Initially the current user in the accessibility manager service is the
owner. This is correct since the system should be able to respond to
queries immediately and their result depends on the current user. However,
the system is calling the user switch callback with the current user
which is the same as the one we initialized with. Switching the user
causes clearing state for the old user winch is in case the current
one. Hence, we are losing state for the current user. This behavior was
masked from the fact that accidentally no events in the system were
fired before the first use user switch call.
repo Losing current user state puts the manager service in an inconsistent
state and it binds to accessibility services more than once. As a result
the accessibility layer starts to misbehave rendering the device useless
to a blind user.
Now we are ignoring user switch callbacks if the new user is the same
as the current one. Since we can no longer initialize at the first user
switch, this change adds explicit system ready method called from
the system server at the right moment.
bug:9496697
Change-Id: Icb39e929ea44e6c0360aba7ddc12f941ca2c9f98
This was a regression caused by a recent change to use String8
instead of char*. We missed the implicit null check here.
Bug: 9377604
Change-Id: I7eff138096622c47b5d45678010373dc82138384
The specific bug is this: SIM PIN unlock attempt toasts are
sent from com.android.settings/.IccLockSettings which runs
as the phone process; NoMan wasn't having any of that and
was blocking the toasts.
With this change we treat SYSTEM_UID and PHONE_UID the same
for all security checks, and furthermore we guarantee that
all notifications and toasts from those UIDs will be
permitted.
Bug: 9098802
Change-Id: Idc788527aa2cb38e015fe92773766a514167999e
When building commands to send across NativeDaemonConnector, scrub
sensitive arguments to prevent them from being logged.
Bug: 8609800
Change-Id: I84b16791749264a010f7e59f9918f68d71bac6b9
There were some paths in LocationManagerService where
mRecivers was being accessed/modified without the lock held.
Update method names to indicate they need to be called with
lock held to make it more clear in the future when such a
problem may happen.
Change-Id: Ie2a9d019155ac7cedd1db298caca75b8fe382ca7
Otherwise there's nothing to kick us into scanable modes unless
the user toggles wifi.
bug:9217455
Change-Id: I6460305e3f299646433546598412f817579cc805
Telehony seems to sometimes be reporting the country before boot
is completed so can't persist the data at that time. Remember
and write it on BOOT_COMPLETED
Also, there are permission issues around writing the setting.
bug:9225156
Change-Id: Ifdf2243da71b0d2ce5743267842597937d790ef5
This commit splits LocationManagerService's monolithic WakeLock into
per-LocationManagerService.Receiver WakeLocks, for better WorkSource
accounting. This should make it easier to debug location-related
power issues.
Change-Id: I0d2897c305a38099f9663dc1bc9354ce4bbe1077
(cherry picked from commit 0aa28602d51bf41e46d18ffefe724ebc3ff7a704)
Fix for the synchronization issue leading to access of an array
Index out of bounds. Issue occurs due to race condition between
removing the activities of a crashed process from history stack
and resuming a separate activity.
Change-Id: I14bb5834e778c15b674248e46fe93b0ce9f37967
The new one is often null when disconnected, so using the new fails.
In other situations, it's the connections on the old network we want
to reset anyway, so the old code when it would work would also do
the wrong thing (unless new iface == old iface).
bug:9112928
Change-Id: I1fcae89cc3aa9d712e516e7c97cece0b89869bd9
Hacks to enable gms to see accounts were interfering with checking for
explicit package name.
Bug: 9095427
Change-Id: Iae65ea2a678d7403d023fa1a61c01347b77c3b5a
This is WRONG WRONG WRONG but certain apps rely on it to
poke the LED and so forth. In a future release this will
stop working.
Bug: 8623399
Change-Id: I49bb8ccc6891b1398ceec94c64d6c3a510ad1c38
With the new scan-always mode we will get scan results when disabled.
Don't tell the user about open AP when they've turned off wifi.
bug:9067982
Change-Id: I7a303c2a6f87a331443eeb54fa8ff5b467c0a1ff
We don't want to save all the heavyweight stuff, but we
can't lighten the notification in place because there may be
other active references.
Bug: 8839246
Change-Id: Id6dcc1cf5ae196ee3e42908d6b3ec4e928e9f0c5
If deferred action msg was delivered promptly it would end up
looking like it too needed to be deferred and then nothing
would end up getting done.
bug:9014777
Change-Id: I243df70463ebb9333efd6df510f71f9d8d9bdf91
PACKAGE_CHANGED is called when either the application or any
of its components change. Rework old early exit code to
handle component-level disabling wrt listeners.
Note PACKAGE_CHANGED can take quite a while to reach NMS.
Up to ten seconds (!) after component-disabling in app.
Bug:8982243
Change-Id: Ia325516140dd9288466a663192a5326019f071ab
For UI test automation purposes we register a fake accessibility service
and suspend all other services. When the UI automation serivce is unregistered
we restore the suspended ones. Since the UI automation serivce is fake and
incomplete, for example it has not resolve info, it should not be reported
to clients as being installed or enabled.
bug:8871034
Change-Id: I66792cd028159c1652d3c8a2982164821282ab24
We sometimes have issues if disable, enable seq happens to fast.
This should only slow down those fast cases.
bug:8715336
Change-Id: I4f7fe9708b7c3c2300c441511838e8a70eaad5d1
When an upgraded system-bundled app is reverted to the factory
version, we now maintain restricted profiles' access policies
w.r.t. that app.
Bug 8905395
Change-Id: I3f230ac3f70a77fb14c0180b44c9b42fc49250a2
implement error recovery when bluetooth stack return error with error
transition state from TURNING_ON to OFF due to ENABLE_TIMEOUT.
issue 7566317
Change-Id: I709b475bbfbd90e4f6e423f9eedc5340c89b2cb0
Loading the avatar icons and drawing them into the sized bitmap
turns out to be quite expensive and the cost increases with number
of users. Caching them shaves off several hundred milliseconds from
Keyguard inflation time during user switching on the lockscreen.
For instance, 15ms vs. 750ms with 3 avatars on a certain 7" tablet.
Bug: 7986933
Change-Id: I3e2065bfa25aa263133ba204ca364c3b04d7c0ff
