OSDN Git Service

android-x86/frameworks-base.git
7 years agoMerge "Removed legacy code used to read/write java arrays from HwBinder parcels."
Treehugger Robot [Tue, 20 Sep 2016 20:28:11 +0000 (20:28 +0000)]
Merge "Removed legacy code used to read/write java arrays from HwBinder parcels."

7 years agoMerge "[NAN] Rename to make discovery explicit"
Etan Cohen [Tue, 20 Sep 2016 16:41:04 +0000 (16:41 +0000)]
Merge "[NAN] Rename to make discovery explicit"

7 years agoMerge "WifiScanner: retrieve single scan results"
Rebecca Silberstein [Tue, 20 Sep 2016 15:50:07 +0000 (15:50 +0000)]
Merge "WifiScanner: retrieve single scan results"

7 years ago[NAN] Rename to make discovery explicit
Etan Cohen [Wed, 14 Sep 2016 21:41:56 +0000 (14:41 -0700)]
[NAN] Rename to make discovery explicit

Pure rename operation - no other functional changes.

(cherry-pick of commit ddb5f646562f31d7e78778a978e546f85312b009)

Bug: 31470256
Test: unit tests, integrated (sl4a) tests.
Change-Id: I5110d1a673b7ba4e39350d723b5984586fec585a

7 years agoRemoved legacy code used to read/write java arrays from HwBinder parcels.
Andreas Huber [Mon, 19 Sep 2016 19:36:19 +0000 (12:36 -0700)]
Removed legacy code used to read/write java arrays from HwBinder parcels.

More generalized functionality is now available through HwBlob.

Change-Id: Id6c713fb01af7593ea3ff8405f3a9e5325877412
Bug: 31438033
Test: run updated "hidl_test" and "hidl_test_java"

7 years agoMerge "[NAN] Update API: replace looper with handler"
Etan Cohen [Mon, 19 Sep 2016 20:47:36 +0000 (20:47 +0000)]
Merge "[NAN] Update API: replace looper with handler"

7 years ago[NAN] Update API: replace looper with handler
Etan Cohen [Fri, 9 Sep 2016 20:36:52 +0000 (13:36 -0700)]
[NAN] Update API: replace looper with handler

Per API council guidelines.

(cherry-pick of commit 0ff5bf0092eb25f579e2e93fb2f93bc953b39211)

Bug: 31377060
Test: unit tests, integrated (sl4a) tests
Change-Id: I05142001fb593399fb4675fbec965d7e08e0fac1

7 years agoMerge "Fix parameter description"
Treehugger Robot [Mon, 19 Sep 2016 20:02:27 +0000 (20:02 +0000)]
Merge "Fix parameter description"

7 years agoMerge "fix wrong algorithm in WifiPowerPerPacket()"
Treehugger Robot [Mon, 19 Sep 2016 18:28:46 +0000 (18:28 +0000)]
Merge "fix wrong algorithm in WifiPowerPerPacket()"

7 years agoMerge "[NAN] Make parcelable class final - per API linter"
Etan Cohen [Mon, 19 Sep 2016 16:03:59 +0000 (16:03 +0000)]
Merge "[NAN] Make parcelable class final - per API linter"

7 years ago[NAN] Make parcelable class final - per API linter
Etan Cohen [Fri, 9 Sep 2016 23:17:31 +0000 (16:17 -0700)]
[NAN] Make parcelable class final - per API linter

(cherry-pick of feef60406e961afcda8893c7560382382c358910)

Bug: 31380952
Test: build + sl4a integration test (WifiNanManagerTest:test_nan_discovery_session)
Change-Id: I74c079763d6b3a4f0119b16acd7a18014a10ebbd

7 years agofix wrong algorithm in WifiPowerPerPacket()
yuanhuihui [Mon, 12 Sep 2016 02:34:53 +0000 (10:34 +0800)]
fix wrong algorithm in WifiPowerPerPacket()

step1: PowerProfile.POWER_WIFI_ACTIVE represents energy consumption(mAh) per hour
devied by 3600, then WIFI_POWER ==> energy consumption(mAh) per second

step2: WIFI_BPS represents 1000000 bit per second
then (double)WIFI_BPS) / 8  ==> 1000000/8 Byte per second

step3: as upload and download, so divided  by 2;
then   (((double)WIFI_BPS) / 8 / 2048))  ==> 1000000/8/2048 KB per second
==> packet  per second (where 1 packet = 2 KB)

so WIFI_POWER / (((double)WIFI_BPS) / 8 / 2048) represents  mAh per Packet where 1 packet = 2 K.

when  divided by (60*60) again , that make WifiPowerEstimator narrow 3600 times.

Change-Id: Ic055a5145b6dfb1129c8969826329a3024c9e2b6
Signed-off-by: yuanhuihui <yuanhuihui@xiaomi.com>
7 years agoWifiScanner: retrieve single scan results
Rebecca Silberstein [Thu, 15 Sep 2016 05:07:04 +0000 (22:07 -0700)]
WifiScanner: retrieve single scan results

Add a new CMD_GET_SINGLE_SCAN_RESULTS message type and
getSingleScanResults method that allow WifiService to get
scan results directly from WifiScanner.

