快捷導(dǎo)航

Android StrictMode運行流程(推薦)

更新時間：2018年01月10日 10:04:42 作者：fighting_goat

strictmode是android在 API9后引入的檢測影響app運行流暢性的一種機制。這篇文章給大家介紹了android strictmode運行流程，需要的朋友參考下吧

什么是 StrictMode(嚴(yán)苛模式)

strictmode是android在 API9后引入的檢測影響app運行流暢性的一種機制,例如我們都知道的主線程中不允許有網(wǎng)絡(luò)操作這條規(guī)則就是嚴(yán)苛模式規(guī)則的一種.

strictmode.java 這個類中設(shè)定了許多detect標(biāo)志位例如 DETECT_NETWORK ，還有許多 penalty標(biāo)志位例如 PENALTY_NETWORK ， DETECT標(biāo)志位決定strictmode是否要對這項內(nèi)容進行檢測，PENALTY標(biāo)志位決定了在這項內(nèi)容發(fā)生時是否要拋出異常（相當(dāng)于一種懲罰機制，PENALTY的意思就是懲罰）.

StrictMode 類的作用之一就是對這些標(biāo)志位進行管理，通過 setThreadPolicy() 方法可以設(shè)定 Policy 變量中的mask值.

之后會將 POLICY 變量傳入 BlockGuard 中，BlockGuard 運行在 Dalvik虛擬機中，對所有的異常操作進行統(tǒng)一的管理.

Android官方文檔中對于strict mode 給出的解釋

strictmode 是一種開發(fā)工具，引入它可以使你發(fā)現(xiàn)在開發(fā)過程中產(chǎn)生的問題，并修復(fù)它們.

在 application main thread中常有UI相關(guān)的操作和動畫發(fā)生，strictmode可以在主線程中檢測硬盤和網(wǎng)絡(luò)相關(guān)的操作.將硬盤讀寫操作和網(wǎng)絡(luò)相關(guān)操作挪出主線程可以使你的app更加流暢和具有響應(yīng)性.同時為了使app更加響應(yīng)性，你可以屏蔽ANR發(fā)生時彈出的dialog.

需要注意的是，盡管android設(shè)備的硬盤類型大多為 flash memory,建立在這種存儲介質(zhì)上的文件系統(tǒng)的并發(fā)性仍然是非常有限的(速度上肯定是RAM比較快).

大部分情況下，硬盤的讀寫操作都是非常快的，但在某些情況下，后臺進程中會運行耗費很大的I/O操作，在這種情況下，app的響應(yīng)速度會下降很多.

一.setThreadPolicy()流程

StrictMode類的文檔中給出的strictmode啟動方式

 * public void onCreate() {
 * if (DEVELOPER_MODE) {
 * StrictMode.setThreadPolicy(new {@link ThreadPolicy.Builder StrictMode.ThreadPolicy.Builder}()
 *  .detectDiskReads()
 *  .detectDiskWrites()
 *  .detectNetwork() // or .detectAll() for all detectable problems
 *  .penaltyLog()
 *  .build());
 * StrictMode.setVmPolicy(new {@link VmPolicy.Builder StrictMode.VmPolicy.Builder}()
 *  .detectLeakedSqlLiteObjects()
 *  .detectLeakedClosableObjects()
 *  .penaltyLog()
 *  .penaltyDeath()
 *  .build());
 * }
 * super.onCreate();

1.

在執(zhí)行了 setThreadPolicy()函數(shù)后會調(diào)用 setThreadPolicyMask()方法.

 public static void setThreadPolicy(final ThreadPolicy policy) {
 setThreadPolicyMask(policy.mask);
 }

2.

在 setThreadPolicyMask()方法中,除了在java層的threadLocal中設(shè)置外,還需要在Native層也進行一個設(shè)置.

private static void setThreadPolicyMask(final int policyMask) {
 // In addition to the Java-level thread-local in Dalvik's
 // BlockGuard, we also need to keep a native thread-local in
 // Binder in order to propagate the value across Binder calls,
 // even across native-only processes. The two are kept in
 // sync via the callback to onStrictModePolicyChange, below.
 setBlockGuardPolicy(policyMask);
 // And set the Android native version...
 Binder.setThreadStrictModePolicy(policyMask);
 }

3.

首先分析java層的 setBlockGuardPolicy()方法.

如果policyMask==0,會返回一個默認(rèn)policy,默認(rèn)policy不進行任何設(shè)置和檢測,policy對象存儲在threadLocal變量中(每個線程保存一個policy的對象),首次運行該方法會生成一個默認(rèn)policy(mMask=0)保存在threadLocal中,這里的policy對象是AndroidBlockGuardPolicy類型.

