014
Changes to ART are causing nano-proto's generated java base classes
to conflict with their duplicates in other projects, leading to a
NullPointerException.
Moving wifi.proto to the same build as other framework service protos in order
for them to share and not generate a duplicate CodedOutputByteBufferNano
classes.
wifi.proto previously lived at:
frameworks/opt/net/wifi/service/proto/wifi.proto
Merged-In: Ia30391453e8d6614edc7f4397e9c17cfe95cf63e
Bug: 32261442
Test: Builds, and existing Unit tests pass
Change-Id: Ia30391453e8d6614edc7f4397e9c17cfe95cf63e
Traditionally, to initiate a new call with the ImsService APIs,
you have to "open()" the interface, which returns a session ID.
This is not needed for querying other information, such as the
provisioning status of the ImsService slot.
Test: Manual
Change-Id: I155b291a49b9aa5ed9599f5a78b1e8d7ff9f8e1c
ro.kernel.qemu based check to determine whether Bluetooth should be
started or not was an old hack used when "features" was not available to
determine hardware availability. This was later fixed with the
hasSystemFeature check. However, the old check was not removed.
The presence of the old check causes Bluetooth to not start on Android
Things devices which have ro.kernel.qemu set for graphics usage. It's
also preferable to depend on the system API rather than on build
configs. Related CLs modify the emulator builds to correct the config in
builds.
Bug: 35361545
Change-Id: I219ce5e14709c76dcba43af4693093e985269d43
Test: Verified by launching bluetooth on local devices after removing
the old check. Also verified by booting up the emulator and verifying
that Bluetooth Services do not start.