BUG: 31444878
Test: manually tested on angler
Test: frameworks/opt/net/wifi/tests/wifitests/runtests.sh on angler
Test: ran GTS CtsNetTestCases on angler
Change-Id: I705233d7dc4ae7e8480c53ff0f9e1b6c8875fecb

7 years agoMerge "WifiScanner: Change hidden/PNO scan params"
Roshan Pius [Fri, 16 Sep 2016 22:34:57 +0000 (22:34 +0000)]
Merge "WifiScanner: Change hidden/PNO scan params"

7 years agoMerge "Collect status on uncrypt time cost"
Tianjie Xu [Fri, 16 Sep 2016 18:19:22 +0000 (18:19 +0000)]
Merge "Collect status on uncrypt time cost"

7 years agoMerge "[AM] Skip unnessary ANR when process already died."
Treehugger Robot [Fri, 16 Sep 2016 18:16:22 +0000 (18:16 +0000)]
Merge "[AM] Skip unnessary ANR when process already died."

7 years agoMerge "Move art::Runtime::InitZygote logic from art into frameworks/base."
Treehugger Robot [Fri, 16 Sep 2016 17:23:41 +0000 (17:23 +0000)]
Merge "Move art::Runtime::InitZygote logic from art into frameworks/base."

7 years ago[AM] Skip unnessary ANR when process already died.
Mark Lu [Wed, 23 Mar 2016 11:08:44 +0000 (19:08 +0800)]
[AM] Skip unnessary ANR when process already died.

When app process been killed by AMS or lowmemkiller just before ANR report,
because process record info has been cleared after received death recipient,
it also cannot dump trace log because process already dead,
so report ANR & show ANR UI to let user wait seems is unnecessary.
(compare normal ANR case, if kill app process by command,
ANR dialog will also dismissed, it seems reasonable.)

To check above condition, if ANR process record killed set as true,
it means process already dead & can skip report this ANR.

Change-Id: I483cb02bacb10c32db80ca1097310b02abbac24d

7 years agoMerge "Stop including libcore/include/ in most of frameworks/base."
Treehugger Robot [Fri, 16 Sep 2016 08:36:44 +0000 (08:36 +0000)]
Merge "Stop including libcore/include/ in most of frameworks/base."

7 years agoMerge "Remove synchronization from getCurrentInputMethodSubtype()"
Treehugger Robot [Fri, 16 Sep 2016 05:21:33 +0000 (05:21 +0000)]
Merge "Remove synchronization from getCurrentInputMethodSubtype()"

7 years agoRemove synchronization from getCurrentInputMethodSubtype()
Gopal Krishna Shukla [Thu, 30 Jun 2016 06:04:21 +0000 (11:34 +0530)]
Remove synchronization from getCurrentInputMethodSubtype()

getCurrentInputMethodSubtype() acquires InputManagerService.mMethodMap
within its body. There seems to be no reason for holding
InputMethodManager.mH to call getCurrentInputMethodSubtype(). Holding mH
can cause potential deadlock b/w two threads acquiring mH and mMethodMap
in different orders.

Bug: 31247871
Bug: 31273203
Bug: b.android.com/218423
Change-Id: I20cf2c20f49b1b02c0f7a18257b49d4bcc081b5d

7 years agoStop including libcore/include/ in most of frameworks/base.
Elliott Hughes [Thu, 15 Sep 2016 20:13:39 +0000 (13:13 -0700)]
Stop including libcore/include/ in most of frameworks/base.

frameworks/base/core/jni/android_text_StaticLayout.cpp does actually use
ScopedIcuLocale, so I've left that for now.

Change-Id: I6458cd133871281a747a9da2e304da10b445051a

7 years agoMove art::Runtime::InitZygote logic from art into frameworks/base.
Robert Sesek [Tue, 13 Sep 2016 23:13:01 +0000 (19:13 -0400)]
Move art::Runtime::InitZygote logic from art into frameworks/base.

Test: m
Test: aosp_angler-eng boots
Change-Id: I964c70576136e7424797140b2f7d179e70298c44

7 years agoWifiScanner: Change hidden/PNO scan params
Roshan Pius [Mon, 12 Sep 2016 20:46:32 +0000 (13:46 -0700)]
WifiScanner: Change hidden/PNO scan params

Since networks are no longer configured in wpa_supplicant, network
IDs/priorities being passed around for hidden/PNO scanning does not
make any sense.
Changing the wifi scanner interface to remove all network ID/priority
references. These will be replaced by SSID of the corresponding
networks for hidden networks.

BUG: 29503772
TEST: Existing unit tests.
TEST: Successfully connected to a hidden network on device
Change-Id: I5d877e0345d574fb513b6eecca14f65f2ee4e7ab

7 years agoMerge "Skip ANR for processes that have been killed"
Treehugger Robot [Wed, 14 Sep 2016 13:54:36 +0000 (13:54 +0000)]
Merge "Skip ANR for processes that have been killed"

7 years agoMerge "app_main: failure to PR_SET_NO_NEW_PRIVS is always fatal."
Treehugger Robot [Wed, 14 Sep 2016 13:07:05 +0000 (13:07 +0000)]
Merge "app_main: failure to PR_SET_NO_NEW_PRIVS is always fatal."

