直接从FingerprintManager.java类开始讲起。
首先从enroll方法作为入口
/**
* Request fingerprint enrollment. This call warms up the fingerprint hardware
* and starts scanning for fingerprints. Progress will be indicated by callbacks to the
* {@link EnrollmentCallback} object. It terminates when
* {@link EnrollmentCallback#onEnrollmentError(int, CharSequence)} or
* {@link EnrollmentCallback#onEnrollmentProgress(int) is called with remaining == 0, at
* which point the object is no longer valid. The operation can be canceled by using the
* provided cancel object.
* @param token a unique token provided by a recent creation or verification of device
* credentials (e.g. pin, pattern or password).
* @param cancel an object that can be used to cancel enrollment
* @param userId the user to whom this fingerprint will belong to
* @param callback an object to receive enrollment events
* @param shouldLogMetrics a flag that indicates if enrollment failure/success metrics
* should be logged.
* @hide
*/
@RequiresPermission(MANAGE_FINGERPRINT)
public void enroll(byte [] hardwareAuthToken, CancellationSignal cancel, int userId,
EnrollmentCallback callback, @EnrollReason int enrollReason) {
if (userId == UserHandle.USER_CURRENT) {
userId = getCurrentUserId();
}
if (callback == null) {
throw new IllegalArgumentException("Must supply an enrollment callback");
}
if (cancel != null) {
if (cancel.isCanceled()) {
Slog.w(TAG, "enrollment already canceled");
return;
} else {
cancel.setOnCancelListener(new OnEnrollCancelListener());
}
}
if (mService != null) {
try {
mEnrollmentCallback = callback;
mService.enroll(mToken, hardwareAuthToken, userId, mServiceReceiver,
mContext.getOpPackageName(), enrollReason);
} catch (RemoteException e) {
Slog.w(TAG, "Remote exception in enroll: ", e);
// Though this may not be a hardware issue, it will cause apps to give up or try
// again later.
callback.onEnrollmentError(FINGERPRINT_ERROR_HW_UNAVAILABLE,
getErrorString(mContext, FINGERPRINT_ERROR_HW_UNAVAILABLE,
0 /* vendorCode */));
}
}
}见第37行，变量mService
下面看看变量mService对象是哪个？通过souce insight 工具搜索
有这个定义private IFingerprintService mService;
然后继续往下找在哪儿定义的
public FingerprintManager(Context context, IFingerprintService service) {
mContext = context;
mService = service;
if (mService == null) {
Slog.v(TAG, "FingerprintService was null");
}
mHandler = new MyHandler(context);
}在其他地方初始化 FingerprintManager的时候传进来的，然后在KeyguardUpdate类中找到以下代码
if (mContext.getPackageManager().hasSystemFeature(PackageManager.FEATURE_FINGERPRINT)) {
mFpm = (FingerprintManager) context.getSystemService(Context.FINGERPRINT_SERVICE);
}显然mFpm是个系统服务，继续往下找。搜到类SystemServieRegistry.java
registerService(Context.FINGERPRINT_SERVICE, FingerprintManager.class,
new CachedServiceFetcher<FingerprintManager>() {
@Override
public FingerprintManager createService(ContextImpl ctx) throws ServiceNotFoundException {
final IBinder binder;
if (ctx.getApplicationInfo().targetSdkVersion >= Build.VERSION_CODES.O) {
binder = ServiceManager.getServiceOrThrow(Context.FINGERPRINT_SERVICE);
} else {
binder = ServiceManager.getService(Context.FINGERPRINT_SERVICE);
}
IFingerprintService service = IFingerprintService.Stub.asInterface(binder);
return new FingerprintManager(ctx.getOuterContext(), service);
}});mService变量实际上就是通过aidl进行访问。
service 就是继承IFingerprintService的FingerprintService类。
就是说，上面的mService 就是 要到FingerprintService类中去寻找对象实例了。
然后来到FingerprintService类，查找对应的enroll方法
mService.enroll(mToken, hardwareAuthToken, userId, mServiceReceiver,
mContext.getOpPackageName(), enrollReason);
看看FingerprintService.java
@Override // Binder call
public void enroll(final IBinder token, @NonNull final byte[] hardwareAuthToken,
final int userId, final IFingerprintServiceReceiver receiver,
final String opPackageName, @FingerprintManager.EnrollReason int enrollReason) {
