The main problem here was a mistake when turning a single process
structure to a multi-package-process structure with a common
process. When we cloned the original process state, if there were
any services already created for the process for that package, they
would be left with their process pointer still referencing the
original now common process instead of the package-specific process,
allowing the active counts to get bad. Now we switch any of those
processes over to the new package-specific process.
There was also another smaller issue with how ServiceRecord is
associated with a ServiceState -- we could be waiting for an
old ServiceRecord to be destroyed while at the same time creating
a new ServiceRecord for that same service class. These would share
the same ServiceState, so when the old record finally finished
destroying itself it would trample over whatever the new service
is doing.
This is fixed by changing the model to instead of using an "active"
reference count, we have an object identifying the current owner
of the ServiceState. Then when the old ServiceRecord is cleaning
up, we know if it is still the owner at that point.
Also some other small things along the way -- new Log.wtfStack()
method that is convenient, new suite of Slog.wtf methods, fixed
some services to use Slog.wtf when catching exceptions being
returned to the caller so that we actually know about them.
Change-Id: I75674ce38050b6423fd3c6f43d1be172b470741f
- Call in all circumstances but only set launchHomeTaskNext for
focused stack. Previous version didn't call handleAppDiedLocked for
non-focused stack.
- Rearrange logic to run down the top task and make sure that all
remaining activities belong to the dying app. Previous version just
looked at the top non-finishing activity and based its behavior on
that.
Fixes bug 11029560.
Change-Id: Ic3a7c873c4c975577d6b390a8955ff41729bdfde
Should fix the array index out of bounds. valueAt() is when we have the
index, not the key.
Bug: 11014210
Change-Id: Icb53fe763782befbd5de1b3df6630b87cda72e84
If two activities are started at the same time the first activity can
add a starting window but never start. In that case there is no event
that will clear the starting window. This change adds a 10 second
timeout for the starting window to be cleared after which it will
clear the starting window automatically.
Fixes bug 10797865.
Change-Id: I1d59c3058c63367ff688d426474e8a6f006b2e0d
All ContentProvider calls are currently blocking, making it hard for
an app to recover when a remote provider is wedged. This change adds
hidden support to ContentProviderClient to timeout remote calls,
treating them as ANRs. This behavior is disabled by default.
Update DocumentsUI to use a 20 second timeout whenever interacting
with a storage provider.
Bug: 10993301, 10819461, 10852518
Change-Id: I10fa3c425c6a7225fff9cb7a0a07659028230cd3
Fixes jank exposed in 10881705. Specifically background activity
animating up along with translucent activity. Repro steps on manta:
1. From home start Settings.
2. Press home.
3. From home start Downloads (translucent activity that takes 85% of
screen).
4. Observe that as Downloads zooms up the 15% boundary that should be
dimly transparent are showing Settings.
The cause was that there is a finishing activity in the Downloads task
that was used to launch the DownloadsActivity. The existence of that
activity kept the logic from recognizing that the home activity was
behind the DownloadsActivity, not the Settings activity.
This fix descends through all of the activities in a task sitting on
home and makes sure that they only keep home from showing if such
activities are not finishing and visible.
Change-Id: I607afce6b0000b4db634f2ce40a6c37fcee369d7
Not dealing with the case where there is a null list.
Also fixed some bugs I found while looking at this:
- When resetting the stats, we would use a newly computed time stamp
for the total durations rather than the one we used to reset the
proc/service entries. This would result in them being able to be
slightly > 100%.
- There was a bug in how we split a single process state into its
per-package representation, where we would but the cloned process
state into the new package's entry (instead of properly for its
own package entry), to be immediately overwritten by the new
process state we make for that package. This could result in
bad data for processes that have multiple packages.
- There was a bug in resetting service stats, where we wouldn't
update the overall run timestamp, allowing that time to sometimes
be > 100%.
- There was a bug in computing pss data for processes with multiple
packages, where the pss data was not distributed across all of the
activity per-package process states.
- There was a bug in computing the zram information that would cause
it to compute the wrong value, and then never be displayed.
Finally a little code refactoring so that ProcessState and ServiceState
can now share a common implementation for the table of duration values.
Change-Id: I5e0f4e9107829b81f395dad9419c33257b4f8902
Do not use the shortcut of the package name to identify the home
activity.
Fixes bug 10963726.
Fixes bug 10920950.
Change-Id: I725781a26672b055a816994aee6ea458a7f07c88
Killing the GEL search results was killing everything in its package.
This fix keeps the home process from being killed when a task in its
package is killed.
Fixes bug 10927223.
Change-Id: I56e75f0a0118885a1604cbd70320bbdb4f8cf1a2
The home activity was being returned to when any activity in a task
that was launched from home crashed. If there were still activities
left in the task then the crash should have brought up those
activities next, not home.
This may be a partial fix for crashes where the back stack was showing
up under launcher icons. Bug 10858941.
