Caused by java.lang.IllegalStateException Not allowed to start service Intent { cmp=com.x.x.x/.x.x.xService }: app is in background uid UidRecord问题原因分析(二)
应用在适配Android 8.0以上系统时，会发现后台启动不了服务，会报出如下异常，并强退：
Fatal Exception: java.lang.IllegalStateException
Not allowed to start service Intent { act=com.xxx.xxx.xxx pkg=com.xxx.xxx (has extras) }: app is in background uid UidRecord{1cbd9ed u0a1967 CEM idle procs:1 seq(0,0,0)}
问题原因分析
Android 8.0 行为变更
https://developer.android.com/about/versions/oreo/android-8.0-changes.html#back-all
Android 8.0 除了提供诸多新特性和功能外，还对系统和 API 行为做出了各种变更。本文重点介绍您应该了解并在开发应用时加以考虑的一些主要变更。
其中大部分变更会影响所有应用，而不论应用针对的是何种版本的 Android。不过，有几项变更仅影响针对 Android 8.0 的应用。为清楚起见，本页面分为两个部分：针对所有 API 级别的应用和针对 Android 8.0 的应用。
针对所有 API 级别的应用
这些行为变更适用于 在 Android 8.0 平台上运行的 所有应用，无论这些应用是针对哪个 API 级别构建。所有开发者都应查看这些变更，并修改其应用以正确支持这些变更（如果适用）。
后台执行限制Android 8.0 为提高电池续航时间而引入的变更之一是，当您的应用进入已缓存状态时，如果没有活动的组件，系统将解除应用具有的所有唤醒锁。
此外，为提高设备性能，系统会限制未在前台运行的应用的某些行为。具体而言：
- 现在，在后台运行的应用对后台服务的访问受到限制。
- 应用无法使用其清单注册大部分隐式广播（即，并非专门针对此应用的广播）。
默认情况下，这些限制仅适用于针对 O 的应用。不过，用户可以从 Settings 屏幕为任意应用启用这些限制，即使应用并不是以 O 为目标平台。
Android 8.0 还对特定函数做出了以下变更：
- 如果针对 Android 8.0 的应用尝试在不允许其创建后台服务的情况下使用
startService()函数，则该函数将引发一个IllegalStateException。 - 新的
Context.startForegroundService()函数将启动一个前台服务。现在，即使应用在后台运行，系统也允许其调用Context.startForegroundService()。不过，应用必须在创建服务后的五秒内调用该服务的startForeground()函数。
如需了解详细信息，请参阅后台执行限制。
出错代码定位
在ContextImpl的startServiceCommon函数中爆出异常，
http://androidxref.com/8.1.0_r33/xref/frameworks/base/core/java/android/app/ContextImpl.java
private ComponentName startServiceCommon(Intent service, boolean requireForeground, UserHandle user) { try { validateServiceIntent(service); service.prepareToLeaveProcess(this); ComponentName cn = ActivityManager.getService().startService( mMainThread.getApplicationThread(), service, service.resolveTypeIfNeeded( getContentResolver()), requireForeground, getOpPackageName(), user.getIdentifier()); if (cn != null) { if (cn.getPackageName().equals("!")) { throw new SecurityException( "Not allowed to start service " + service + " without permission " + cn.getClassName()); } else if (cn.getPackageName().equals("!!")) { throw new SecurityException( "Unable to start service " + service + ": " + cn.getClassName()); //此处就是曝出异常的地方，非法状态，不允许启动服务 } else if (cn.getPackageName().equals("?")) { throw new IllegalStateException( "Not allowed to start service " + service + ": " + cn.getClassName()); } } return cn; } catch (RemoteException e) { throw e.rethrowFromSystemServer(); } }
startServiceCommon这个函数做的操作是AMS的startService，用于启动服务
AMS的startService
接下去看AMS的startService，稍微注意一下传递的参数，里面有一个前台后台相关的requireForeground，可能跟问题有关系。
AMS代码位置
