| /* |
| * Copyright (C) 2010 The Android Open Source Project |
| * |
| * Licensed under the Apache License, Version 2.0 (the "License"); |
| * you may not use this file except in compliance with the License. |
| * You may obtain a copy of the License at |
| * |
| * http://www.apache.org/licenses/LICENSE-2.0 |
| * |
| * Unless required by applicable law or agreed to in writing, software |
| * distributed under the License is distributed on an "AS IS" BASIS, |
| * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| * See the License for the specific language governing permissions and |
| * limitations under the License. |
| */ |
| |
| #ifndef _LIBINPUT_INPUT_TRANSPORT_H |
| #define _LIBINPUT_INPUT_TRANSPORT_H |
| |
| #pragma GCC system_header |
| |
| /** |
| * Native input transport. |
| * |
| * The InputChannel provides a mechanism for exchanging InputMessage structures across processes. |
| * |
| * The InputPublisher and InputConsumer each handle one end-point of an input channel. |
| * The InputPublisher is used by the input dispatcher to send events to the application. |
| * The InputConsumer is used by the application to receive events from the input dispatcher. |
| */ |
| |
| #include <string> |
| #include <unordered_map> |
| |
| #include <android-base/chrono_utils.h> |
| #include <android-base/result.h> |
| #include <android-base/unique_fd.h> |
| |
| #include <binder/IBinder.h> |
| #include <binder/Parcelable.h> |
| #include <input/Input.h> |
| #include <sys/stat.h> |
| #include <ui/Transform.h> |
| #include <utils/BitSet.h> |
| #include <utils/Errors.h> |
| #include <utils/RefBase.h> |
| #include <utils/Timers.h> |
| |
| |
| namespace android { |
| class Parcel; |
| |
| /* |
| * Intermediate representation used to send input events and related signals. |
| * |
| * Note that this structure is used for IPCs so its layout must be identical |
| * on 64 and 32 bit processes. This is tested in StructLayout_test.cpp. |
| * |
| * Since the struct must be aligned to an 8-byte boundary, there could be uninitialized bytes |
| * in-between the defined fields. This padding data should be explicitly accounted for by adding |
| * "empty" fields into the struct. This data is memset to zero before sending the struct across |
| * the socket. Adding the explicit fields ensures that the memset is not optimized away by the |
| * compiler. When a new field is added to the struct, the corresponding change |
| * in StructLayout_test should be made. |
| */ |
| struct InputMessage { |
| enum class Type : uint32_t { |
| KEY, |
| MOTION, |
| FINISHED, |
| FOCUS, |
| CAPTURE, |
| DRAG, |
| TIMELINE, |
| }; |
| |
| struct Header { |
| Type type; // 4 bytes |
| uint32_t seq; |
| } header; |
| |
| // For keys and motions, rely on the fact that std::array takes up exactly as much space |
| // as the underlying data. This is not guaranteed by C++, but it simplifies the conversions. |
| static_assert(sizeof(std::array<uint8_t, 32>) == 32); |
| |
| // For bool values, rely on the fact that they take up exactly one byte. This is not guaranteed |
| // by C++ and is implementation-dependent, but it simplifies the conversions. |
| static_assert(sizeof(bool) == 1); |
| |
| // Body *must* be 8 byte aligned. |
| union Body { |
| struct Key { |
| int32_t eventId; |
| uint32_t empty1; |
| nsecs_t eventTime __attribute__((aligned(8))); |
| int32_t deviceId; |
| int32_t source; |
| int32_t displayId; |
| std::array<uint8_t, 32> hmac; |
| int32_t action; |
| int32_t flags; |
| int32_t keyCode; |
| int32_t scanCode; |
| int32_t metaState; |
| int32_t repeatCount; |
| uint32_t empty2; |
| nsecs_t downTime __attribute__((aligned(8))); |
| |
| inline size_t size() const { return sizeof(Key); } |
| } key; |
| |
| struct Motion { |
| int32_t eventId; |
| uint32_t empty1; |
| nsecs_t eventTime __attribute__((aligned(8))); |
| int32_t deviceId; |
| int32_t source; |
| int32_t displayId; |
| std::array<uint8_t, 32> hmac; |
| int32_t action; |
| int32_t actionButton; |
| int32_t flags; |
| int32_t metaState; |
| int32_t buttonState; |
| MotionClassification classification; // base type: uint8_t |
| uint8_t empty2[3]; // 3 bytes to fill gap created by classification |