Change-Id: I840a25bd8395bfce46f4e21b112d78b12884706d
If an activity is visible behind the keyguard when it is launched
by another activity then there would be no call to dismissKeyguard.
Because the other activity is pausing the call to dismissKeyguard
is skipped in startActivityLocked(). And because it is already
visible the call to ActivityRecord.windowsVisible() is never made and
the call to reportActivityVisibleLocked() which calls
dismissKeyguard() is also never made.
This change recognizes when an activity is resumed and visible and
calls dismissKeyguard() in that case.
Fixes bug 10732489.
Change-Id: I3de1350a55231aaa14dadc8709fd0fcf4960742c
Back out changes from CLs ag/363992 and ag/363859. These introduced
the bugs found in bug 10917435 which is now fixed. Note that backing
out these changes reintroduces bug 10732489.
Change-Id: Ic5105dd4cfc8bf79c6f06188283d1ee3680c370c
Check and throw if callers request invalid grant flags. Add API to
test if a Uri is backend by a DocumentsProvider.
Bug: 10919391, 10935608
Change-Id: Ifa6afefb95983558c8c64dc15ddf650e9fe07080
Now when memory low, if a service's process is above
a selected pss, then the process is not allowed to go
in to the service a list.
Also simplified the normal meminfo details dump to not
include the shared dirty and shared clean sizes by
default, since these can be very confusing. You will
still get to see them with the "-a" flag.
Finally some small steps to better managing service
processes in the LRU list, so hopefully we can some
day be better about letting them drop down in the list
when there isn't really much interesting happening in
the process. Not yet used at this point.
Change-Id: I654bfd6d05de2a63120185ebb15ffda8cbeb5dac
Change our Intent flag to indicate that a Uri permission grant is
persistable, but don't actually persist it until explicitly taken by
the receiving app. This prevents apps from spamming each other if
persisted permissions aren't really required.
Remember the last time a persisted grant was taken by an app, and
use this to prune away the oldest grants when the number of grants
grows too large. Allow apps to query persisted grants they are
holding, and allow them to release previously persisted grants. Add
public UriPermission class to return grant details and timestamp.
Track various permission strengths separately, and combine together
after each mutation pass. Persistable grants are currently treated
like global grants, but they could be moved to have owners in the
future. Require that grant holders trying to extend a persistable
permission actually hold a persistable permission themselves.
Bug: 10835779
Change-Id: I95b2f797c04ce7fd2612f9a644685dbd44e03759
The previous fix for keeping activities from running on startup,
ag/363992, was keeping the home task from launching when the
keyguard should have allowed it.
This fix permits the home activity to launch in such situations.
Fixes bug 10916877.
Change-Id: I429f0d5a13e06a247b9b6b7241f9a3514044c371
The problem was that the ResolverActivity filters some activities
out of the list it shows, but it uses that display list as the
list of components the preference is set against when ultimately
setting it on the package manager... but that filtered list is *not*
the right component set, since it is not the same as the package
manager's view on it.
The fix here is to retain the original set of matching components
and use that when setting the preferred activity. Note that this
does mean that in very unusual cases where filtering is happeing
(such as one of the activities not being exported but being seen
as a possible completion from another app), then you will be setting
the preference for the complete set. Ultimately we probably need
to have the package manager apply these filtering rules up-front so
this is all consistent, but this is a very rare case so not that
important.
And then most of the change here is just improving the debug
output for intent resolution.
Change-Id: Ie35ac2c05a45946439951bbf41433c8b7de79c05
Following boot the initial activity was automatically resumed even if
a lockscreen is obscuring it. Refer to CL 363859 for why this breaks
things.
This fix pauses all activities the first time a lockscreen appears.
Completes the fix for bug 10732489.
Change-Id: I6fcac14b574c495aa0e16d798cddc1263c6b4c25
...ActivityManagerService.updateLruProcessInternalLocked on bluetooth
Don't try to move process records associated with dead service
connections.
Technically we should probably be clearing the binding/service's
app entry so we don't get into this case, but the least intrusive
change for now is this check.
Change-Id: I6683e692eb5a8fa4f8ec1fa31bd63ec3d7f878ef
...activity chooser from being shown
Add more useful output when intent filter debugging is enabled.
Change-Id: I3722b03ed625046398e81233cf7fb6aa5ded5eca
TLDR: Having a resumed activity behind keyguard can cause the keyguard
not to be dismissed.
Swiping the home button to launch Google Now causes an ASSIST intent
to be launched. The ASSIST intent starts SearchActivity which then
launches GEL. If an activity is resumed behind the keyguard when this
happens then that activity will be paused.
Because that activity is PAUSING, ActivityStackSupervisor
startActivityLocked() doesn't call dismissKeyguard() immediately.
Instead dismissKeyguard will be called later when GEL switches from
not-visible to visible. However, if the paused activity happens to be
GEL then there is never a not-visible to visible transition and
dismissKeyguard never gets called.