@Override public ComponentName startService(IApplicationThread caller, Intent service, String resolvedType, boolean requireForeground, String callingPackage, int userId) throws TransactionTooLargeException { enforceNotIsolatedCaller("startService"); // Refuse possible leaked file descriptors if (service != null && service.hasFileDescriptors() == true) { throw new IllegalArgumentException("File descriptors passed in Intent"); } if (callingPackage == null) { throw new IllegalArgumentException("callingPackage cannot be null"); } if (DEBUG_SERVICE) Slog.v(TAG_SERVICE, "*** startService: " + service + " type=" + resolvedType + " fg=" + requireForeground); synchronized(this) { final int callingPid = Binder.getCallingPid(); final int callingUid = Binder.getCallingUid(); final long origId = Binder.clearCallingIdentity(); ComponentName res; try { //调用ActiveServices的startServiceLocked res = mServices.startServiceLocked(caller, service, resolvedType, callingPid, callingUid, requireForeground, callingPackage, userId); } finally { Binder.restoreCallingIdentity(origId); } return res; } }
会调用ActiveServices的startServiceLocked
ActiveServices的startServiceLocked
ComponentName startServiceLocked(IApplicationThread caller, Intent service, String resolvedType, int callingPid, int callingUid, boolean fgRequired, String callingPackage, final int userId) throws TransactionTooLargeException { //... // 启动服务之前有2个判断一个是startRequested，一个是fgRequired。 // startRequested代表的是：是否已经启动过服务，一般出现问题都是启动一个没有运行的服务， // 那么这个就是false。 // fgRequired这个就是启动服务传递的requireForeground， if (!r.startRequested && !fgRequired) { // 这里面有个关键函数getAppStartModeLocked，判断是否运行启动服务 // 注意此处传递的最后2个参数：alwaysRestrict和disabledOnly都是false final int allowed = mAm.getAppStartModeLocked(r.appInfo.uid, r.packageName, r.appInfo.targetSdkVersion, callingPid, false, false); // 如果不允许启动服务则会运行到里面 if (allowed != ActivityManager.APP_START_MODE_NORMAL) { //... UidRecord uidRec = mAm.mActiveUids.get(r.appInfo.uid); // 此处就是不允许运行服务返回的原因"app is in background" // 找到了原因就在这里 return new ComponentName("?", "app is in background uid " + uidRec); } } //... }
这里面由于出现错误，那么startRequested==false而且fgRequired==false，说明这个服务是第一次启动，而且是后台请求启动服务。
至于为什么不允许启动服务，我们还需要查看AMS的getAppStartModeLocked函数。
AMS判断并返回服务启动模式
int getAppStartModeLocked(int uid, String packageName, int packageTargetSdk,
int callingPid, boolean alwaysRestrict, boolean disabledOnly) {
UidRecord uidRec = mActiveUids.get(uid);
if (DEBUG_BACKGROUND_CHECK) Slog.d(TAG, "checkAllowBackground: uid=" + uid + " pkg="
+ packageName + " rec=" + uidRec + " always=" + alwaysRestrict + " idle="
+ (uidRec != null ? uidRec.idle : false));
if (uidRec == null || alwaysRestrict || uidRec.idle) {
boolean ephemeral;
if (uidRec == null) {
ephemeral = getPackageManagerInternalLocked().isPackageEphemeral(
UserHandle.getUserId(uid), packageName);
} else {
ephemeral = uidRec.ephemeral;
}
if (ephemeral) {
// We are hard-core about ephemeral apps not running in the background.
return ActivityManager.APP_START_MODE_DISABLED;
} else {
if (disabledOnly) {
// The caller is only interested in whether app starts are completely
// disabled for the given package (that is, it is an instant app). So
// we don't need to go further, which is all just seeing if we should
// apply a "delayed" mode for a regular app.