Utils.checkPermission(getContext(), MANAGE_FINGERPRINT);
final Pair<Integer, ServiceProvider> provider = getSingleProvider();
if (provider == null) {
Slog.w(TAG, "Null provider for enroll");
return;
}
provider.second.scheduleEnroll(provider.first, token, hardwareAuthToken, userId,
receiver, opPackageName, enrollReason);
}第13行的部分，provider.second 的详细分析，见下面博客
生物解锁--指纹服务注册流程_liujun3512159的博客-CSDN博客
return provider.second.scheduleEnroll方法调用，实际上就是调用Fingerprint21对象实例。注意 这里的second的用法，可以参照pair使用规则。
然后继续看Fingerprint21.java类的scheduleEnroll方法。
@Override
public void scheduleEnroll(int sensorId, @NonNull IBinder token,
@NonNull byte[] hardwareAuthToken, int userId,
@NonNull IFingerprintServiceReceiver receiver, @NonNull String opPackageName,
@FingerprintManager.EnrollReason int enrollReason) {
mHandler.post(() -> {
scheduleUpdateActiveUserWithoutHandler(userId);
final FingerprintEnrollClient client = new FingerprintEnrollClient(mContext,
mLazyDaemon, token, new ClientMonitorCallbackConverter(receiver), userId,
hardwareAuthToken, opPackageName, FingerprintUtils.getLegacyInstance(mSensorId),
ENROLL_TIMEOUT_SEC, mSensorProperties.sensorId, mUdfpsOverlayController,
mSidefpsController, enrollReason);
mScheduler.scheduleClientMonitor(client, new BaseClientMonitor.Callback() {
@Override
public void onClientStarted(@NonNull BaseClientMonitor clientMonitor) {
mFingerprintStateCallback.onClientStarted(clientMonitor);
}
@Override
public void onClientFinished(@NonNull BaseClientMonitor clientMonitor,
boolean success) {
mFingerprintStateCallback.onClientFinished(clientMonitor, success);
if (success) {
// Update authenticatorIds
scheduleUpdateActiveUserWithoutHandler(clientMonitor.getTargetUserId(),
true /* force */);
}
}
});
});
}见第14行，实际上是调用类BiometricScheduler下的方法scheduleClientMonitor
public void scheduleClientMonitor(@NonNull BaseClientMonitor clientMonitor,
@Nullable BaseClientMonitor.Callback clientCallback) {
// If the incoming operation should interrupt preceding clients, mark any interruptable
// pending clients as canceling. Once they reach the head of the queue, the scheduler will
// send ERROR_CANCELED and skip the operation.
if (clientMonitor.interruptsPrecedingClients()) {
for (Operation operation : mPendingOperations) {
if (operation.mClientMonitor instanceof Interruptable
&& operation.mState != Operation.STATE_WAITING_IN_QUEUE_CANCELING) {
Slog.d(getTag(), "New client incoming, marking pending client as canceling: "
+ operation.mClientMonitor);
operation.mState = Operation.STATE_WAITING_IN_QUEUE_CANCELING;
}
}
}
mPendingOperations.add(new Operation(clientMonitor, clientCallback));
Slog.d(getTag(), "[Added] " + clientMonitor
+ ", new queue size: " + mPendingOperations.size());
// If the new operation should interrupt preceding clients, and if the current operation is
// cancellable, start the cancellation process.
if (clientMonitor.interruptsPrecedingClients()
&& mCurrentOperation != null
&& mCurrentOperation.mClientMonitor instanceof Interruptable
&& mCurrentOperation.mState == Operation.STATE_STARTED) {
Slog.d(getTag(), "[Cancelling Interruptable]: " + mCurrentOperation);
cancelInternal(mCurrentOperation);
}
startNextOperationIfIdle();
}第17行，很重要，后面会说到
第31行，继续往下
protected void startNextOperationIfIdle() {
if (mCurrentOperation != null) {
Slog.v(getTag(), "Not idle, current operation: " + mCurrentOperation);
return;
}
if (mPendingOperations.isEmpty()) {
Slog.d(getTag(), "No operations, returning to idle");
return;
}
mCurrentOperation = mPendingOperations.poll();
final BaseClientMonitor currentClient = mCurrentOperation.mClientMonitor;
Slog.d(getTag(), "[Polled] " + mCurrentOperation);
// If the operation at the front of the queue has been marked for cancellation, send
// ERROR_CANCELED. No need to start this client.
