This eliminates the need for the ConnectivityService.VpnCallback class.
This requires shifting VPNs to the new "network" netd API.
VpnService.protect() is modified to no longer go through ConnectivityService.
NetworkCapabilities is extended to add a transport type for VPNs and a
capability requiring a non-VPN (so the default NetworkRequest isn't satisfied
by a VPN).
bug:15409918
Change-Id: Ic4498f1961582208add6f375ad16ce376ee9eb95
Permits apps with permission
android.permission.ACCESS_PERSISTENT_PARTITION to obtain
a read and write data blocks to the PST partition.
Only one block ever exists at one time in PST. When
a client writes another block, the previous one is
overwritten.
This permits storing a block of data that will live
across factory resets.
Change-Id: I8f23df3531f3c0512118eb4b7530eff8a8e81c83
Since DEVICE_AVAILABLE event fires right after initialization, it
could happen before TvInputHal.mPtr is set, which causes calls to
null pointer. Fix the bug by allowing TvInputHal to wait for mPtr to
be set before calling native methods.
Change-Id: Id07f15130beb69f77c16bf3c735285c31ae4a3a4
TvInputHardwareManager detects AudioPort specified by tv_input and
connect source to sink via AudioManager.createAudioPatch().
Bug: 15177175
Change-Id: I2252eb0df2d8287889ed28cc7d76dc1a659fd08b
Routing control action is initiated in various cases, such as manual
TV input port switching, routing change of a different CEC switch,
and so on. The action determines the device to be a new active source.
Change-Id: I1efcd6ff1919dd94d6fa0e0ffa6e430c48d4e9c6
DeviceSelectAction is the main handler for the API deviceSelect() which is
used to choose a new active source among logical devices on the bus.
Change-Id: I77582a1f873423fc316d89f67a89a867461a76b2
Local device is in charge of handling incoming messages
and for some messages such as <set menum language> or
<report physical address> each device has slightly different
behavior. Instread of checking destination address and
local device status from service, this change leaves it
to each local device.
Note that some messages are still left on service
and will be refactored in the following changes.
Along with this, following changes are included.
1. add missing jin interfaces
set_option
set_audio_return_channel
is_connected
Note that get_port_info is under review of jinsuk's change
2. if tv device receives <Report Physical Address>,
starts NewDeviceAction.
3. add constants variables related to new interfaces.
4. add two helper methods for physical address and vendor
id handling
Change-Id: I9c9d04744bc09fbf38431ecfa6e77097b0618a37
The service is replaced with HdmiControlService. Removing all the related
classes and the initialization of the service.
Change-Id: Ic7baaddffb9873613ddd1096e874f226da983939
HdmiCecService will soon be deprecated and replaced with HdmiControlService.
Stopped publishing the service, and turned the corresponding manager to
stub so that it will do nothing but print out a warning msg instead.
Will help the app transition to the new service in the meantime. Once
the migration is finished, will gut out all the HdmiCecService/
HdmiCecManager stuff.
Change-Id: Ic34a69c4fb99f4e5a2c8323538195204ab16a545
Notification for incoming cec message can be issued
from any thread but jni can call java method
in the thread that jni knows like service thread.
So this change delegates incoming message to
jni-friendly thread, service thread by exploit
looper of service thread.
Change-Id: If3b141d917df3e209912af1778eb509777199969
Instead of storing a direct SkCanvas reference, Canvas now tracks
an opaque native wrapper class. The native wrapper can be used to
store additional info for emulating deprecated Skia features
(at this point it only stores a canvas).
Some notes:
* all native handle -> SkCanvas conversions are routed through a
handful of native utility methods.
* safeCanvasSwap() refactored as a lower level setNativeBitmp() - which
is what clients need.
* removed unused get_thread_msec() (Canvas.cpp)
Change-Id: I715a5a6f1e1621c1cfc1e510ae4f2ea15cf11114
HdmiCecService will soon be deprecated and replaced with HdmiControlService.
Stopped publishing the service, and turned the corresponding manager to
stub so that it will do nothing but print out a warning msg instead.
Will help the app transition to the new service in the meantime. Once
the migration is finished, will gut out all the HdmiCecService/
HdmiCecManager stuff.
Change-Id: Ic34a69c4fb99f4e5a2c8323538195204ab16a545
Notification for incoming cec message can be issued
from any thread but jni can call java method
in the thread that jni knows like service thread.
So this change delegates incoming message to
jni-friendly thread, service thread by exploit
looper of service thread.
Change-Id: If3b141d917df3e209912af1778eb509777199969
ARC, Audio Return Channel is newly introduced in HDMI CEC 1.4.
Here is a list of new command
<Initiate ARC>
<Report ARC Initiated>
<Report ARC Terminated>
<Request ARC Initiation>
<Request ARC Termination>
<Terminate ARC>
Along with it replaced <Get OSD Name> with <Give OSD Name>
which is renamed in CEC 1.4.
Change-Id: If4a7aedef839c8413a8e3b5f66f3df0c7234104b
Currently, the native canvas is accessed/manipulated from several
unrelated classes.
