Since ContentProvider file operations can end up doing substantial
network I/O before returning the file, allow clients to cancel their
file requests with CancellationSignal.
Ideally this would only be needed for openFile(), but ContentResolver
heavily relies on openAssetFile() and openTypedAssetFile() for common
cases.
Also improve documentation to mention reliable ParcelFileDescriptors
and encourage developers to move away from "rw" combination modes,
since they restrict provider flexibility. Mention more about places
where pipes or socket pairs could be returned.
Improve DocumentsContract documentation.
Bug: 10329944
Change-Id: I49b2825ea433eb051624c4da3b77612fe3ffc99c
When Activity.convert{To|From}Translucent() is called the
ViewRootImpl is now notified when it is safe to convert the Canvas
from translucent to opaque and back to translucent. This will make
it possible to save resources when compositing opaque layers.
Fixes bug 10349536.
Change-Id: I7282aee1d54601fb00611d20be204bf164d873f6
Having a single mHomeProcess in ActivityManagerService ignores the
possibility of multiple processes serving as home. Particularly when
we have a home activity that only serves to launch the true home
activity.
Fixes bug b10296766
Change-Id: I9e5c9c955c64f4b3df856b6bc66d6a0d93f85aab
The warning about an ActivityRecord not being assigned a task was
incorrect in the location modified by this fix. In this case an
existing ActivityRecord is relaunched so it is not necessary to
assign the task to the passed ActivityRecord.
Addresses the warning associated with bug 10181389.
Change-Id: I76d5066c320bf9da2663bc34bcaca801ad4953bc
Now that granted Uri permissions can be persisted across reboots,
offer APIs to discover them. Returns incoming or outgoing grants
matching the requested flags and mask. Add helper method to discover
"open" documents using this new API and filtering for non-documents.
Require that callers own at least of the filtering packages to avoid
exposing all grants. Switch internal grant tracking to use ArrayMap.
Change-Id: I0a755f221d0d160b411f8d3cfc48279b64345733
The new wallpaper positioning associated with multiple activity stacks
put animation backgrounds in front of the wallpaper. This caused
hideous jank. Testing for visible wallpaper and moving the background
behind it fixes the jank.
Fixes bug 10078282.
Possibly fixes bug 10247094.
Change-Id: I3f4e07accd7276d59725192081904b791f77781f
For process state, if a top process is making use of another
process, we should probably count the second process as top
as well (instead of IMPORTANT_FOREGROUND). Specially, when
chrome binds to render processes we want those render
processes to be TOP. Otherwise, they end up in the
important foreground state and it looks like they are running
for some other reason in the background.
Change-Id: Id115dbb65dc2b403ffa4bbe4e7837564eb3b9cb5
Also fix a bug where, when parceling the stats, we were
computing the final duration values too late. We need to
do that before we write the long table.
Change-Id: Idb6c1ed95417448c56973fe5866bfb3570e525f4
Make sure screen layout changes are taken into account in Configuration diffs.
Initialize the SystemProperty from Global settings on startup of
ActivityManagerService.
TextUtils checks the override flag to decide if the default layout direction
should be forced to RTL.
Bug: 10244047
Change-Id: I23a2583d790a355060d0d898ba44e5f7dc896b46
PAC (Proxy auto-config) files contain a single javascript function,
FindProxyForURL(url, host). It gets called to determine what proxy should be
used for a specific request.
This adds PAC support to the system. The ProxyProperties has been modified
to hold the PAC file when one is present. The Proxy method
setHttpProxySystemProperty has been modified to insert a PacProxySelector
as the default ProxySelector when it is required. This new ProxySelector
makes calls to the ConnectivityService to parse the PAC file.
The ConnectivityService and the WifiConfigStore have been modified to support
saving the extra PAC file data.
The ConnectivityService now has a class attached (PacProxyNative) that
interfaces to the native calls for PAC files. The parsing of the PAC file
is handled by libpac (which is being added to external/) which utilizes
libv8 to parse the javascript.
As a fallback to applications that don't use the java ProxySelector, the proxy
is setup to point to a local proxy server that will handle the pac parsing.
bug:10182711
Change-Id: I5eb8df893c632fd3e1b732385cb7720ad646f401
In order to determine if the topmost activity needs to be moved out
of the way, compare its task to the task of the source activity
rather than compare the activities themeselves.
Fixes bug 10182783.
Change-Id: Idd4542d7ddce8b7586ee47f100366bc47e07d54f
Add Service.onProvideAssistData(Bundle) which will be
called on foreground Services that have the new attr
in their manifest of provideAssistData = true;
Rename private reference to e.g. "getTopActivityExtras"
as "getAssistContextExtras" - do not rename the relevant
permission, since it is already public.
In ActivityManagerService, request extras both from the
top activity and from any foreground services with the
above attribute. Extend PendingActivityExtras as
PendingAssistExtras with a list of Services from which
extras are expected.
Reduce the timeout to or reporting extras from 4 sec to
just 500 ms.