if (mCurrentOperation.mState == Operation.STATE_WAITING_IN_QUEUE_CANCELING) {
Slog.d(getTag(), "[Now Cancelling] " + mCurrentOperation);
if (!(currentClient instanceof Interruptable)) {
throw new IllegalStateException("Mis-implemented client or scheduler, "
+ "trying to cancel non-interruptable operation: " + mCurrentOperation);
}
final Interruptable interruptable = (Interruptable) currentClient;
interruptable.cancelWithoutStarting(getInternalCallback());
// Now we wait for the client to send its FinishCallback, which kicks off the next
// operation.
return;
}
if (mGestureAvailabilityDispatcher != null
&& mCurrentOperation.mClientMonitor instanceof AcquisitionClient) {
mGestureAvailabilityDispatcher.markSensorActive(
mCurrentOperation.mClientMonitor.getSensorId(),
true /* active */);
}
// Not all operations start immediately. BiometricPrompt waits for its operation
// to arrive at the head of the queue, before pinging it to start.
final boolean shouldStartNow = currentClient.getCookie() == 0;
if (shouldStartNow) {
if (mCurrentOperation.isUnstartableHalOperation()) {
final HalClientMonitor<?> halClientMonitor =
(HalClientMonitor<?>) mCurrentOperation.mClientMonitor;
// Note down current length of queue
final int pendingOperationsLength = mPendingOperations.size();
final Operation lastOperation = mPendingOperations.peekLast();
Slog.e(getTag(), "[Unable To Start] " + mCurrentOperation
+ ". Last pending operation: " + lastOperation);
// For current operations, 1) unableToStart, which notifies the caller-side, then
// 2) notify operation's callback, to notify applicable system service that the
// operation failed.
halClientMonitor.unableToStart();
if (mCurrentOperation.mClientCallback != null) {
mCurrentOperation.mClientCallback.onClientFinished(
mCurrentOperation.mClientMonitor, false /* success */);
}
// Then for each operation currently in the pending queue at the time of this
// failure, do the same as above. Otherwise, it's possible that something like
// setActiveUser fails, but then authenticate (for the wrong user) is invoked.
for (int i = 0; i < pendingOperationsLength; i++) {
final Operation operation = mPendingOperations.pollFirst();
if (operation == null) {
Slog.e(getTag(), "Null operation, index: " + i
+ ", expected length: " + pendingOperationsLength);
break;
}
if (operation.isHalOperation()) {
((HalClientMonitor<?>) operation.mClientMonitor).unableToStart();
}
if (operation.mClientCallback != null) {
operation.mClientCallback.onClientFinished(operation.mClientMonitor,
false /* success */);
}
Slog.w(getTag(), "[Aborted Operation] " + operation);
}
// It's possible that during cleanup a new set of operations came in. We can try to
// run these. A single request from the manager layer to the service layer may
// actually be multiple operations (i.e. updateActiveUser + authenticate).
