in the frameworks/base repo.
Author: Rubin Xu <
rub...@google.com>
Date: Tue Nov 22 15:18:32 2016 +0000
Fix boot loop when upgrading direclty from L to N
A second attempt to fix the upgrade problem due to SID == 0
in the above upgrade path. The previous fix contains a bug
where it would cause future attempts to unify work challenge
to silently fail, and crash SystemUi when unlocking.
This fix adds a check for non-zero SID before doing the initial work
profile unification (which caused the upgrade crash when SID == 0).
This means the initial work profile unification would only happen when
the user has unlocked the lockscreen and SID is generated.
Bug: 32490092
Bug: 33050562
Change-Id: Ib28951b2ec26b4f091df7763d9902f55616fcb5c
(cherry picked from commit bfc7faaf353ea75ab04e986edbc79478679d40f6)
It seems this commit was reverted quietly:
$ git diff ad760e11 android-7.1.1_r6
diff --git a/services/core/java/com/android/server/LockSettingsService.java
b/services/core/java/com/android/server/LockSettingsService.java
index 03744f8..a91e205 100644
--- a/services/core/java/com/android/server/LockSettingsService.java
+++ b/services/core/java/com/android/server/LockSettingsService.java
@@ -240,17 +240,6 @@ public class LockSettingsService extends
ILockSettings.Stub {
if (DEBUG) Slog.v(TAG, "Parent does not have a screen lock");
return;
}
- // Do not tie when the parent has no SID (but does have a screen lock).
- // This can only happen during an upgrade path where SID is yet to be
- // generated when the user unlocks for the first time.
- try {
- if (getGateKeeperService().getSecureUserId(parentId) == 0) {
- return;
- }
- } catch (RemoteException e) {
- Slog.e(TAG, "Failed to talk to GateKeeper service", e);
- return;
- }
if (DEBUG) Slog.v(TAG, "Tie managed profile to parent now!");
byte[] randomLockSeed = new byte[] {};
try {
I'm curious why.
If this commit should be reverted, there should be
a commit which explicitly reverted it. Shouldn't it?
Or is this commit was incorrectly reverted
due to wrong merge reset?