Do not reuse the same ProcessRecord if it's not fully gone yet
When force-stopping apps, it's eligible for restart, and during that we might be reusing the old ProcessRecord before that process is really gone. Now don't reuse that. And also, remove the ProcessRecord from the LRU list when we know it's really gone. Bug: 286299952 Test: Manual - Clear app data and make sure it could restart correctly. Change-Id: I2c71c64b962cb9b4ae15881add50c35fd9e4a7af
This commit is contained in:
@@ -2572,9 +2572,14 @@ public final class ProcessList {
|
||||
// and did the cleanup before the actual death notification. Check the dying processes.
|
||||
predecessor = mDyingProcesses.get(processName, info.uid);
|
||||
if (predecessor != null) {
|
||||
if (app != null) {
|
||||
// The process record could have existed but its pid is set to 0. In this case,
|
||||
// the 'app' and 'predecessor' could end up pointing to the same instance;
|
||||
// so make sure we check this case here.
|
||||
if (app != null && app != predecessor) {
|
||||
app.mPredecessor = predecessor;
|
||||
predecessor.mSuccessor = app;
|
||||
} else {
|
||||
app = null;
|
||||
}
|
||||
Slog.w(TAG_PROCESSES, predecessor.toString() + " is attached to a previous process "
|
||||
+ predecessor.getDyingPid());
|
||||
@@ -5195,6 +5200,8 @@ public final class ProcessList {
|
||||
mDyingProcesses.remove(app.processName, app.uid);
|
||||
app.setDyingPid(0);
|
||||
handlePrecedingAppDiedLocked(app);
|
||||
// Remove from the LRU list if it's still there.
|
||||
removeLruProcessLocked(app);
|
||||
return true;
|
||||
}
|
||||
return false;
|
||||
|
||||
Reference in New Issue
Block a user