In order to facilitate SaveFlags emulation, this CL encapsulates
the field and refactors its external users.
Two main changes:
* new getNativeCanvas() getter for use in Java-level clients.
* JNI canvas swappers (GraphicsBuffers, Surface, TextureView &
AssetAtlasService) are refactored based on the exising/equivalent
safeCanvasSwap() Canvas method.
Change-Id: I966bd4898f0838fb3699e226d3d3d51e0224ea97
HAL interface for CEC has more apis that haven't
connected with HdmiCecController.
Here is a list of apis.
1. addLogicalAddress
2. clearLogicalAddress
3. getPhysicalAddress
4. getVersion
5. getVendorId
Verified that it does not affect to other components in nakasi
Change-Id: I535bde47c43dd25bbede002c0a2aefbe86779076
This change includes native jni implementation for
incoming and outgoing message of CEC.
For incoming message, native layer converts it into three
pieces, source address (initiator), destination address (follower)
and data body which includes opcode. In Java layer, it is delegated to
main io thread. For now all messages are rejected by sending <Feature Abort>
to initiator.
For outoging message, all messages are sent to io thread and it delegates
it into native layer. Native logic converts it into cec_message and
pass it to HAL so that HAL performs sending message.
In order to handle <Feature Abort> message, added [Abort reason] which is
defined in hdmi_cec.h.
Change-Id: If9fd74745f476105e5cfae964e39c78bae69d3e2
This consolidates all of the information that was in the native
KeyEvent and the KeyLayout files into the managed KeyEvent class.
It also moves the definition for all of the key names to the native
side, rather than having them in both places.
Change-Id: I172e3b554e7eb52c79ae2ec406ef4332e8b25ffa
HdmiCecService is a system service handling HDMI-CEC features
and command. Recently we found out that industry has more
requirements to support HDMI-CEC. Also, MHL is another
standard should be in our pocket. Basically, MHL is
a standard to support communication between mobile device
and TV or Av device. As CEC is a control standard over HDMI
cable, MHL has control channel for peer device.
There behavior is very similiar. Both have commands that
can change Tv's current input and can send/receive key
to other device to control other deivce or TV.
In order to cover both CEC and MHL, current HdmiCecService
implementation has limitation. We had several
session of discussion and decided to refactor
HdmiCecService into HdmiControlService.
For each standard it will have separate controller instance
like HdmiCecController and HdmiMhlController.
In this change I didn't touch original HdmiCecService
because some component, like cast receiver, uses HdmiCecService.
For a while we will keep HdmiCecService until HdmiControlService
accomodates all features of HdmiCecService.
Change-Id: I5485280ab803dbf071d898bfbe34be0b11ce7958
Logical address logic is independent from hardware, and the algorithm is
standardized in the spec. Moved the logic to service so that all HAL
implementations doesn't have to duplicated logic inside.
Related change in HAL definition is in:
https://googleplex-android-review.git.corp.google.com/#/c/447951/
Change-Id: I9699309719f3214fc2dbdf219faad484c0c1c5bf
Most of the methods in the interface IHdmiCecService should be implemented
based on the device type. This CL makes a change such that the HdmiCecDevice
just has stub methods that should be overriden by subclasses.
Other changes:
- Fixed a bug of <Inactive Source> not sending its physical address
in its message body. Also the command should have been sent to TV
only rather than broadcast.
- Put back sendGiveDevicePowerStatus interface method. It allows the client
to keep track of the other device(like TV) power status more closely.
Devices goes through the status from standby -> transient to on -> on
but the CEC spec doesn't require that they broacast it actively.
The restored method can be used to let the playback device to get
up-to-date power status of TV/display when it is booting up.
This method should work the same across all the device types. So it was
implemented in the service, not delegated to HdmiCecDevice.
- Send <Report Physical Address> when a new logical device is registered,
which is required by CEC spec: "it should report the association between
its logical and physical address by broadcasting <Report Physical
Address>
Change-Id: Iac1d2cf5783d947f2dcd6965a54670fbdb8e6a63
Declare a new method, Display.getState() to retrieve the actual
power state of a display.
Improved documentation for Intent.ACTION_SCREEN_ON and
Intent.ACTION_SCREEN_OFF to clarify what they really mean in
terms of the interactive state of the device.
Deprecated PowerManager.isScreenOn() and replaced it with
PowerManager.isInteractive() with a more suggestive name and
better documentation.
Redirect display power state changes to go through the display
manager first and only then head over to the power manager for
legacy compatibility.
Eliminated the bright here and woke here policy flags since they
were unused. Simplified the input dispatch policy somewhat.
Ensure that screen wake locks are respected up until the point
when dozing really begins.
Fixed a regression in DreamService where onDreamingStarted
might be called before onWindowAttached.
Bug: 13133142
Bug: 13472578
Bug: 13929355
Bug: 13760290
Change-Id: Iabef96921dd554ce3768fb18619cefc3230b5fb0