return ActivityManager.APP_START_MODE_NORMAL;
}
// 此处alwaysRestrict==false，于是调用的是appServicesRestrictedInBackgroundLocked
final int startMode = (alwaysRestrict)
? appRestrictedInBackgroundLocked(uid, packageName, packageTargetSdk)
: appServicesRestrictedInBackgroundLocked(uid, packageName,
packageTargetSdk);
if (DEBUG_BACKGROUND_CHECK) Slog.d(TAG, "checkAllowBackground: uid=" + uid
+ " pkg=" + packageName + " startMode=" + startMode
+ " onwhitelist=" + isOnDeviceIdleWhitelistLocked(uid));
if (startMode == ActivityManager.APP_START_MODE_DELAYED) {
// This is an old app that has been forced into a "compatible as possible"
// mode of background check. To increase compatibility, we will allow other
// foreground apps to cause its services to start.
if (callingPid >= 0) {
ProcessRecord proc;
synchronized (mPidsSelfLocked) {
proc = mPidsSelfLocked.get(callingPid);
}
if (proc != null &&
!ActivityManager.isProcStateBackground(proc.curProcState)) {
// Whoever is instigating this is in the foreground, so we will allow it
// to go through.
return ActivityManager.APP_START_MODE_NORMAL;
}
}
}
return startMode;
}
}
return ActivityManager.APP_START_MODE_NORMAL;
}
根据alwaysRestrict的值会调用appRestrictedInBackgroundLocked或者appServicesRestrictedInBackgroundLocked；
其中appRestrictedInBackgroundLocked是直接根据应用sdk进行判断，
appServicesRestrictedInBackgroundLocked会进行条件过滤，直接运行部分应用启动服务，其它的进行应用sdk的判断。
接下来看appServicesRestrictedInBackgroundLocked进行条件过滤,允许部分启动服务
int appServicesRestrictedInBackgroundLocked(int uid, String packageName, int packageTargetSdk) { // Persistent app? // 如果是常驻内存的，可以直接启动服务 if (mPackageManagerInt.isPackagePersistent(packageName)) { if (DEBUG_BACKGROUND_CHECK) { Slog.i(TAG, "App " + uid + "/" + packageName + " is persistent; not restricted in background"); } return ActivityManager.APP_START_MODE_NORMAL; } // Non-persistent but background whitelisted? // 如果是非常驻内存的话，但是在白名单列表里面的uid也是允许的 // 目前这个白名单里面就只有一个：蓝牙BLUETOOTH_UID = 1002 if (uidOnBackgroundWhitelist(uid)) { if (DEBUG_BACKGROUND_CHECK) { Slog.i(TAG, "App " + uid + "/" + packageName + " on background whitelist; not restricted in background"); } return ActivityManager.APP_START_MODE_NORMAL; } // Is this app on the battery whitelist? // 如果是在电源相关的白名单里面，也是允许启动服务的 if (isOnDeviceIdleWhitelistLocked(uid)) { if (DEBUG_BACKGROUND_CHECK) { Slog.i(TAG, "App " + uid + "/" + packageName + " on idle whitelist; not restricted in background"); } return ActivityManager.APP_START_MODE_NORMAL; } // None of the service-policy criteria apply, so we apply the common criteria return appRestrictedInBackgroundLocked(uid, packageName, packageTargetSdk); }
分别对于：
1. 是否常驻内存应用，常驻内存允许启动服务
2.如果是蓝牙也是允许启动服务
3.是在电源相关的DeviceIdle白名单里面，允许启动服务的
4.如果都不是则执行默认策略appRestrictedInBackgroundLocked
再看appServicesRestrictedInBackgroundLocked进行条件过滤,允许部分启动服务