| int32_t edgeFlags; |
| nsecs_t downTime __attribute__((aligned(8))); |
| float dsdx; |
| float dtdx; |
| float dtdy; |
| float dsdy; |
| float tx; |
| float ty; |
| float xPrecision; |
| float yPrecision; |
| float xCursorPosition; |
| float yCursorPosition; |
| uint32_t displayOrientation; |
| int32_t displayWidth; |
| int32_t displayHeight; |
| uint32_t pointerCount; |
| /** |
| * The "pointers" field must be the last field of the struct InputMessage. |
| * When we send the struct InputMessage across the socket, we are not |
| * writing the entire "pointers" array, but only the pointerCount portion |
| * of it as an optimization. Adding a field after "pointers" would break this. |
| */ |
| struct Pointer { |
| PointerProperties properties; |
| PointerCoords coords; |
| } pointers[MAX_POINTERS] __attribute__((aligned(8))); |
| |
| int32_t getActionId() const { |
| uint32_t index = (action & AMOTION_EVENT_ACTION_POINTER_INDEX_MASK) |
| >> AMOTION_EVENT_ACTION_POINTER_INDEX_SHIFT; |
| return pointers[index].properties.id; |
| } |
| |
| inline size_t size() const { |
| return sizeof(Motion) - sizeof(Pointer) * MAX_POINTERS |
| + sizeof(Pointer) * pointerCount; |
| } |
| } motion; |
| |
| struct Finished { |
| bool handled; |
| uint8_t empty[7]; |
| nsecs_t consumeTime; // The time when the event was consumed by the receiving end |
| |
| inline size_t size() const { return sizeof(Finished); } |
| } finished; |
| |
| struct Focus { |
| int32_t eventId; |
| // The following 3 fields take up 4 bytes total |
| bool hasFocus; |
| bool inTouchMode; |
| uint8_t empty[2]; |
| |
| inline size_t size() const { return sizeof(Focus); } |
| } focus; |
| |
| struct Capture { |
| int32_t eventId; |
| bool pointerCaptureEnabled; |
| uint8_t empty[3]; |
| |
| inline size_t size() const { return sizeof(Capture); } |
| } capture; |
| |
| struct Drag { |
| int32_t eventId; |
| float x; |
| float y; |
| bool isExiting; |
| uint8_t empty[3]; |
| |
| inline size_t size() const { return sizeof(Drag); } |
| } drag; |
| |
| struct Timeline { |
| int32_t eventId; |
| uint32_t empty; |
| std::array<nsecs_t, GraphicsTimeline::SIZE> graphicsTimeline; |
| |
| inline size_t size() const { return sizeof(Timeline); } |
| } timeline; |
| } __attribute__((aligned(8))) body; |
| |
| bool isValid(size_t actualSize) const; |
| size_t size() const; |
| void getSanitizedCopy(InputMessage* msg) const; |
| }; |
| |
| /* |
| * An input channel consists of a local unix domain socket used to send and receive |
| * input messages across processes. Each channel has a descriptive name for debugging purposes. |
| * |
| * Each endpoint has its own InputChannel object that specifies its file descriptor. |
| * |
| * The input channel is closed when all references to it are released. |
| */ |
| class InputChannel : public Parcelable { |
| public: |
| static std::unique_ptr<InputChannel> create(const std::string& name, |
| android::base::unique_fd fd, sp<IBinder> token); |
| InputChannel() = default; |
| InputChannel(const InputChannel& other) |
| : mName(other.mName), mFd(::dup(other.mFd)), mToken(other.mToken){}; |
| InputChannel(const std::string name, android::base::unique_fd fd, sp<IBinder> token); |
| ~InputChannel() override; |
| /** |
| * Create a pair of input channels. |
| * The two returned input channels are equivalent, and are labeled as "server" and "client" |
| * for convenience. The two input channels share the same token. |
| * |
| * Return OK on success. |
| */ |
| static status_t openInputChannelPair(const std::string& name, |
| std::unique_ptr<InputChannel>& outServerChannel, |
| std::unique_ptr<InputChannel>& outClientChannel); |
| |
| inline std::string getName() const { return mName; } |
| inline const android::base::unique_fd& getFd() const { return mFd; } |
| inline sp<IBinder> getToken() const { return mToken; } |
| |
| /* Send a message to the other endpoint. |
| * |
| * If the channel is full then the message is guaranteed not to have been sent at all. |
| * Try again after the consumer has sent a finished signal indicating that it has |
| * consumed some of the pending messages from the channel. |
| * |
| * Return OK on success. |
