When kernel is configured with CONFIG_VMAP_STACK=y kernel stacks are mapped
into process space and therefore already accounted as part of the value
returned by ReadVmallocInfo and reported under MEMINFO_VM_ALLOC_USED
category. Introduce isVmapStack function to check if kernel is configured
with that option and if so do not exclude kernel stack size in the
calculation of memory used by kernel.
Bug: 146088882
Test: dumpsys meminfo with CONFIG_VMAP_STACK enabled and disabled
Change-Id: I9230bbf0ab05524cf52ded35171fd9ec3da16fa0
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Provide means for system_server to query ION memory usage. The new API
allows to get total size of ION heaps, pools and part of ION heaps that
is mapped into user space.
Bug: 138148041
Test: dumpsys meminfo
Change-Id: I6d76d534bbab5ee1646e8e3e3fe5210dcf76bc23
Merged-In: I6d76d534bbab5ee1646e8e3e3fe5210dcf76bc23
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
If device doesn't support userspace reboot then call to
PowerManager.reboot("userspace") will throw an
UnsupportedOperationException.
For the case of upgrading devices, Treble enforces that previous vendor
image should work with the new system image. Since userspace reboot
requires services to be stopped and restarted cleanly, which upgrading
devices might not be able to comply with, we need to have an API to
distinguish between devices that support userspace reboot and the ones
that do not.
Test: atest PowerManagerTest
Test: atest CtsUserspaceRebootHostSideTestCases
Bug: 138605273
Bug: 135984674
Change-Id: I8303f43ab29499eb2995f0256854c787055d9560
Such checks uses old libvintf to parse new VINTF metadata, which is not
supported. The check is now moved to OTA package build time, so there is
no need to check it on the device.
Test: builds
Bug: 139300422
Change-Id: Ifaec06b5298f698958e5727316bf661d1960d0aa
With new KReclaimable stat available in /proc/meminfo include it in
getMemInfo results so that it can be accounted in dumpsys meminfo reports.
Bug: 138148041
Test: dumpsys meminfo
Change-Id: Id475fcc0a63c6fc12ea64e08cec5cd24c83848a5
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Test: Java in process API still work correctly - tested locally with test app + adb overrides.
Test: atest com.android.internal.os
Test: atest com.android.server.am
Bug: 145743810
Change-Id: I6645a7ec767400cd579207f0eed1e6e7136c8a6f
Merged-In: I6645a7ec767400cd579207f0eed1e6e7136c8a6f
As we are elimating the Binder notifications for the sysprop update for
atrace, we no longer have a callback that can be used to read the new
value of the enabled tags.
@CritivalNative calls are very fast (25 ns) so the overhead of always
going to native code to read the tags is negligible.
Test: flash & boot
Test: adb shell su root atrace -t 10 ss
Test: adb shell su root atrace -t 10 wm
Bug: 137366208
This is a cherry-pick of 9fd210224a
Change-Id: I1a07fefd751ee28ca9a632a3d78a2925e8827b9c
Merged-In: I1a07fefd751ee28ca9a632a3d78a2925e8827b9c
This class only exists for UnsupportedAppUsage. These methods and implementations are required for compilation but should never be called.
Bug: 143108344
Test: Manual (compilation and boot)
Change-Id: I27111ae35ca57f7cb80b0439f78edd7f204408e9
This parcelable is used to pass incfs fds from vold to system_server.
Test: builds
Change-Id: I7f0af8798d12815fbae404afabf4a88229a705da
Merged-In: Ifc3a43633356951b239486d66ee80d120bc4ab76
Exposing constants that were originally hidden
Test: build
Bug: 140908357
Change-Id: I12b8d338a58c0ffe294137171b8bac7d80a0b776
Merged-In: I12b8d338a58c0ffe294137171b8bac7d80a0b776
* changes:
DO NOT MERGE: Fixes NPE when preparing app data during init
Support privileged apps installed in APEX.
Support non-privileged APKs in APEX.