// Sets the policy in Dalvik/libcore (BlockGuard)
 private static void setBlockGuardPolicy(final int policyMask) {
 if (policyMask == 0) {
 BlockGuard.setThreadPolicy(BlockGuard.LAX_POLICY);
 return;
 }
 final BlockGuard.Policy policy = BlockGuard.getThreadPolicy();
 final AndroidBlockGuardPolicy androidPolicy;
 if (policy instanceof AndroidBlockGuardPolicy) {
 androidPolicy = (AndroidBlockGuardPolicy) policy;
 } else {
 androidPolicy = threadAndroidPolicy.get();
 BlockGuard.setThreadPolicy(androidPolicy);
 }
 androidPolicy.setPolicyMask(policyMask);
 }

4.

再看Native層的代碼:

設(shè)置了policy

386void IPCThreadState::setStrictModePolicy(int32_t policy)
387{
388 mStrictModePolicy = policy;
389}

二.StrictMode如何檢測問題.

1.

CloseGuard檢測游標(biāo)是否正常關(guān)閉:

當(dāng)使用ContentResolver來查詢數(shù)據(jù)庫的時候,會返回一個CursorWrapperInner類型的Cursor對象.

mCursor = mResolver.query(mUri, null, null, null, null);

CloseGuard對CursorWrapperInner是否正常關(guān)閉的檢測的邏輯在finalize()函數(shù)中,finalize()會在gc執(zhí)行垃圾回收的時候被調(diào)用(垃圾回收使用了GcRoot算法)

如果沒有執(zhí)行CursorWrapperInner的close()函數(shù)，僅將CursorWrapperInner對象置為null,當(dāng)主動觸發(fā)gc的時候( Systemgc()),finalize()函數(shù)被調(diào)用 ,"Cursor finalized without prior close()"這段log被打印.但如果沒有將CursorWrapperInner對象置為null,這時主動觸發(fā)gc并不會引起 finalize()函數(shù)的執(zhí)行,因為CursorWrapperInner對象被強引用,垃圾回收器在回收時不會考慮回收強引用對象，即使最后內(nèi)存不足而崩潰.

經(jīng)過測試程序的測試，發(fā)現(xiàn)"Cursor finalized without prior close()"這段log在 CursorWrapperInner對象置空并執(zhí)行 System.gc()后是會打印出來的.

但是 CloseGuard中的 warnIfOpen()函數(shù)始終沒有執(zhí)行

在 CursorWrapperInner的構(gòu)造函數(shù)中,mCloseGuard執(zhí)行 open()函數(shù),在 open函數(shù)中allocationSite被賦值,而 ENABLED 變量是默認(rèn)為true的，唯一改變它的setEnabled()方法在源碼中也并沒有被調(diào)用,所以應(yīng)該是會在REPORTER中打印SystemLog的,但最后SystemLog并沒有打印,具體原因分析不出來.

@Override
 protected void finalize() throws Throwable {
 try {
 if (mCloseGuard != null) {
  mCloseGuard.warnIfOpen();
 }
 if (!mProviderReleased && mContentProvider != null) {
  // Even though we are using CloseGuard, log this anyway so that
  // application developers always see the message in the log.
  Log.w(TAG, "Cursor finalized without prior close()");
  ContentResolver.this.releaseProvider(mContentProvider);
 }
 } finally {
 super.finalize();
 }
 }
 }
 public void warnIfOpen() {
 if (allocationSite == null || !ENABLED) {
 return;
 }
 String message = 
 ("A resource was acquired at attached stack trace but never released. "
  + "See java.io.Closeable for information on avoiding resource leaks.");
 REPORTER.report(message, allocationSite);
 }
 @Override public void report (String message, Throwable allocationSite) {
 System.logW(message, allocationSite);
 }
 CursorWrapperInner(Cursor cursor, IContentProvider icp) {
 super(cursor);
 mContentProvider = icp;
 mCloseGuard.open("close");
 }

2.

onSqliteObjectsLeaked()也是用來檢測數(shù)據(jù)庫游標(biāo)有沒有正常關(guān)閉,但這個函數(shù)檢測的是通過SqliteDataBase. query()得到的SqliteCursor游標(biāo)對象.

檢測位置也是在 finalize()函數(shù)中.

 /**
 * Release the native resources, if they haven't been released yet.
 */
 @Override
 protected void finalize() {
 try {
 // if the cursor hasn't been closed yet, close it first
 if (mWindow != null) {
 if (mStackTrace != null) {
  String sql = mQuery.getSql();
  int len = sql.length();
  StrictMode.onSqliteObjectLeaked(
  "Finalizing a Cursor that has not been deactivated or closed. " +
  "database = " + mQuery.getDatabase().getLabel() +
  ", table = " + mEditTable +
  ", query = " + sql.substring(0, (len > 1000) ? 1000 : len),
  mStackTrace);
 }
 close();
 }
 } finally {
 super.finalize();
 }
 }

流程如下圖