7 years agoapp_main: failure to PR_SET_NO_NEW_PRIVS is always fatal.
Narayan Kamath [Tue, 13 Sep 2016 15:19:33 +0000 (16:19 +0100)]
app_main: failure to PR_SET_NO_NEW_PRIVS is always fatal.

Don't make allowances for older kernels. Kernels must handle this
properly since it is now a CTS requirement.

Also remove some unreachable return statements.

Change-Id: I280c4ed5e461d37672236747b5dbb7d1e97ecfec

7 years agoMerge "fix NullPointerException in RegisteredServicesCache"
Treehugger Robot [Tue, 13 Sep 2016 20:57:10 +0000 (20:57 +0000)]
Merge "fix NullPointerException in RegisteredServicesCache"

7 years agoFix parameter description
Cuihtlauac ALVARADO [Tue, 13 Sep 2016 12:49:41 +0000 (14:49 +0200)]
Fix parameter description

Correct parameter description text according to method name.

Change-Id: If6bd8254386a84efd338b58797c2c338ccb513b7

7 years agoMerge "DO NOT MERGE. Listen to ACTION_UID_REMOVED only."
Treehugger Robot [Mon, 12 Sep 2016 21:07:39 +0000 (21:07 +0000)]
Merge "DO NOT MERGE. Listen to ACTION_UID_REMOVED only."

7 years agoMerge "Add a flag to indicate if a ScanData is for a scan for all channels"
Mitchell Wills [Mon, 12 Sep 2016 20:38:20 +0000 (20:38 +0000)]
Merge "Add a flag to indicate if a ScanData is for a scan for all channels"

7 years agoMerge changes I104e6133,I54f04c03
Treehugger Robot [Mon, 12 Sep 2016 19:42:01 +0000 (19:42 +0000)]
Merge changes I104e6133,I54f04c03

* changes:
  Split the zygote logic out of android.os.Process into a new ZygoteProcess class.
  Refactor ZygoteInit to support a WebView-specific zygote.

7 years agoAdd a flag to indicate if a ScanData is for a scan for all channels
Mitchell Wills [Tue, 6 Sep 2016 22:08:36 +0000 (15:08 -0700)]
Add a flag to indicate if a ScanData is for a scan for all channels

BUG=30897947

Change-Id: I919d7fa9c3bf196371e506d23e393a9d609a5b4b
(cherry picked from commit 4bfb391c91b3539d2ec20ffd7e01056dafa0984d)

7 years agoMerge "Move frameworks/base to std::unique_ptr."
Elliott Hughes [Mon, 12 Sep 2016 16:50:17 +0000 (16:50 +0000)]
Merge "Move frameworks/base to std::unique_ptr."

7 years agoMove frameworks/base to std::unique_ptr.
Elliott Hughes [Sun, 11 Sep 2016 21:50:12 +0000 (14:50 -0700)]
Move frameworks/base to std::unique_ptr.

Bug: http://b/22403888
Change-Id: Iae2c68b00be7cd9dd215a75cf80b16d7403f6328

7 years agoMerge "Fix google-explicit-constructor warnings in media utils."
Chih-hung Hsieh [Sat, 10 Sep 2016 17:07:53 +0000 (17:07 +0000)]
Merge "Fix google-explicit-constructor warnings in media utils."

7 years agoMerge "Fix issue that Overview key doesn't work at secondary user"
Treehugger Robot [Fri, 9 Sep 2016 21:46:04 +0000 (21:46 +0000)]
Merge "Fix issue that Overview key doesn't work at secondary user"

7 years agoCollect status on uncrypt time cost
Tianjie Xu [Fri, 9 Sep 2016 18:27:32 +0000 (11:27 -0700)]
Collect status on uncrypt time cost

Read and report time cost of uncrypt (in seconds) from last_install.

Bug: 31383361

Change-Id: I283970c33b2a0324f2f51cde328e3e527eff7c58

7 years agoSplit the zygote logic out of android.os.Process into a new ZygoteProcess class.
Robert Sesek [Fri, 18 Mar 2016 20:52:57 +0000 (16:52 -0400)]
Split the zygote logic out of android.os.Process into a new ZygoteProcess class.

There is no functional change. This is to support adding new types of zygotes
that all operate using the same protocol.

Bug: 21643067
(cherry picked from commit 94e824bc1b5f83024b7c9a228a8f2d0cd1347d11)

Merged-In: Ie673ee816cae34ac20ffb8c774ec3e6461c3fd0a
Change-Id: I104e6133a90dc93a9854836b5e92d3cd542163a3

7 years agoRefactor ZygoteInit to support a WebView-specific zygote.
Tobias Sargeant [Tue, 19 Jan 2016 16:34:54 +0000 (16:34 +0000)]
Refactor ZygoteInit to support a WebView-specific zygote.

This is a non-functional change that separates out functionality
that should be shared between the system zygote and the WebView
zygote from that which is system zygote specific.

* Move MethodAndArgsCaller to Zygote.
* Split out server socket functions into ZygoteServer.
* Add a new (stub, for now) WebViewZygoteInit class.