This fix removes an unnecessary call to resumeTopActivitiesLocked
which was causing activities behind the lockscreen to be resumed.
This fixes bug 10732489 except immediately after boot. Pausing the
initial activity if the lockscreen is visible after boot is deferred
for another CL.
Change-Id: I323262596ae41bc5a2700bae5942f6a4fba80936
...in ActivityManagerService.updateLruProcessInternalLocked on bluetooth
Add more debug output to help track down what is going on.
Also fix a little problem where, when a service ANRs, if you ask to
wait and it still wasn't responding, the ANR dialog wouldn't be
shown again.
Change-Id: I5be2b1705a0a39ca2992624ae683945c5f38065d
Gah I messed up when refactoring so it would always be told
RAM is low.
Also slightly tune the low memory parameters to go into low
memory states a bit more aggressively.
Change-Id: I5f970349760ad349d515a85c266ab21b387ee353
When transitioning from activity-over-launcher to task-over-launcher
ensureActivitiesVisibleLocked() was too aggressive in showing the
launcher. If there were any non-fullscreen activities in a task that
sits over the launcher then the launcher would be shown.
This fix adds a test to make sure the launcher will only be shown if
the bottommost activity in such a task is non-fullscreen.
Fixes bug 10840919.
Change-Id: I5dcd63be3fa2865ae38cbb921332937dfa4b5d47
...be uncached and too large
When the device is in a low RAM state, when we go to pull a cached
process out to use for some background operation, we can now kill
the current process if we consider its size to be too large.
Note that the current implementation for killing processes is to
just use the same killUnneededProcessLocked() method that we already
have for other things like too many cached processes. This is a
little wrong here, though, because in this case we are at the
point where the caller is actually looking for a process to use.
This current code is not actually removing or cleaning up the
process, so we still need to return the now killed ProcessRecord
and let things fall out from there, which typically means the caller
trying to make an IPC on it and failing and falling into its "oh
no the process died unexpectedly" path. All code using this
*should* be able to handle this correctly, anyway, since processes
really can be killed at any time.
At some point we may to make this implementation cleaner, where it
actually tears down the process right in the call and returns a
null ProcessRecord. That is very dangerous however (we'd need to
go through all paths into this to make sure they are going to be
okay with process state changing on them like that), and I'm not
sure it is really worthwhile. This intention is that killing
processes like this is unusual, due to processes being too large,
and anyway as I wrote all of our incoming code paths must already
be able to handle the process being killed at this point and one
could argue this is just another way to excercise those code paths.
Really, the main negative to this is that we will often have spam
in the log with exceptions about processes dying unexpectedly.
If that is the only issue, we could just add some conditions to
quiet that up at in this case.
We don't want to compute the size of the process each time we try
to evaluate it here (it takes 10s or ms to do so), so there is now
a new field associated with the process to give us the last pss
size we computed for it while it was in the cached state.
To be able to have better cached pss data when we now need it, the
timing for computing process pss has been tuned to use a much
shorter delay for the situations when the process has first switch
into a new state. This may result in us having a fair amount more
pss data overall, which is good, as long as it doesn't cause us to
be computing pss excessively and burning cpu.
Procstats now also has new state to keep track of the number of
times each process has been killed by this new system, along with
the min, avg, max pss of all the times it has happened. This has
slightly changed the checkin format to include this additional data
at the end of pkgkills/prockills lines.
Other changes here:
- Fixed a problem where GPU RAM was not being seen when dumping
the full RAM details of a process. This was because in that
case the system would ask the process to compute its own MemInfo,
which it returned, but the process doesn't have permission to
access the files containing the GPU RAM data. So now the system
always computes the MemInfo and hands it to the app.
- Improved broadcast delays to not apply the delay if the next receiver
of the broadcast is going to run in the same process as the last
one. A situation I was seeing was an application that had two
receivers, one of which started a service; we are better off letting
the second receiver run while the service is running.
- Changed the alarm manager's TIME_TICK broadcast to be a foreground
broadcast. This really should have been anyway (it is supposed to
go out even minute, on the minute, very accurately, for UI elements
to update), and is even more important now that we are doing more
things to delay background broadcasts.
- Reworked how we maintain the LRU process list. It is now divided
into the two parts, the top always containing the processes holding
activities. This better matches the semantics we want (always try
to keep those around modulated by the LRU order we interleave with
other cached processes), and we now know whether a process is being
moved on the LRU list because of an activity operation so we can
only change the order of these activity processes when user operations
happen. Further, this just makes that common code path a lot simpler
and gets rid of all the old complexity that doesn't make sense any
more.
Change-Id: I04933ec3931b96db70b2b6ac109c071698e124eb
At some point during refactoring of ActivityStack, the code to pause the current
activity got deleted. Added back that line of code. Activity will now pause
as soon as the screen is turned off, rather than after 5 seconds (sleep timeout).
Bug: 10632898
Change-Id: If3cc8708d692d29a13dbd8cfd9edda8883b38c2e