Bug: 9526331
Change-Id: Ia03b96e8189033a68ae9c514c8cea0199a19bce8
ProcessStats is now called ProcessCpuTracker.
ProcessTracker is now ProcessStatsService, and its inner State
class is broken out into a separate top-level ProcessStats class.
This ProcessStats is moved to the framework, so we will be able
to use it elsewhere.
Change-Id: I6a127bcb835b6b474b72647c0b99b82c2137e5c5
Add more locations where mLastPausedActivity must be set null before
calling resumeTopActivityLocked().
Also separate the "No History" activity from mLastPausedActivity by
adding mLastNoHistoryActivity. This is necessary because now
mLastPausedActivity will be null when mLastNoHistoryActivity
is checked.
Continues to fix bug 10022212.
Change-Id: Ida543b0638c865127691dc51a6e0635062539659
Moving an activity to the back used to move it behind the launcher.
This meant that the test in resumeTopActivityLocked() for
mLastPausedActivity being equal to the top activity would fail after
moveTaskToBack(). In the new activity stack model the launcher is
on a different stack so if there is only one task then the test
now passes. Clearing mLastPausedActivity when moving an activity
forward now passes the test.
Fixes bug 10022212.
Change-Id: I17932d5c87d2483f13c29ae85a7067202a08f34e
This is a process that is not actually running, but would be
if we had enough RAM.
Also rework how service stats are stored to used the nice
compact structure we are already using for processes.
And fix a bug where when we committed the current state data,
we would mistakenly write it under the name of the *next* state
that is now starting. Ouch.
Change-Id: I9f2b84b8b3f305301fd48dcd9d4e6c232abe8ef9
The goal is to collect PSS data from processes when they
are in as stable a state as possible. We now have tables
for the durations to use for PSS collection based on
process state, and scheduling of collection is all driven
by process state changes.
Change-Id: I95e076f6971b9c093118e53dc8240fb3f5bad3b2
Fix a problem where we would not respect the device's
override display density/size if those had been reset at
some point (to an empty string).
Tweak the tuning of low RAM devices to consider 480x800 as
the base line resolution. That is our target these days.
Improve output of oom and meminfo to include this tuning
information.
Change-Id: Ic9a85b0391b075178c4ac17e64086ef21889f430
In preperation for DEFAULT_TO_EXPERIMENTAL_WEBVIEW flag-flip, make this
field and WEBVIEW_EXPERIMENTAL_PROPERTY private so we know all users are
querying the property in a consistent way.
Change-Id: Iabaffb8ac5a34e323b170339d47f872eb4c83042
If a service is being executed due to a call from a background
process, there is no reason for its process to be brought up
to the foreground scheduling group to do its work. So, don't.
Change-Id: I0f4f9f2fdcc3481ca4ad2e4e3deb43b8edbbd897
Set the sys.sysctl.extra_free_kbytes property, which will proxy
through init to the /proc/sys/vm/extra_free_kbytes tunable (on
kernels where it exists). This will ask the kernel to keep
more memory free, causing lowmemorykiller to fire earlier and
allocations to succeed faster.
The size is set to 3 full screen 32bpp buffers. Products can
override the size with an absolute or adjustment value using
overlays config_extraFreeKbytesAbsolute and
config_extraFreeKbytesAdjust.
Bug: 10024467
Change-Id: Ib4d4507513ec3c1f4d4ceeb81ed632d1ad643437
We now include in the app ops information about the device runtime
config that is relevant. Currently this is the dalvik runtime in
use and the webview. In checkin, this data looks like:
config,libdvm.so,webview
Change-Id: I85fd53418bd43595468f23ec8619e772fdaee7e1
Add overlays config_lowMemoryKillerMinFreeKbytesAbsolute and
config_lowMemoryKillerMinFreeKbytesAdjust to allow a product to
modify the size of the lowmemorykiller minfree buckts. The absolute
overlay sets the size of the largest bucket to the specified value
and scales the smaller buckets proportionally. The adjust overlay
(which can be negative) directly adds to the size of the largest
bucket and to the the smaller buckets proportionally.
Change-Id: I0d6b7662be12fd151deb2bf9591f2c7a8b1cb6f7
We were now propagating the screen on state when updating
all process states, so they would get pushed into the screen
off bucket always even if the screen was on. Oops!
Also improve the summary output when dumping a single package
to be more summary-like.
Change-Id: I16c640f9dc02d6db8c66aeb1c720f67beab60635
When safely resetting stats after committing them, we were
mistakenly clearing the current memory state so we would lose
that total memory time until the memory state changes again.
Also improve the summary output to print percentages, which
make more sense for that display.
Change-Id: I0fe45fd78e97ec8b94976170dd42f4ed345a5899
Activities from the home package were causing tasks types to change from
application to home. This was not the intention of setting the task type
when adding an activity. This change sets the task type to the inherent
type of the first activity added. All subsequent activities added to the
task then inherent the task's type overriding the inherent type of the
task.
Fixes bug 9972495.
Change-Id: Ib77675aea790ea64d4f166af62c7138e89356c13