Bug: 22084679
Bug: 21643067
(cherry picked from commit ba816e0c9efd8cd2aeef618a819a2ad46b742f87)

Merged-In: I4c508a42af7ab7b53d10570ad53b846df7782cc4
Change-Id: I54f04c03443d10dabe6426697d1ff8a0cc66b985

7 years agoMerge "char16 strings returned by JNIEnv::GetStringCritical are not '\0' terminated."
Treehugger Robot [Fri, 9 Sep 2016 15:47:44 +0000 (15:47 +0000)]
Merge "char16 strings returned by JNIEnv::GetStringCritical are not '\0' terminated."

7 years agoDO NOT MERGE. Listen to ACTION_UID_REMOVED only.
Felipe Leme [Fri, 9 Sep 2016 00:09:08 +0000 (17:09 -0700)]
DO NOT MERGE. Listen to ACTION_UID_REMOVED only.

NMPS was listening to both ACTION_UID_REMOVED and
ACTION_PACKAGE_REMOVED, which not only was redundant but would also
cause problems when a package with shared UIDs was removed.

BUG: 31375684

Change-Id: I78da95ef168321a837ed85ce2787a66a16830a78

7 years agochar16 strings returned by JNIEnv::GetStringCritical are not '\0' terminated.
Andreas Huber [Thu, 8 Sep 2016 20:52:36 +0000 (13:52 -0700)]
char16 strings returned by JNIEnv::GetStringCritical are not '\0' terminated.

Properly construct String16s from both pointer and length instead.

Bug: 31373735
Change-Id: Ia89df7840a48c4fce06e08fc3ae373e0fadfb44e

7 years agoMerge "Add Base64 from OpenJDK 8u60"
Tobias Thierer [Thu, 8 Sep 2016 13:10:14 +0000 (13:10 +0000)]
Merge "Add Base64 from OpenJDK 8u60"

7 years agoMerge "java.security.Provider: port to jdk8u60 (second part)"
Sergio Giro [Thu, 8 Sep 2016 12:18:43 +0000 (12:18 +0000)]
Merge "java.security.Provider: port to jdk8u60 (second part)"

7 years agoMerge "Fix SuppressWarnings typo."
Paul Duffin [Thu, 8 Sep 2016 12:17:05 +0000 (12:17 +0000)]
Merge "Fix SuppressWarnings typo."

7 years agoFix SuppressWarnings typo.
Ian Rogers [Fri, 27 May 2016 21:31:19 +0000 (14:31 -0700)]
Fix SuppressWarnings typo.

As Error Prone states:
Suppressing "deprecated" is probably a typo for "deprecation"
Bug: 27723540

(cherry picked from commit a7f834f1ce5ec52acdde5a8619d91a18f3376598)

Change-Id: I0c6a9fc0a160769955cccf97ec7decb1f2b9b8fb

7 years agoMerge "Track libcore change 76fa41aa4ac891670446592243b654fac7deb53f"
Yi Kong [Thu, 8 Sep 2016 09:59:54 +0000 (09:59 +0000)]
Merge "Track libcore change 76fa41aa4ac891670446592243b654fac7deb53f"

7 years agofix NullPointerException in RegisteredServicesCache
liulvping [Thu, 8 Sep 2016 01:16:48 +0000 (09:16 +0800)]
fix NullPointerException in RegisteredServicesCache

https://code.google.com/p/android/issues/detail?id=221932

Change-Id: I67c0b799d10f4b9c8fa0779069d23cc49916b395
Signed-off-by: liulvping <liulvping@xiaomi.com>
7 years agoMerge "WifiManager: Remove unused methods"
Treehugger Robot [Wed, 7 Sep 2016 21:43:19 +0000 (21:43 +0000)]
Merge "WifiManager: Remove unused methods"

7 years agoTrack libcore change 76fa41aa4ac891670446592243b654fac7deb53f
Yi Kong [Mon, 29 Aug 2016 15:47:00 +0000 (16:47 +0100)]
Track libcore change 76fa41aa4ac891670446592243b654fac7deb53f

Bug: 30992227
Change-Id: I33c25562468f5803e211910f50116b501b66fe56

7 years agoFix issue that Overview key doesn't work at secondary user
dooyoung.hwang [Wed, 31 Aug 2016 05:15:22 +0000 (14:15 +0900)]
Fix issue that Overview key doesn't work at secondary user

If user captures screenshot, screenshot process is started. Screenshot
process also starts services that is defined SERVICES_PER_USER.
As a result Recents class's object is unexpectedly constucted from
screenshot process, and cause to connect with system-user's SystemUI
with mUserToSystemServiceConnection. So, binder proxy from system-user's
SystemUI to secondary-user's SystemUI is replaced with proxy to
screenshot process. In this case Overview key doesn't work at all
until reboot.