mCurrentOperation = null;
startNextOperationIfIdle();
} else {
Slog.d(getTag(), "[Starting] " + mCurrentOperation);
currentClient.start(getInternalCallback());
mCurrentOperation.mState = Operation.STATE_STARTED;
}
} else {
try {
mBiometricService.onReadyForAuthentication(currentClient.getCookie());
} catch (RemoteException e) {
Slog.e(getTag(), "Remote exception when contacting BiometricService", e);
}
Slog.d(getTag(), "Waiting for cookie before starting: " + mCurrentOperation);
mCurrentOperation.mState = Operation.STATE_WAITING_FOR_COOKIE;
}
}第88行代码，需要先弄清楚currentClient变量。
首先是
final BaseClientMonitor currentClient = mCurrentOperation.mClientMonitor;
然后继续追踪mCurrentOperation。然后搜索到前面的代码中
mPendingOperations.add(new Operation(clientMonitor, clientCallback));
Operation(
@NonNull BaseClientMonitor clientMonitor,
@Nullable BaseClientMonitor.Callback callback
) {
this(clientMonitor, callback, STATE_WAITING_IN_QUEUE);
}很容易分析出clientMonitor 就是mCurrentOperation.mClientMonitor 对象实例
那继续追踪clientMonitor，在方法scheduleClientMonitor中传递进来的。
搜索 mScheduler.scheduleClientMonitor ，发现在Fingerprint21.java类的scheduleEnroll方法定义的。见文章开头部分。
最终FingerprintEnrollClient 对象实例就是clientMonitor ，也就是mClientMonitor。
所以，第88行的
currentClient.start(getInternalCallback());
实际上是调用的FingerprintEnrollClient对象下的start方法。
然后在其父类中找到其方法
@Override
public void start(@NonNull Callback callback) {
super.start(callback);
if (hasReachedEnrollmentLimit()) {
Slog.e(TAG, "Reached enrollment limit");
callback.onClientFinished(this, false /* success */);
return;
}
mEnrollmentStartTimeMs = System.currentTimeMillis();
startHalOperation();
}第12行，startHalOperation方法，实际上是调用的FingerprintEnrollClient下的。
@Override
protected void startHalOperation() {
mSensorOverlays.show(getSensorId(), getOverlayReasonFromEnrollReason(mEnrollReason), this);
BiometricNotificationUtils.cancelBadCalibrationNotification(getContext());
try {
mCancellationSignal = getFreshDaemon().enroll(
HardwareAuthTokenUtils.toHardwareAuthToken(mHardwareAuthToken));
} catch (RemoteException e) {
Slog.e(TAG, "Remote exception when requesting enroll", e);
onError(BiometricFingerprintConstants.FINGERPRINT_ERROR_UNABLE_TO_PROCESS,
0 /* vendorCode */);
mCallback.onClientFinished(this, false /* success */);
}
}第7行，实际上就是跨进程调用远程对象BiometricsFingerprint.cpp下的方法enroll方法。
Return<RequestStatus> BiometricsFingerprint::enroll(const hidl_array<uint8_t, 69>& hat,
uint32_t gid, uint32_t timeoutSec) {
const hw_auth_token_t* authToken =
reinterpret_cast<const hw_auth_token_t*>(hat.data());
return ErrorFilter(mDevice->enroll(mDevice, authToken, gid, timeoutSec));
}到这里，指纹录入流程，已经关联到指纹模组了，下面再分析，指纹录入（可以说是指纹按压）的流程。
下面的部分，是我猜测的，这段代码，我感觉应该是hal层或者说硬件厂商指纹接受录入的指纹，然后回调到这里的，然后有这个地方层层上报到最上层，告诉用户，正在录入指纹。
BiometricsFingerprint.cpp
void BiometricsFingerprint::notify(const fingerprint_msg_t *msg) {
BiometricsFingerprint* thisPtr = static_cast<BiometricsFingerprint*>(
BiometricsFingerprint::getInstance());
std::lock_guard<std::mutex> lock(thisPtr->mClientCallbackMutex);
if (thisPtr == nullptr || thisPtr->mClientCallback == nullptr) {
ALOGE("Receiving callbacks before the client callback is registered.");
return;
}
const uint64_t devId = reinterpret_cast<uint64_t>(thisPtr->mDevice);
switch (msg->type) {
case FINGERPRINT_ERROR: {
int32_t vendorCode = 0;
FingerprintError result = VendorErrorFilter(msg->data.error, &vendorCode);
ALOGD("onError(%d)", result);
if (!thisPtr->mClientCallback->onError(devId, result, vendorCode).isOk()) {
ALOGE("failed to invoke fingerprint onError callback");
}
}
break;
case FINGERPRINT_ACQUIRED: {
int32_t vendorCode = 0;
FingerprintAcquiredInfo result =
VendorAcquiredFilter(msg->data.acquired.acquired_info, &vendorCode);
ALOGD("onAcquired(%d)", result);
if (!thisPtr->mClientCallback->onAcquired(devId, result, vendorCode).isOk()) {