int appRestrictedInBackgroundLocked(int uid, String packageName, int packageTargetSdk) { // Apps that target O+ are always subject to background check // 如果apk的sdk版本大于AndroidO的话，那么默认是不允许启动服务的 if (packageTargetSdk >= Build.VERSION_CODES.O) { if (DEBUG_BACKGROUND_CHECK) { Slog.i(TAG, "App " + uid + "/" + packageName + " targets O+, restricted"); } return ActivityManager.APP_START_MODE_DELAYED_RIGID; } // ...and legacy apps get an AppOp check // 如果是之前版本的apk，会查看AppOps是否允许后台运行权限， // 由于我们sdk版本肯定会升级的，这个就暂时不考虑了 int appop = mAppOpsService.noteOperation(AppOpsManager.OP_RUN_IN_BACKGROUND, uid, packageName); if (DEBUG_BACKGROUND_CHECK) { Slog.i(TAG, "Legacy app " + uid + "/" + packageName + " bg appop " + appop); } switch (appop) { case AppOpsManager.MODE_ALLOWED: return ActivityManager.APP_START_MODE_NORMAL; case AppOpsManager.MODE_IGNORED: return ActivityManager.APP_START_MODE_DELAYED; default: return ActivityManager.APP_START_MODE_DELAYED_RIGID; } }
如果apk的sdk版本大于AndroidO的话，那么默认是不允许启动服务的，那么要适配Android O/GO以后的版本，此处是绕不过去的坎，建议尽早处理。
网上所传的notification隐藏是否可以？
有的需求是启动服务，但是又不想有通知，这和Google定的规则有冲突呀，有没有什么办法呢？网上2年前的方案是启动两个Service，一个Service干活，另外一个Service把通知隐藏掉。
Android O Google应该考虑到这个漏洞了：
private void cancelForegroundNotificationLocked(ServiceRecord r) { if (r.foregroundId != 0) { // First check to see if this app has any other active foreground services // with the same notification ID. If so, we shouldn't actually cancel it, // because that would wipe away the notification that still needs to be shown // due the other service. ServiceMap sm = getServiceMapLocked(r.userId); if (sm != null) { for (int i = sm.mServicesByName.size()-1; i >= 0; i--) { ServiceRecord other = sm.mServicesByName.valueAt(i); if (other != r && other.foregroundId == r.foregroundId && other.packageName.equals(r.packageName)) { // Found one! Abort the cancel. return; } } } r.cancelNotification(); } }
如果前台服务的通知还有被占用，那就别想用其他服务把它干掉了
框架规避方案
修改方案（仅供参考）：ActiveServices.java如下加一个packageName的crash的规避，anr同理，发出消息的地方可以修改为不发出timeout消息，也可以在startForeground的时候就移除。(如果Service耗时小于5s，Service在stop流程的时候会将anr消息移除，可不修改)
// Check to see if the service had been started as foreground, but being // brought down before actually showing a notification. That is not allowed. if (r.fgRequired) { Slog.w(TAG_SERVICE, "Bringing down service while still waiting for start foreground: " + r); r.fgRequired = false; r.fgWaiting = false; mAm.mHandler.removeMessages( ActivityManagerService.SERVICE_FOREGROUND_TIMEOUT_MSG, r); if (r.app != null && !"packageName".equals(r.packageName)) { Message msg = mAm.mHandler.obtainMessage( ActivityManagerService.SERVICE_FOREGROUND_CRASH_MSG); msg.obj = r.app; mAm.mHandler.sendMessage(msg); } }
修改原理：
编译一个service.jar，打印报错堆栈
01-01 07:02:17.669 918 1334 W ActivityManager: Bringing down service while still waiting for start foreground: ServiceRecord{2d44a2d u0 packageName/.servicename} 01-01 07:02:17.669 918 1334 W ActivityManager: java.lang.Throwable 01-01 07:02:17.669 918 1334 W ActivityManager: at com.android.server.am.ActiveServices.bringDownServiceLocked(ActiveServices.java:2612) 01-01 07:02:17.669 918 1334 W ActivityManager: at com.android.server.am.ActiveServices.bringDownServiceIfNeededLocked(ActiveServices.java:2559) 