Bug:30340532
Change-Id: I84b8b4a02ac3ff781e06d57be19cff56efa76521

7 years agoMerge "WifiEnterpriseConfig: Refactor set/get field values"
Roshan Pius [Wed, 7 Sep 2016 15:47:39 +0000 (15:47 +0000)]
Merge "WifiEnterpriseConfig: Refactor set/get field values"

7 years agoSkip ANR for processes that have been killed
Tobias Lindskog [Tue, 3 Feb 2015 10:45:58 +0000 (11:45 +0100)]
Skip ANR for processes that have been killed

If a controller is attached and decides to kill a process after an ANR,
other ANR reports for that app that are queued up won't be handled until
after the app has died.  This can create a report without the relevant
callstacks because the app is dead by the time the traces are dumped.
Since the trace file is global, the traces recorded for the first ANR
are overwritten, leaving us with no clue as to what happened.

After the app has been killedByAm, it is not interesting to handle ANRs
for that app until it is started again, so killedByAm can be used to
filter out these spurious reports.

Change-Id: I34ba790f6d29d563c819dc2f6ac71a3c8955bb76

7 years agoMerge "java.security.cert.X509CRL: add new method verify(PublicKey, Provider)"
Sergio Giro [Wed, 7 Sep 2016 12:08:27 +0000 (12:08 +0000)]
Merge "java.security.cert.X509CRL: add new method verify(PublicKey, Provider)"

7 years agoMerge "java.security.PKCS12Attribute: port class from jdk8u60"
Sergio Giro [Wed, 7 Sep 2016 12:07:13 +0000 (12:07 +0000)]
Merge "java.security.PKCS12Attribute: port class from jdk8u60"

7 years agoMerge "java.security.SecureRandom: add getInstanceStrong() method"
Sergio Giro [Wed, 7 Sep 2016 12:05:35 +0000 (12:05 +0000)]
Merge "java.security.SecureRandom: add getInstanceStrong() method"

7 years agoMerge "java.security: port DomainLoadStoreParameter from jdk8u60"
Sergio Giro [Wed, 7 Sep 2016 12:04:20 +0000 (12:04 +0000)]
Merge "java.security: port DomainLoadStoreParameter from jdk8u60"

7 years agoMerge "Add a test app for suppressed exceptions"
Yohann Roussel [Wed, 7 Sep 2016 09:52:49 +0000 (09:52 +0000)]
Merge "Add a test app for suppressed exceptions"

7 years agoMerge changes from topic 'enable_nan'
Etan Cohen [Wed, 7 Sep 2016 04:19:51 +0000 (04:19 +0000)]
Merge changes from topic 'enable_nan'

* changes:
  Give WakeupMessage the ability to transport an object as well. [DO NOT MERGE]
  [NAN] Add transport type for NAN. [DO NOT MERGE]

7 years agoGive WakeupMessage the ability to transport an object as well. [DO NOT MERGE]
Lorenzo Colitti [Tue, 7 Jun 2016 07:27:10 +0000 (16:27 +0900)]
Give WakeupMessage the ability to transport an object as well. [DO NOT MERGE]

(cherry-pick of commit eab17da5882e59caff25c73aa6bcd0587aca9dd4)

Bug: 23113288
Change-Id: Ic98e3bcb3ea8b0b28a309bd647fb4178311c46a3

7 years ago[NAN] Add transport type for NAN. [DO NOT MERGE]
Etan Cohen [Mon, 20 Jun 2016 16:27:12 +0000 (09:27 -0700)]
[NAN] Add transport type for NAN. [DO NOT MERGE]

Add TRANSPORT_WIFI_NAN - specifies a peer-to-peer Wi-Fi NAN data
connectivity request.

While NAN uses Wi-Fi L1 PHY and part of the MAC - it is a different
protocol and constitutes a different transport.

(cherry-pick of commit 7b84987a6acf2f0a950c334a17a6ee085db3ba4f)

Bug: 26564277
Change-Id: I975ebc72d8f97a592d18038b3d6465b7a40efa75

7 years agoMerge "WifiConfiguration: Fix isEnterprise"
Treehugger Robot [Tue, 6 Sep 2016 22:51:42 +0000 (22:51 +0000)]
Merge "WifiConfiguration: Fix isEnterprise"

7 years agoMerge "Avoid profile guided compilation for vmSafeMode"
Mathieu Chartier [Tue, 6 Sep 2016 22:19:03 +0000 (22:19 +0000)]
Merge "Avoid profile guided compilation for vmSafeMode"

7 years agoMerge "fixes mStartedUsers shoule be locked by mLock, not mVolumes"
Treehugger Robot [Tue, 6 Sep 2016 21:09:05 +0000 (21:09 +0000)]
Merge "fixes mStartedUsers shoule be locked by mLock, not mVolumes"

7 years agoMerge changes from topic 'cherry_pms_installd'
Andreas Gampe [Tue, 6 Sep 2016 21:02:53 +0000 (21:02 +0000)]
Merge changes from topic 'cherry_pms_installd'

* changes:
  Release mPackages lock earlier
  OtaDexopt: Special-case system app classpaths
  OtaDexoptService: Run dexopt on all apps
  Frameworks/base: Add compiler stats to Package Manager
  [RFC] Special case system apps for profile optimizations
  Compile core-apps according to their own "reason" for A/B
  Frameworks/base: Add new flow to OtaDexoptService
  Otadexopt: Expose progress percentage
  cp preopted files from B partition during PM init.