ALOGE("failed to invoke fingerprint onAcquired callback");
}
}
break;
case FINGERPRINT_TEMPLATE_ENROLLING:
ALOGD("onEnrollResult(fid=%d, gid=%d, rem=%d)",
msg->data.enroll.finger.fid,
msg->data.enroll.finger.gid,
msg->data.enroll.samples_remaining);
if (!thisPtr->mClientCallback->onEnrollResult(devId,
msg->data.enroll.finger.fid,
msg->data.enroll.finger.gid,
msg->data.enroll.samples_remaining).isOk()) {
ALOGE("failed to invoke fingerprint onEnrollResult callback");
}
break;
case FINGERPRINT_TEMPLATE_REMOVED:
ALOGD("onRemove(fid=%d, gid=%d, rem=%d)",
msg->data.removed.finger.fid,
msg->data.removed.finger.gid,
msg->data.removed.remaining_templates);
if (!thisPtr->mClientCallback->onRemoved(devId,
msg->data.removed.finger.fid,
msg->data.removed.finger.gid,
msg->data.removed.remaining_templates).isOk()) {
ALOGE("failed to invoke fingerprint onRemoved callback");
}
break;
case FINGERPRINT_AUTHENTICATED:
if (msg->data.authenticated.finger.fid != 0) {
ALOGD("onAuthenticated(fid=%d, gid=%d)",
msg->data.authenticated.finger.fid,
msg->data.authenticated.finger.gid);
const uint8_t* hat =
reinterpret_cast<const uint8_t *>(&msg->data.authenticated.hat);
const hidl_vec<uint8_t> token(
std::vector<uint8_t>(hat, hat + sizeof(msg->data.authenticated.hat)));
if (!thisPtr->mClientCallback->onAuthenticated(devId,
msg->data.authenticated.finger.fid,
msg->data.authenticated.finger.gid,
token).isOk()) {
ALOGE("failed to invoke fingerprint onAuthenticated callback");
}
} else {
// Not a recognized fingerprint
if (!thisPtr->mClientCallback->onAuthenticated(devId,
msg->data.authenticated.finger.fid,
msg->data.authenticated.finger.gid,
hidl_vec<uint8_t>()).isOk()) {
ALOGE("failed to invoke fingerprint onAuthenticated callback");
}
}
break;
case FINGERPRINT_TEMPLATE_ENUMERATING:
ALOGD("onEnumerate(fid=%d, gid=%d, rem=%d)",
msg->data.enumerated.finger.fid,
msg->data.enumerated.finger.gid,
msg->data.enumerated.remaining_templates);
if (!thisPtr->mClientCallback->onEnumerate(devId,
msg->data.enumerated.finger.fid,
msg->data.enumerated.finger.gid,
msg->data.enumerated.remaining_templates).isOk()) {
ALOGE("failed to invoke fingerprint onEnumerate callback");
}
break;
}
}见第35行，!thisPtr->mClientCallback->onEnrollResult
这里的mClientCallback，在下面的博客中，有详细的介绍。直接直接给出结果，实际上就是HalResultController类对象。
指纹设别流程_liujun3512159的博客-CSDN博客
直接看源码
public static class HalResultController extends IBiometricsFingerprintClientCallback.Stub {
/**
* Interface to sends results to the HalResultController's owner.
*/
public interface Callback {
/**
* Invoked when the HAL sends ERROR_HW_UNAVAILABLE.
*/
void onHardwareUnavailable();
}
private final int mSensorId;
@NonNull private final Context mContext;
@NonNull final Handler mHandler;
@NonNull final BiometricScheduler mScheduler;
@Nullable private Callback mCallback;
HalResultController(int sensorId, @NonNull Context context, @NonNull Handler handler,
@NonNull BiometricScheduler scheduler) {
mSensorId = sensorId;
mContext = context;
mHandler = handler;
mScheduler = scheduler;
}
public void setCallback(@Nullable Callback callback) {
mCallback = callback;
}
@Override
public void onEnrollResult(long deviceId, int fingerId, int groupId, int remaining) {
mHandler.post(() -> {
final BaseClientMonitor client = mScheduler.getCurrentClient();
if (!(client instanceof FingerprintEnrollClient)) {
Slog.e(TAG, "onEnrollResult for non-enroll client: "
+ Utils.getClientName(client));
return;
}
final int currentUserId = client.getTargetUserId();
final CharSequence name = FingerprintUtils.getLegacyInstance(mSensorId)
.getUniqueName(mContext, currentUserId);
final Fingerprint fingerprint = new Fingerprint(name, groupId, fingerId, deviceId);
final FingerprintEnrollClient enrollClient = (FingerprintEnrollClient) client;