01-01 07:02:17.669 918 1334 W ActivityManager: at com.android.server.am.ActiveServices.stopServiceTokenLocked(ActiveServices.java:792) 01-01 07:02:17.669 918 1334 W ActivityManager: at com.android.server.am.ActivityManagerService.stopServiceToken(ActivityManagerService.java:18789) 01-01 07:02:17.669 918 1334 W ActivityManager: at android.app.IActivityManager$Stub.onTransact(IActivityManager.java:759) 01-01 07:02:17.669 918 1334 W ActivityManager: at com.android.server.am.ActivityManagerService.onTransact(ActivityManagerService.java:3080) 01-01 07:02:17.669 918 1334 W ActivityManager: at android.os.Binder.execTransact(Binder.java:697)
找到对应抛出Context.startForegroundService() did not then call Service.startForeground()的逻辑代码：
ActiveServices.java bringDownServiceLocked
// Check to see if the service had been started as foreground, but being // brought down before actually showing a notification. That is not allowed. if (r.fgRequired) { Slog.w(TAG_SERVICE, "Bringing down service while still waiting for start foreground: " + r); r.fgRequired = false; r.fgWaiting = false; mAm.mHandler.removeMessages( ActivityManagerService.SERVICE_FOREGROUND_TIMEOUT_MSG, r); if (r.app != null) { Message msg = mAm.mHandler.obtainMessage( ActivityManagerService.SERVICE_FOREGROUND_CRASH_MSG); msg.obj = r.app; mAm.mHandler.sendMessage(msg); } }
走到这里面继而会由ams发出一个service_foreground_crash_msg的消息，导致crash。
至于为嘛会走到这里呢，都是id = 0 的过，既没有走前台服务的流程也没有将r.fgRequired设为false,anr的msg也没有移除掉。
private void setServiceForegroundInnerLocked(ServiceRecord r, int id, Notification notification, int flags) { if (id != 0) { if (notification == null) { throw new IllegalArgumentException("null notification"); } // Instant apps need permission to create foreground services. ... if (r.fgRequired) { if (DEBUG_SERVICE || DEBUG_BACKGROUND_CHECK) { Slog.i(TAG, "Service called startForeground() as required: " + r); } r.fgRequired = false; r.fgWaiting = false; mAm.mHandler.removeMessages( ActivityManagerService.SERVICE_FOREGROUND_TIMEOUT_MSG, r); } if (r.foregroundId != id) { cancelForegroundNotificationLocked(r); r.foregroundId = id; } notification.flags |= Notification.FLAG_FOREGROUND_SERVICE; r.foregroundNoti = notification; if (!r.isForeground) { final ServiceMap smap = getServiceMapLocked(r.userId); if (smap != null) { ActiveForegroundApp active = smap.mActiveForegroundApps.get(r.packageName); if (active == null) { active = new ActiveForegroundApp(); active.mPackageName = r.packageName; active.mUid = r.appInfo.uid; active.mShownWhileScreenOn = mScreenOn; if (r.app != null) { active.mAppOnTop = active.mShownWhileTop = r.app.uidRecord.curProcState <= ActivityManager.PROCESS_STATE_TOP; } active.mStartTime = active.mStartVisibleTime = SystemClock.elapsedRealtime(); smap.mActiveForegroundApps.put(r.packageName, active); requestUpdateActiveForegroundAppsLocked(smap, 0); } active.mNumActive++; } r.isForeground = true; } r.postNotification(); if (r.app != null) { updateServiceForegroundLocked(r.app, true); } getServiceMapLocked(r.userId).ensureNotStartingBackgroundLocked(r); mAm.notifyPackageUse(r.serviceInfo.packageName, PackageManager.NOTIFY_PACKAGE_USE_FOREGROUND_SERVICE); } else { if (r.isForeground) { final ServiceMap smap = getServiceMapLocked(r.userId); if (smap != null) { decActiveForegroundAppLocked(smap, r); } r.isForeground = false; if (r.app != null) { mAm.updateLruProcessLocked(r.app, false, null); updateServiceForegroundLocked(r.app, true); } } if ((flags & Service.STOP_FOREGROUND_REMOVE) != 0) { cancelForegroundNotificationLocked(r); r.foregroundId = 0; r.foregroundNoti = null; } else if (r.appInfo.targetSdkVersion >= Build.VERSION_CODES.LOLLIPOP) { r.stripForegroundServiceFlagFromNotification(); if ((flags & Service.STOP_FOREGROUND_DETACH) != 0) { r.foregroundId = 0; r.foregroundNoti = null; } } } }
anr的时限为嘛是5s呢？
void scheduleServiceForegroundTransitionTimeoutLocked(ServiceRecord r) { if (r.app.executingServices.size() == 0 || r.app.thread == null) { return; } Message msg = mAm.mHandler.obtainMessage( ActivityManagerService.SERVICE_FOREGROUND_TIMEOUT_MSG); msg.obj = r; r.fgWaiting = true; mAm.mHandler.sendMessageDelayed(msg, SERVICE_START_FOREGROUND_TIMEOUT); } // How long the startForegroundService() grace period is to get around to // calling startForeground() before we ANR + stop it. static final int SERVICE_START_FOREGROUND_TIMEOUT = 5*1000;
这种timeout流程就很熟悉了。
修改方案
有上面可知，问题原因主要是后台启动了服务，在这部分Android O/GO做了限制，
根据章节2.4的过滤条件可以提供如下修改方案：
1) 提升应用优先级到常驻内存级别 (不建议应用采纳这种方式，会导致手机出现很多性能问题)
=> 在AndroidManifest.xml添加android:persistent=”true”
并且签上系统签名
2) 类似与蓝牙BLUETOOTH_UID一样放在白名单里面(需要拥有源码修改权限，而且修改了源码，不利于apk的版本兼容，不建议采纳)
3) 添加在电源相关的DeviceIdle白名单(不建议添加，可能导致功耗增加)
按照上面的都说是不建议采取，是否没有办法了呢？
我们继续往源头找找看看是否有办法：
在ContextImpl的startServiceCommon、ActiveServices的startServiceLocked有一个参数requireForeground/fgRequired，是否前台请求，如果requireForeground/fgRequired为false才会进行后台请求判断，如果是true的话，是可以直接绕过去的
回到ActiveServices的startServiceLocked=>
ComponentName startServiceLocked(IApplicationThread caller, Intent service, String resolvedType, int callingPid, int callingUid, boolean fgRequired, String callingPackage, final int userId) throws TransactionTooLargeException { //... // fgRequired是true可以直接绕过 if (!r.startRequested && !fgRequired) { //.. } //... }
那么方案4，我们可以采取如下方式
4) 通过Context(activity、service的this都是包含context的，故不用担心调用方式)，将之前的startService，修改成ContextImpl的startForegroundService或者startForegroundServiceAsUser方法，启动一个前台服务。
//frameworks/base/core/java/android/app/ContextImpl.java
@Override public ComponentName startForegroundService(Intent service) { warnIfCallingFromSystemProcess(); return startServiceCommon(service, true, mUser); } @Override public ComponentName startForegroundServiceAsUser(Intent service, UserHandle user) { return startServiceCommon(service, true, user); }
ps：注意上面的方法是启动前台服务，你的服务需要是前台的，这个怎么做呢，下面提供2种方法：
1) 在service中调用startForeground (最常见方法)
2) 设置service为前台，可以使用AMS的setProcessImportant设置优先级别 (优点是：不会在通知栏中出现通知图标。缺点是：需要相应的权限)