7 years agoWifiManager: Remove unused methods
Roshan Pius [Tue, 6 Sep 2016 20:38:31 +0000 (13:38 -0700)]
WifiManager: Remove unused methods

Remove the unused hidden methods from WifiManager.

BUG: 29449615
TEST: Compiles, boots up.
Change-Id: Ie7e6f438b23a368e44d35390267d5505e12bb75e

7 years agoRelease mPackages lock earlier
Todd Kennedy [Thu, 18 Aug 2016 18:22:52 +0000 (11:22 -0700)]
Release mPackages lock earlier

We cannot hold mPackages when calling into generatePackageDexopts().
This method takes Package Manager's mInstallLock. By holding mPackages,
we have lock inversion and hilarity ensues.

Change-Id: Ia11a158677051e3511702f38cde6780e75b256fb
Fixes: 30927731
(cherry picked from commit a8d4f489974f3ea8f73990cbabbce205343fb926)

7 years agoOtaDexopt: Special-case system app classpaths
Andreas Gampe [Tue, 2 Aug 2016 05:08:26 +0000 (22:08 -0700)]
OtaDexopt: Special-case system app classpaths

Specialize system apps' shared libraries so that classpath checks
are elided.

Bug: 30568568
Change-Id: Id1f2555ef43984b616e01f8596701ccba77895b3
(cherry picked from commit c7e02c1d2a53e0f738d2a8383e03658b39d91ce6)

7 years agoOtaDexoptService: Run dexopt on all apps
Andreas Gampe [Thu, 28 Jul 2016 23:54:41 +0000 (16:54 -0700)]
OtaDexoptService: Run dexopt on all apps

To avoid work after reboot, dexopt all apps. However, unimportant
apps are optimized with the first-boot reason (which is usually
interpret-only) for space reasons.

Bug: 30468718
Change-Id: Ia05d879957388967c69f9380e50d8a9834afe07d
(cherry picked from commit 115514b236ebe7cb3c90892891954b8c7ba69335)

7 years agoFrameworks/base: Add compiler stats to Package Manager
Andreas Gampe [Wed, 13 Jul 2016 05:42:41 +0000 (22:42 -0700)]
Frameworks/base: Add compiler stats to Package Manager

Add a simple class for storing compiler statistics. Capture compile
times for code paths from a package.

Bug: 29223204
Change-Id: I1b066de6a83a739470a42480eee0bfef88423eea
(cherry picked from commit 37e5fdc6b4963f3533caecdd92b129f79da69dd8)

7 years ago[RFC] Special case system apps for profile optimizations
Calin Juravle [Tue, 12 Jul 2016 14:56:41 +0000 (15:56 +0100)]
[RFC] Special case system apps for profile optimizations

If the OTA updates a system app which was previously preopted to a non-
preopted state the app might end up being verified at runtime. That's
because by default the apps are verify-profile but for preopted apps
there's no profile.

Do a hacky check to ensure that if we have no profiles (a reasonable
indication that before the OTA the app was preopted) system apps get
compiled with a non-profile filter (by default interpret-only).

Bug: 30032273
Test: Andreas "has verified that Calin's change to A/B works as expected
and promotes things like SystemUI to speed. From my side, that's
ready to be merged"

Change-Id: I7a052a8ea76cab7f649dc993237ea05534d6c4b9
(cherry picked from commit 0bd776207999ccba17e5adb163710bd7b16ac907)

7 years agoCompile core-apps according to their own "reason" for A/B
Calin Juravle [Tue, 12 Jul 2016 12:56:34 +0000 (13:56 +0100)]
Compile core-apps according to their own "reason" for A/B

If the core apps are not preopted in the B OTA, and REASON_AB_OTA is not
speed (by default is speed-profile) they will be interepreted/JITed.
This in itself is not a problem as we will end up doing profile guided
compilation. However, some core apps may be loaded by system server
which doesn't JIT and we need to make sure we don't interpret-only.

Bug: 30032273
Change-Id: Idb11b1c01c4c4ceba043feb3a9ccc6958035d3b7
(cherry picked from commit 565621328dd4fc2e179a34494d0f8b8dc1b1eae9)

7 years agoFrameworks/base: Add new flow to OtaDexoptService
Andreas Gampe [Fri, 24 Jun 2016 03:27:12 +0000 (20:27 -0700)]
Frameworks/base: Add new flow to OtaDexoptService

Add functionality to capture/intercept installd communication, and
use this to return the full communication for dexopt. These parameters
can be used to drive otapreopt_chroot directly.

Keep the old direct invocation alive until devices have transitioned
to a service that exposes this API.

In preparation for renaming of A/B OTA artifacts to include target
slot names.

Bug: 25612095
Bug: 28069686
Change-Id: I14728ee1266f3882cada8f08dd21891ed5f7a0cb
(cherry picked from commit cc241a580cb9b753d0dde0cea578feb74ad517e7)

7 years agoOtadexopt: Expose progress percentage
Andreas Gampe [Fri, 10 Jun 2016 22:21:39 +0000 (15:21 -0700)]
Otadexopt: Expose progress percentage

To be able to report progress of an A/B OTA dexopt, expose a
progress function that the script can query.

