1. 需求分析与架构设计
核心需求：
在系统级实现应用列表过滤功能，满足以下要求：
- 动态维护过滤规则（支持黑白名单）
- 多用户隔离策略
- 高性能过滤（<5ms延迟）
- 兼容主流Android版本（8.0+）
架构设计：
plaintext
+----------------+
| SystemUI |
+----------------+
|
↓
+----------------+
| PackageManager|
+----------------+
|
↓
+----------------+
| PMS Extension |
+----------------+
|
+------------+------------+
↓ ↓
+-------------------+ +-------------------+
| Filter Config | | Filter Engine |
| (XML/DB/Settings) | | (Real-time Filter)|
+-------------------+ +-------------------+
2. 核心功能实现
2.1 过滤规则管理服务
新增系统服务AppFilterManagerService：
java
public class AppFilterManagerService extends SystemService {
private static final String TAG = "AppFilterManager";
// 过滤规则存储路径
private static final String FILTER_XML = "/data/system/app_filter.xml";
// 内存缓存规则
private final ArrayMap<Integer, AppFilterPolicy> mUserPolicies = new ArrayMap<>();
@Override
public void onStart() {
publishBinderService(Context.APP_FILTER_SERVICE, new BinderService());
}
// 按用户加载过滤规则
private void loadPolicyForUser(int userId) {
AppFilterPolicy policy = new AppFilterPolicy();
File configFile = new File(FILTER_XML);
if (configFile.exists()) {
// XML解析实现
try (FileInputStream fis = new FileInputStream(configFile)) {
XmlPullParser parser = Xml.newPullParser();
parser.setInput(fis, StandardCharsets.UTF_8.name());
// ...解析逻辑
}
}
mUserPolicies.put(userId, policy);
}
// Binder接口实现
private final class BinderService extends IAppFilterManager.Stub {
@Override
public void addFilterRule(int userId, AppFilterRule rule) {
// 权限校验
enforceManageAppFilterPermission();
synchronized (mUserPolicies) {
AppFilterPolicy policy = mUserPolicies.get(userId);
policy.addRule(rule);
persistPolicy(userId);
}
}
@Override
public boolean shouldFilterApp(int userId, String packageName) {
synchronized (mUserPolicies) {
AppFilterPolicy policy = mUserPolicies.get(userId);
return policy != null && policy.matches(packageName);
}
}
}
}
2.2 PMS查询逻辑增强
修改PackageManagerService.queryIntentActivitiesInternal：
java
List<ResolveInfo> queryIntentActivitiesInternal(...) {
List<ResolveInfo> result = /* 原始查询逻辑 */;
// 新增过滤逻辑
IAppFilterManager filterManager = IAppFilterManager.Stub.asInterface(
ServiceManager.getService(Context.APP_FILTER_SERVICE));
Iterator<ResolveInfo> iterator = result.iterator();
while (iterator.hasNext()) {
ResolveInfo info = iterator.next();
if (filterManager.shouldFilterApp(userId, info.activityInfo.packageName)) {
iterator.remove();
}
}
return result;
}
3. 过滤规则配置方案
3.1 XML规则格式
xml
<filter-policy user="0">
<blacklist>
<item>com.example.hiddenapp1</item>
<item>com.example.hiddenapp2</item>
</blacklist>
<whitelist>
<item>com.example.essentialapp</item>
</whitelist>
</filter-policy>
3.2 动态配置接口
通过cmd命令实现实时更新：
bash
adb shell cmd app_filter add --user 0 --pkg com.example.hiddenapp
adb shell cmd app_filter remove --user 0 --pkg com.example.hiddenapp
3.3 多用户支持
java
public boolean shouldFilterApp(int userId, String packageName) {
// 获取当前用户的父用户策略（适用于Work Profile）
int parentUserId = getParentUserId(userId);
AppFilterPolicy parentPolicy = mUserPolicies.get(parentUserId);
// 合并策略逻辑
return userPolicy.matches(pkg) || (parentPolicy != null && parentPolicy.matches(pkg));
}
4. 性能优化措施
4.1 内存缓存优化
使用LRU缓存策略：
java
private static final int MAX_CACHED_USERS = 5;
private LruCache<Integer, AppFilterPolicy> mPolicyCache = new LruCache<>(MAX_CACHED_USERS) {
protected void entryRemoved(boolean evicted, Integer key,
AppFilterPolicy oldValue, AppFilterPolicy newValue) {
persistPolicy(key); // 策略淘汰时自动持久化
}
};
4.2 快速匹配算法
使用Trie树进行包名匹配：
java
class AppFilterPolicy {
private TrieNode blacklistTrie;
private Set<String> whitelist;
boolean matches(String packageName) {
// 白名单优先
if (whitelist.contains(packageName)) return false;
// Trie树匹配黑名单
return blacklistTrie.contains(packageName);
}
}
5. 安全与权限控制
5.1 权限声明
新增系统权限：
xml
<permission android:name="android.permission.MANAGE_APP_FILTER"
android:protectionLevel="signature|privileged"/>
5.2 Binder调用校验
java
private void enforceManageAppFilterPermission() {
if (Binder.getCallingUid() != Process.SYSTEM_UID) {
throw new SecurityException("Requires MANAGE_APP_FILTER permission");
}
}
6. 兼容性处理
6.1 版本适配方案
通过@SystemApi控制接口可见性：
java
@SystemApi(client = SystemApi.Client.MODULE_LIBRARIES)
public interface IAppFilterManager {
void addFilterRule(int userId, AppFilterRule rule);
boolean shouldFilterApp(int userId, String packageName);
}
6.2 降级策略
当过滤服务不可用时自动跳过：
java
try {
if (filterManager != null && filterManager.shouldFilterApp(...)) {
iterator.remove();
}
} catch (RemoteException e) {
Slog.w(TAG, "AppFilter service unavailable", e);
}
7. 验证与测试
7.1 单元测试用例
java
@RunWith(AndroidJUnit4.class)
public class AppFilterTest {
@Test
public void testBasicFiltering() {
// 添加过滤规则
mFilterManager.addFilterRule(0, new AppFilterRule("com.example.hidden"));
// 查询应用列表
List<ResolveInfo> apps = mPms.queryIntentActivities(...);
// 验证过滤结果
assertFalse(containsPackage(apps, "com.example.hidden"));
}
}
7.2 性能基准测试
使用Jetpack Macrobenchmark：
kotlin
@RunWith(AndroidJUnit4::class)
class FilterPerformanceTest {
@get:Rule
val rule = MacrobenchmarkRule()
@Test
fun queryAppsWithFilter() {
rule.measureRepeated(
packageName = "com.android.systemui",
metrics = listOf(StartupTimingMetric()),
iterations = 10,
setup = { addFilterRules() }
) {
startActivityAndWait(Intent(Intent.ACTION_MAIN).apply {
addCategory(Intent.CATEGORY_HOME)
})
}
}
}
通过以上系统级方案，可在不修改应用层代码的前提下，实现灵活高效的应用列表过滤功能。建议采用动态配置与内存缓存相结合的方式，在保证性能的同时满足企业级设备管理的多样化需求。