| * Return WOULD_BLOCK if the channel is full. |
| * Return DEAD_OBJECT if the channel's peer has been closed. |
| * Other errors probably indicate that the channel is broken. |
| */ |
| status_t sendMessage(const InputMessage* msg); |
| |
| /* Receive a message sent by the other endpoint. |
| * |
| * If there is no message present, try again after poll() indicates that the fd |
| * is readable. |
| * |
| * Return OK on success. |
| * Return WOULD_BLOCK if there is no message present. |
| * Return DEAD_OBJECT if the channel's peer has been closed. |
| * Other errors probably indicate that the channel is broken. |
| */ |
| status_t receiveMessage(InputMessage* msg); |
| |
| /* Return a new object that has a duplicate of this channel's fd. */ |
| std::unique_ptr<InputChannel> dup() const; |
| |
| void copyTo(InputChannel& outChannel) const; |
| |
| status_t readFromParcel(const android::Parcel* parcel) override; |
| status_t writeToParcel(android::Parcel* parcel) const override; |
| |
| /** |
| * The connection token is used to identify the input connection, i.e. |
| * the pair of input channels that were created simultaneously. Input channels |
| * are always created in pairs, and the token can be used to find the server-side |
| * input channel from the client-side input channel, and vice versa. |
| * |
| * Do not use connection token to check equality of a specific input channel object |
| * to another, because two different (client and server) input channels will share the |
| * same connection token. |
| * |
| * Return the token that identifies this connection. |
| */ |
| sp<IBinder> getConnectionToken() const; |
| |
| bool operator==(const InputChannel& inputChannel) const { |
| struct stat lhs, rhs; |
| if (fstat(mFd.get(), &lhs) != 0) { |
| return false; |
| } |
| if (fstat(inputChannel.getFd(), &rhs) != 0) { |
| return false; |
| } |
| // If file descriptors are pointing to same inode they are duplicated fds. |
| return inputChannel.getName() == getName() && inputChannel.getConnectionToken() == mToken && |
| lhs.st_ino == rhs.st_ino; |
| } |
| |
| private: |
| base::unique_fd dupFd() const; |
| |
| std::string mName; |
| android::base::unique_fd mFd; |
| |
| sp<IBinder> mToken; |
| }; |
| |
| /* |
| * Publishes input events to an input channel. |
| */ |
| class InputPublisher { |
| public: |
| /* Creates a publisher associated with an input channel. */ |
| explicit InputPublisher(const std::shared_ptr<InputChannel>& channel); |
| |
| /* Destroys the publisher and releases its input channel. */ |
| ~InputPublisher(); |
| |
| /* Gets the underlying input channel. */ |
| inline std::shared_ptr<InputChannel> getChannel() { return mChannel; } |
| |
| /* Publishes a key event to the input channel. |
| * |
| * Returns OK on success. |
| * Returns WOULD_BLOCK if the channel is full. |
| * Returns DEAD_OBJECT if the channel's peer has been closed. |
| * Returns BAD_VALUE if seq is 0. |
| * Other errors probably indicate that the channel is broken. |
| */ |
| status_t publishKeyEvent(uint32_t seq, int32_t eventId, int32_t deviceId, int32_t source, |
| int32_t displayId, std::array<uint8_t, 32> hmac, int32_t action, |
| int32_t flags, int32_t keyCode, int32_t scanCode, int32_t metaState, |
| int32_t repeatCount, nsecs_t downTime, nsecs_t eventTime); |
| |
| /* Publishes a motion event to the input channel. |
| * |
| * Returns OK on success. |
| * Returns WOULD_BLOCK if the channel is full. |
| * Returns DEAD_OBJECT if the channel's peer has been closed. |
| * Returns BAD_VALUE if seq is 0 or if pointerCount is less than 1 or greater than MAX_POINTERS. |
| * Other errors probably indicate that the channel is broken. |
| */ |
| status_t publishMotionEvent(uint32_t seq, int32_t eventId, int32_t deviceId, int32_t source, |
| int32_t displayId, std::array<uint8_t, 32> hmac, int32_t action, |
| int32_t actionButton, int32_t flags, int32_t edgeFlags, |
| int32_t metaState, int32_t buttonState, |
| MotionClassification classification, const ui::Transform& transform, |
| float xPrecision, float yPrecision, float xCursorPosition, |
| float yCursorPosition, uint32_t displayOrientation, |