Bug: 25612095
Bug: 29223204
Change-Id: Ie8162946d18f6fa78649a40ad5d3949d31a181cd
(cherry picked from commit bf06232f4d440ced8230662a77ca0e8ece6383ca)

7 years agoWifiConfiguration: Fix isEnterprise
Roshan Pius [Tue, 6 Sep 2016 17:13:40 +0000 (10:13 -0700)]
WifiConfiguration: Fix isEnterprise

Add a check for the eap method in Enterpriseconfig associated with the
WifiConfiguration.

BUG: 31275240
Change-Id: I21aabb09bdd21e29a80d4b9c6ea0e2b1f0d7e2df

7 years agoWifiEnterpriseConfig: Refactor set/get field values
Roshan Pius [Fri, 2 Sep 2016 18:05:17 +0000 (11:05 -0700)]
WifiEnterpriseConfig: Refactor set/get field values

WifiEnterpriseConfig has a few keys which are generated internally
have unquoted values. However, the public setFieldValue() always quotes
the value when set.  So, this causes a problem when restoring
these field values from config store. Since this is an internal
disctinction that only WifiEnterpriseConfig understands, add a list to
check if a particular field value needs to be quoted or not. Also,
remove any direct "mFields.put" invocations with |setFieldValue|.

While there,
Refactor the existing |setFieldValue| & |getFieldValue| methods.

BUG: 31246524
TEST: Unit tests
TEST: Integrated with config store and verified that a previous TLS EAP
connection failure is no longer seen.

Change-Id: Ib85f3bce378720a6a6c2ae1439837a8e866a088d

7 years agocp preopted files from B partition during PM init.
Alex Light [Thu, 23 Jun 2016 00:56:37 +0000 (17:56 -0700)]
cp preopted files from B partition during PM init.

This allows us to save space on the system partition while still
having access to the preopted files.

We do this on first boot when the "ro.cp_system_other_odex" property
is set to 1. We do this during package manager initialization before
scanning the system to see which apks need to be optimized again.

Note that a separate script, run by init, is actually responsible for
finding and copying the files. We simply request that it runs.

Bug: 29278988

Change-Id: I8d7c790ad35b32a0ce1d87939f043419bae4d88a
(cherry picked from commit 3dafee6c1820bf0946bab04b290c5a757112d3e7)

7 years agojava.security.cert.X509CRL: add new method verify(PublicKey, Provider)
Sergio Giro [Tue, 6 Sep 2016 13:09:41 +0000 (14:09 +0100)]
java.security.cert.X509CRL: add new method verify(PublicKey, Provider)

New method introduced in Java 8. This method has a "default" implementation
in order to avoid breaking providers who construct concrete subclasses.

Unfortunately this implementation always results in an infinite loop, and
has thus been changed by an UnsupportedMethodOperation, with a TODO note
as to improve the situation if possible.

java/security/cert/X509CRL: add "default" method
sun/security/x509/X509CRLImpl: add declaration for that particular
implementation

Also, cosmetic changes like:
- use of @code tags instead of <code>
- use of static fields using the class name instead of instance name

This completes the port of this package to jdk8u60.

Bug: 29631070
Bug: 31294527
Test: make droid docs; old-cts run cts --class libcore.java.security.cert.X509CRLTest
Change-Id: Iccdfa3625a9ff2e7b808c22bce59ff75ae903e8a

7 years agoAdd Base64 from OpenJDK 8u60
Tobias Thierer [Thu, 1 Sep 2016 11:47:34 +0000 (12:47 +0100)]
Add Base64 from OpenJDK 8u60

Test: make update-api
Bug: 29935305

Change-Id: Idc1cfadd78553dd05a7cb1f9ebd02858a3675e51

7 years agofixes mStartedUsers shoule be locked by mLock, not mVolumes
yuanhuihui [Wed, 13 Jul 2016 13:21:03 +0000 (21:21 +0800)]
fixes mStartedUsers shoule be locked by mLock, not mVolumes

when create/delete multiple user, plug/unplug externel storage(such as sdcard) at the same time.
multiple threads run concurrency , the newly created user can not get state of externel storage.

https://code.google.com/p/android/issues/detail?id=215901

Change-Id: I097f5291108af89a430e12045376065d71d48175
Signed-off-by: yuanhuihui <yuanhuihui@xiaomi.com>
7 years agoMerge "Updated user document for Single-Source RenderScript"
Treehugger Robot [Sat, 3 Sep 2016 06:22:38 +0000 (06:22 +0000)]
Merge "Updated user document for Single-Source RenderScript"

7 years agoAvoid profile guided compilation for vmSafeMode
Mathieu Chartier [Fri, 2 Sep 2016 23:36:42 +0000 (16:36 -0700)]
Avoid profile guided compilation for vmSafeMode

Bug: 30688277
Bug: 30733908

Test: adb shell cmd package compile -m speed-profile comb.BBClient

Change-Id: I6783c80566b6b3b67025a3d258ebbdede0c7ca2d

7 years agoUpdated user document for Single-Source RenderScript
Yang Ni [Wed, 29 Jun 2016 19:44:06 +0000 (12:44 -0700)]
Updated user document for Single-Source RenderScript

Bug: 29875503
Bug: 29879448

Added a section introducing the new single-source feature.