enrollClient.onEnrollResult(fingerprint, remaining);
});
}
@Override
public void onAcquired(long deviceId, int acquiredInfo, int vendorCode) {
onAcquired_2_2(deviceId, acquiredInfo, vendorCode);
}
@Override
public void onAcquired_2_2(long deviceId, int acquiredInfo, int vendorCode) {
mHandler.post(() -> {
final BaseClientMonitor client = mScheduler.getCurrentClient();
if (!(client instanceof AcquisitionClient)) {
Slog.e(TAG, "onAcquired for non-acquisition client: "
+ Utils.getClientName(client));
return;
}
final AcquisitionClient<?> acquisitionClient = (AcquisitionClient<?>) client;
acquisitionClient.onAcquired(acquiredInfo, vendorCode);
});
}
@Override
public void onAuthenticated(long deviceId, int fingerId, int groupId,
ArrayList<Byte> token) {
mHandler.post(() -> {
final BaseClientMonitor client = mScheduler.getCurrentClient();
if (!(client instanceof AuthenticationConsumer)) {
Slog.e(TAG, "onAuthenticated for non-authentication consumer: "
+ Utils.getClientName(client));
return;
}
final AuthenticationConsumer authenticationConsumer =
(AuthenticationConsumer) client;
final boolean authenticated = fingerId != 0;
final Fingerprint fp = new Fingerprint("", groupId, fingerId, deviceId);
authenticationConsumer.onAuthenticated(fp, authenticated, token);
});
}
@Override
public void onError(long deviceId, int error, int vendorCode) {
mHandler.post(() -> {
final BaseClientMonitor client = mScheduler.getCurrentClient();
Slog.d(TAG, "handleError"
+ ", client: " + Utils.getClientName(client)
+ ", error: " + error
+ ", vendorCode: " + vendorCode);
if (!(client instanceof ErrorConsumer)) {
Slog.e(TAG, "onError for non-error consumer: " + Utils.getClientName(client));
return;
}
final ErrorConsumer errorConsumer = (ErrorConsumer) client;
errorConsumer.onError(error, vendorCode);
if (error == BiometricConstants.BIOMETRIC_ERROR_HW_UNAVAILABLE) {
Slog.e(TAG, "Got ERROR_HW_UNAVAILABLE");
if (mCallback != null) {
mCallback.onHardwareUnavailable();
}
}
});
}
@Override
public void onRemoved(long deviceId, int fingerId, int groupId, int remaining) {
mHandler.post(() -> {
Slog.d(TAG, "Removed, fingerId: " + fingerId + ", remaining: " + remaining);
final BaseClientMonitor client = mScheduler.getCurrentClient();
if (!(client instanceof RemovalConsumer)) {
Slog.e(TAG, "onRemoved for non-removal consumer: "
+ Utils.getClientName(client));
return;
}
final Fingerprint fp = new Fingerprint("", groupId, fingerId, deviceId);
final RemovalConsumer removalConsumer = (RemovalConsumer) client;
removalConsumer.onRemoved(fp, remaining);
});
}
@Override
public void onEnumerate(long deviceId, int fingerId, int groupId, int remaining) {
mHandler.post(() -> {
final BaseClientMonitor client = mScheduler.getCurrentClient();
if (!(client instanceof EnumerateConsumer)) {
Slog.e(TAG, "onEnumerate for non-enumerate consumer: "
+ Utils.getClientName(client));
return;
}
final Fingerprint fp = new Fingerprint("", groupId, fingerId, deviceId);
final EnumerateConsumer enumerateConsumer = (EnumerateConsumer) client;
enumerateConsumer.onEnumerationResult(fp, remaining);
});
}
}在第45行，其实这里的enrollClient是在前面的方法scheduleEnroll中传进来的，实际上就是FingerprintEnrollClient
继续追踪 FingerprintEnrollClient下的onEnrollResult 方法
@Override
public void onEnrollResult(BiometricAuthenticator.Identifier identifier, int remaining) {
super.onEnrollResult(identifier, remaining);
mSensorOverlays.ifUdfps(
controller -> controller.onEnrollmentProgress(getSensorId(), remaining));
if (remaining == 0) {
mSensorOverlays.hide(getSensorId());
}
}分析到这里，基本上结束，注意第3行，调用uper.onEnrollResult(identifier, remaining);
其实现逻辑在方法scheduleEnroll中，见最上面。
第6行的
controller.onEnrollmentProgress 就是录入指纹的进度条方法了。抽空也补充下。
487
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