| int32_t displayWidth, int32_t displayHeight, nsecs_t downTime, |
| nsecs_t eventTime, uint32_t pointerCount, |
| const PointerProperties* pointerProperties, |
| const PointerCoords* pointerCoords); |
| |
| /* Publishes a focus event to the input channel. |
| * |
| * Returns OK on success. |
| * Returns WOULD_BLOCK if the channel is full. |
| * Returns DEAD_OBJECT if the channel's peer has been closed. |
| * Other errors probably indicate that the channel is broken. |
| */ |
| status_t publishFocusEvent(uint32_t seq, int32_t eventId, bool hasFocus, bool inTouchMode); |
| |
| /* Publishes a capture event to the input channel. |
| * |
| * Returns OK on success. |
| * Returns WOULD_BLOCK if the channel is full. |
| * Returns DEAD_OBJECT if the channel's peer has been closed. |
| * Other errors probably indicate that the channel is broken. |
| */ |
| status_t publishCaptureEvent(uint32_t seq, int32_t eventId, bool pointerCaptureEnabled); |
| |
| /* Publishes a drag event to the input channel. |
| * |
| * Returns OK on success. |
| * Returns WOULD_BLOCK if the channel is full. |
| * Returns DEAD_OBJECT if the channel's peer has been closed. |
| * Other errors probably indicate that the channel is broken. |
| */ |
| status_t publishDragEvent(uint32_t seq, int32_t eventId, float x, float y, bool isExiting); |
| |
| struct Finished { |
| uint32_t seq; |
| bool handled; |
| nsecs_t consumeTime; |
| }; |
| |
| struct Timeline { |
| int32_t inputEventId; |
| std::array<nsecs_t, GraphicsTimeline::SIZE> graphicsTimeline; |
| }; |
| |
| typedef std::variant<Finished, Timeline> ConsumerResponse; |
| /* Receive a signal from the consumer in reply to the original dispatch signal. |
| * If a signal was received, returns a Finished or a Timeline object. |
| * The InputConsumer should return a Finished object for every InputMessage that it is sent |
| * to confirm that it has been processed and that the InputConsumer is responsive. |
| * If several InputMessages are sent to InputConsumer, it's possible to receive Finished |
| * events out of order for those messages. |
| * |
| * The Timeline object is returned whenever the receiving end has processed a graphical frame |
| * and is returning the timeline of the frame. Not all input events will cause a Timeline |
| * object to be returned, and there is not guarantee about when it will arrive. |
| * |
| * If an object of Finished is returned, the returned sequence number is never 0 unless the |
| * operation failed. |
| * |
| * Returned error codes: |
| * OK on success. |
| * WOULD_BLOCK if there is no signal present. |
| * DEAD_OBJECT if the channel's peer has been closed. |
| * Other errors probably indicate that the channel is broken. |
| */ |
| android::base::Result<ConsumerResponse> receiveConsumerResponse(); |
| |
| private: |
| std::shared_ptr<InputChannel> mChannel; |
| }; |
| |
| /* |
| * Consumes input events from an input channel. |
| */ |
| class InputConsumer { |
| public: |
| /* Creates a consumer associated with an input channel. */ |
| explicit InputConsumer(const std::shared_ptr<InputChannel>& channel); |
| |
| /* Destroys the consumer and releases its input channel. */ |
| ~InputConsumer(); |
| |
| /* Gets the underlying input channel. */ |
| inline std::shared_ptr<InputChannel> getChannel() { return mChannel; } |
| |
| /* Consumes an input event from the input channel and copies its contents into |
| * an InputEvent object created using the specified factory. |
| * |
| * Tries to combine a series of move events into larger batches whenever possible. |
| * |
| * If consumeBatches is false, then defers consuming pending batched events if it |
| * is possible for additional samples to be added to them later. Call hasPendingBatch() |
| * to determine whether a pending batch is available to be consumed. |
| * |
| * If consumeBatches is true, then events are still batched but they are consumed |
| * immediately as soon as the input channel is exhausted. |
| * |
| * The frameTime parameter specifies the time when the current display frame started |
| * rendering in the CLOCK_MONOTONIC time base, or -1 if unknown. |
| * |
| * The returned sequence number is never 0 unless the operation failed. |