Refactors initial directory scan to be dryer
* changes:
[Tether10]Remove UserManagerInternal usage in Tethering
[Tether07] Migrate Tethering into module
[Tether07] Clean up build rule for libtetheroffloadjni
To be the unbundled application, tethering can not use UserManagerInternal
method anymore. Replace it by listening user restriction change intent.
Also fix some cleanup from
https://android-review.googlesource.com/c/platform/frameworks/base/+/1131002
Bug: 143195885
Test: -build, flash, boot
-atest TetheringTests
-manual test. Add a trigger point to call
UserManager.setUserRestriction(DISALLOW_CONFIG_TETHERING) from
Settings. Open hotspot -> restrict tethering from new adding trigger
point -> hotspot is OFF and Tethering settings are not available.
Change-Id: I4ed08d20f8b7476b8ecd5d018b75806dd2f22cc1
Merged-In: I4ed08d20f8b7476b8ecd5d018b75806dd2f22cc1
Now tethering would be run in dedicated service.
TetheringManager is the interface used to communicate with
TetheringService. The new call flow would be: ConnectivityManager
-> ConnectivityService -> TetheringManager -> TetheringService.
Note: the return value of #tether(), #untether() and #setUsbTethering()
APIs would always be no error. Client can use #getLastTetherError()
or #getTetheredIfaces or listen tether state change to check
status of corresponding interface.
Bug: 136040414
Bug: 144742179
Test: -build, flash, boot
-atest TetheringTests
-atest FrameworksNetTests
Change-Id: I7e78c0e0a3e70f940a749ba2a39ece7c7ec5b9b3
Merged-In: I7e78c0e0a3e70f940a749ba2a39ece7c7ec5b9b3
This change adds initial support for scanning APEX directories looking
for APK, and make them available to the system as normal system APKs.
For now privileged apps and resource overlay are not supported, and the
support will be added in a later CL.
Scanning features are inherited from the partition in which the
preinstalled version of a given APEX module is located.
Supports both flattened and un-flattened APEX configurations.
Bug: 138429615
Test: Compiled a test apex with the Snake app installed in it. Verified
app is correctly scanned and usable. Tried also installing the apex in
/vendor and /product.
Exempt-From-Owner-Approval: Approved in
https://googleplex-android-review.git.corp.google.com/c/platform/frameworks/base/+/9533445
Merged-In: I8fb91317da0296567403b2b62babbb71e4b48938
Change-Id: I8fb91317da0296567403b2b62babbb71e4b48938
For the AIDL compiler to call. Without this, the Java framework can't
pass callbacks to HALs.
Bug: 139325468
Test: use vibrator callback from Java
Change-Id: I147d3e3da699ce740f6589264b37f43f8557719b
Convenience utilities for working with HidlMemory.
Change-Id: I85a46d37de7b89fb48c57f69b1552ca59b802c19
Merged-In: I85a46d37de7b89fb48c57f69b1552ca59b802c19
Bug: 143566068
This change adds support for reading/writing a HidlMemory
instance into / out of a HwParcel and HwBlob, in a format that
is compatible with the hidl_memory C++ type.
This paves the way to be able to exchange shared memory blocks
between native and Java via HwBinder, which will be exposed
as adding Java support for the HIDL 'memory' type.
Change-Id: I6cbbf852218c8a631f9014e7caa7a97d17e11889
Bug: 143566068
Merged-In: I6cbbf852218c8a631f9014e7caa7a97d17e11889
Remove dead code from SystemServer that obstensively protects devices
from negative (< 1970) system clock times, but does not.
In reality, SystemServer calls SystemClock.setCurrentTimeMillis(), which
calls through to the AlarmManager.setTime() binder call. When the code
runs the system server hasn't started any services yet, so this would
always fail.
AlarmManager has similar protections in place that do work: if the
device is earlier than a "build time", then the device's clock is forced
forward. That provides the protection that the code being removed from
SystemServer was trying to achieve.
Test: Added logging to SystemClock to prove AlarmManager is null /
manual debugging
Change-Id: I87dca1062829bc1a10202804fa7ee31e0a6f03a3