| * |
| * Returns OK on success. |
| * Returns WOULD_BLOCK if there is no event present. |
| * Returns DEAD_OBJECT if the channel's peer has been closed. |
| * Returns NO_MEMORY if the event could not be created. |
| * Other errors probably indicate that the channel is broken. |
| */ |
| status_t consume(InputEventFactoryInterface* factory, bool consumeBatches, nsecs_t frameTime, |
| uint32_t* outSeq, InputEvent** outEvent); |
| |
| /* Sends a finished signal to the publisher to inform it that the message |
| * with the specified sequence number has finished being process and whether |
| * the message was handled by the consumer. |
| * |
| * Returns OK on success. |
| * Returns BAD_VALUE if seq is 0. |
| * Other errors probably indicate that the channel is broken. |
| */ |
| status_t sendFinishedSignal(uint32_t seq, bool handled); |
| |
| status_t sendTimeline(int32_t inputEventId, |
| std::array<nsecs_t, GraphicsTimeline::SIZE> timeline); |
| |
| /* Returns true if there is a deferred event waiting. |
| * |
| * Should be called after calling consume() to determine whether the consumer |
| * has a deferred event to be processed. Deferred events are somewhat special in |
| * that they have already been removed from the input channel. If the input channel |
| * becomes empty, the client may need to do extra work to ensure that it processes |
| * the deferred event despite the fact that the input channel's file descriptor |
| * is not readable. |
| * |
| * One option is simply to call consume() in a loop until it returns WOULD_BLOCK. |
| * This guarantees that all deferred events will be processed. |
| * |
| * Alternately, the caller can call hasDeferredEvent() to determine whether there is |
| * a deferred event waiting and then ensure that its event loop wakes up at least |
| * one more time to consume the deferred event. |
| */ |
| bool hasDeferredEvent() const; |
| |
| /* Returns true if there is a pending batch. |
| * |
| * Should be called after calling consume() with consumeBatches == false to determine |
| * whether consume() should be called again later on with consumeBatches == true. |
| */ |
| bool hasPendingBatch() const; |
| |
| /* Returns the source of first pending batch if exist. |
| * |
| * Should be called after calling consume() with consumeBatches == false to determine |
| * whether consume() should be called again later on with consumeBatches == true. |
| */ |
| int32_t getPendingBatchSource() const; |
| |
| std::string dump() const; |
| |
| private: |
| // True if touch resampling is enabled. |
| const bool mResampleTouch; |
| |
| std::shared_ptr<InputChannel> mChannel; |
| |
| // The current input message. |
| InputMessage mMsg; |
| |
| // True if mMsg contains a valid input message that was deferred from the previous |
| // call to consume and that still needs to be handled. |
| bool mMsgDeferred; |
| |
| // Batched motion events per device and source. |
| struct Batch { |
| std::vector<InputMessage> samples; |
| }; |
| std::vector<Batch> mBatches; |
| |
| // Touch state per device and source, only for sources of class pointer. |
| struct History { |
| nsecs_t eventTime; |
| BitSet32 idBits; |
| int32_t idToIndex[MAX_POINTER_ID + 1]; |
| PointerCoords pointers[MAX_POINTERS]; |
| |
| void initializeFrom(const InputMessage& msg) { |
| eventTime = msg.body.motion.eventTime; |
| idBits.clear(); |
| for (uint32_t i = 0; i < msg.body.motion.pointerCount; i++) { |
| uint32_t id = msg.body.motion.pointers[i].properties.id; |
| idBits.markBit(id); |
| idToIndex[id] = i; |
| pointers[i].copyFrom(msg.body.motion.pointers[i].coords); |
| } |
| } |
| |
| void initializeFrom(const History& other) { |
| eventTime = other.eventTime; |
| idBits = other.idBits; // temporary copy |
| for (size_t i = 0; i < other.idBits.count(); i++) { |
| uint32_t id = idBits.clearFirstMarkedBit(); |
| int32_t index = other.idToIndex[id]; |
| idToIndex[id] = index; |
| pointers[index].copyFrom(other.pointers[index]); |
| } |
| idBits = other.idBits; // final copy |
| } |
| |
| const PointerCoords& getPointerById(uint32_t id) const { |
| return pointers[idToIndex[id]]; |
| } |
| |
| bool hasPointerId(uint32_t id) const { |
| return idBits.hasBit(id); |
| } |
| }; |
| struct TouchState { |
| int32_t deviceId; |
| int32_t source; |
| size_t historyCurrent; |
| size_t historySize; |
| History history[2]; |
| History lastResample; |
| |
| void initialize(int32_t deviceId, int32_t source) { |
| this->deviceId = deviceId; |
| this->source = source; |
| historyCurrent = 0; |
| historySize = 0; |
| lastResample.eventTime = 0; |
| lastResample.idBits.clear(); |
| } |
| |
| void addHistory(const InputMessage& msg) { |
| historyCurrent ^= 1; |
| if (historySize < 2) { |
| historySize += 1; |
| } |
| history[historyCurrent].initializeFrom(msg); |
| } |
| |
| const History* getHistory(size_t index) const { |
| return &history[(historyCurrent + index) & 1]; |
| } |
| |
| bool recentCoordinatesAreIdentical(uint32_t id) const { |
| // Return true if the two most recently received "raw" coordinates are identical |
| if (historySize < 2) { |
| return false; |
| } |
| if (!getHistory(0)->hasPointerId(id) || !getHistory(1)->hasPointerId(id)) { |
| return false; |
| } |
| float currentX = getHistory(0)->getPointerById(id).getX(); |
| float currentY = getHistory(0)->getPointerById(id).getY(); |
| float previousX = getHistory(1)->getPointerById(id).getX(); |
| float previousY = getHistory(1)->getPointerById(id).getY(); |
| if (currentX == previousX && currentY == previousY) { |
| return true; |
| } |
| return false; |
| } |
| }; |
| std::vector<TouchState> mTouchStates; |
| |
| // Chain of batched sequence numbers. When multiple input messages are combined into |
| // a batch, we append a record here that associates the last sequence number in the |
| // batch with the previous one. When the finished signal is sent, we traverse the |
| // chain to individually finish all input messages that were part of the batch. |
| struct SeqChain { |
| uint32_t seq; // sequence number of batched input message |
| uint32_t chain; // sequence number of previous batched input message |
| }; |
| std::vector<SeqChain> mSeqChains; |
| |
| // The time at which each event with the sequence number 'seq' was consumed. |
| // This data is provided in 'finishInputEvent' so that the receiving end can measure the latency |
| // This collection is populated when the event is received, and the entries are erased when the |
| // events are finished. It should not grow infinitely because if an event is not ack'd, ANR |
| // will be raised for that connection, and no further events will be posted to that channel. |
| std::unordered_map<uint32_t /*seq*/, nsecs_t /*consumeTime*/> mConsumeTimes; |
| |
| status_t consumeBatch(InputEventFactoryInterface* factory, |
| nsecs_t frameTime, uint32_t* outSeq, InputEvent** outEvent); |
| status_t consumeSamples(InputEventFactoryInterface* factory, |
| Batch& batch, size_t count, uint32_t* outSeq, InputEvent** outEvent); |
| |
| void updateTouchState(InputMessage& msg); |
| void resampleTouchState(nsecs_t frameTime, MotionEvent* event, |
| const InputMessage *next); |
| |
| ssize_t findBatch(int32_t deviceId, int32_t source) const; |
| ssize_t findTouchState(int32_t deviceId, int32_t source) const; |
| |
| nsecs_t getConsumeTime(uint32_t seq) const; |
| void popConsumeTime(uint32_t seq); |
| status_t sendUnchainedFinishedSignal(uint32_t seq, bool handled); |
| |
| static void rewriteMessage(TouchState& state, InputMessage& msg); |
| static void initializeKeyEvent(KeyEvent* event, const InputMessage* msg); |
| static void initializeMotionEvent(MotionEvent* event, const InputMessage* msg); |
| static void initializeFocusEvent(FocusEvent* event, const InputMessage* msg); |
| static void initializeCaptureEvent(CaptureEvent* event, const InputMessage* msg); |
| static void initializeDragEvent(DragEvent* event, const InputMessage* msg); |
| static void addSample(MotionEvent* event, const InputMessage* msg); |
| static bool canAddSample(const Batch& batch, const InputMessage* msg); |
| static ssize_t findSampleNoLaterThan(const Batch& batch, nsecs_t time); |
| static bool shouldResampleTool(int32_t toolType); |
| |
| static bool isTouchResamplingEnabled(); |
| }; |
| |
| } // namespace android |
| |
| #endif // _LIBINPUT_INPUT_TRANSPORT_H |