Local staging:
http://yangni.mtv.corp.google.com/guide/topics/renderscript/compute.html

This updates
https://developer.android.com/guide/topics/renderscript/compute.html

Change-Id: I62dda3ab60b1678a9580fd2873f64f33d9696e13
(cherry picked from commit c069ed7f8c3d307c9fed2c8d1a81e28152190b5a)

7 years agoMerge "WifiEnterpriseConfig: Fix copy constructor"
Roshan Pius [Fri, 2 Sep 2016 21:10:10 +0000 (21:10 +0000)]
Merge "WifiEnterpriseConfig: Fix copy constructor"

7 years agoWifiEnterpriseConfig: Fix copy constructor
Roshan Pius [Thu, 1 Sep 2016 20:46:40 +0000 (13:46 -0700)]
WifiEnterpriseConfig: Fix copy constructor

Add missing elements from WifiEnterpriseConfig copy
constructor.

BUG: 31245222
TEST: Ran ACTS WifiEnterpriseTest suite.
Change-Id: Ia6f95fca4131ad9a1facca8f9c4c856dff868058

7 years agoMerge changes from topic 'java_hidl_support'
Iliyan Malchev [Fri, 2 Sep 2016 20:09:25 +0000 (20:09 +0000)]
Merge changes from topic 'java_hidl_support'

* changes:
  Move Status to libhidl (DO NOT MERGE)
  Adds framework support for hidl-gen Java backend. (to support structs) (DO NOT MERGE)
  Add Bool* APIs to HwParcel (DO NOT MERGE)
  Link against libhidl for HidlSupport/svcmgr (DO NOT MERGE)
  Support one-way methods in java support for hardware binder (DO NOT MERGE)
  Initial commit of Java support for hardware binder (DO NOT MERGE)

7 years agoMove Status to libhidl (DO NOT MERGE)
Martijn Coenen [Wed, 31 Aug 2016 23:37:05 +0000 (01:37 +0200)]
Move Status to libhidl (DO NOT MERGE)

Bug: 31226252

Change-Id: Ieb0adcf1fa71d6859324c2d83841685d1e1a22f8
Signed-off-by: Iliyan Malchev <malchev@google.com>
7 years agoAdds framework support for hidl-gen Java backend. (to support structs) (DO NOT MERGE)
Andreas Huber [Thu, 25 Aug 2016 18:21:21 +0000 (11:21 -0700)]
Adds framework support for hidl-gen Java backend. (to support structs) (DO NOT MERGE)

Bug: 30575790
Change-Id: Ida30d8fe7a1b210e98f1a0ea5d429a0112f9ef3f
Signed-off-by: Iliyan Malchev <malchev@google.com>
7 years agoAdd Bool* APIs to HwParcel (DO NOT MERGE)
Andreas Huber [Wed, 24 Aug 2016 23:19:03 +0000 (16:19 -0700)]
Add Bool* APIs to HwParcel (DO NOT MERGE)

Bug: 31045584
Change-Id: I202c4aa7caf92a4cd7e3a45ef6784638e41facc8
Signed-off-by: Iliyan Malchev <malchev@google.com>
7 years agoLink against libhidl for HidlSupport/svcmgr (DO NOT MERGE)
Martijn Coenen [Fri, 19 Aug 2016 12:23:42 +0000 (14:23 +0200)]
Link against libhidl for HidlSupport/svcmgr (DO NOT MERGE)

Bug: 30839546
Change-Id: I52609576aa7f10da608725ef4ea78b778f127eef
Signed-off-by: Iliyan Malchev <malchev@google.com>
7 years agoSupport one-way methods in java support for hardware binder (DO NOT MERGE)
Andreas Huber [Thu, 18 Aug 2016 21:29:40 +0000 (14:29 -0700)]
Support one-way methods in java support for hardware binder (DO NOT MERGE)

Bug: 30922538
Change-Id: I5ff93126a29f6bff42dee3f9868fa794ca7c077f
Signed-off-by: Iliyan Malchev <malchev@google.com>
7 years agoInitial commit of Java support for hardware binder (DO NOT MERGE)
Andreas Huber [Mon, 15 Aug 2016 16:25:02 +0000 (09:25 -0700)]
Initial commit of Java support for hardware binder (DO NOT MERGE)

Change-Id: If1098ab921a11bae8eca2a70a3c3070e4daa0ea2

7 years agoMerge "NetworkSelectionStatus: Fix copy constructor"
Treehugger Robot [Fri, 2 Sep 2016 05:14:41 +0000 (05:14 +0000)]
Merge "NetworkSelectionStatus: Fix copy constructor"

7 years agoNetworkSelectionStatus: Fix copy constructor
Roshan Pius [Tue, 30 Aug 2016 16:48:47 +0000 (09:48 -0700)]
NetworkSelectionStatus: Fix copy constructor

Add one more missing element from NetworkSelectionStatus copy constructor.

BUG: 31102798
Change-Id: Iae18f944e7ef722b280b20a51f128f90b2ff45c7

7 years agoMerge "Fix google-explicit-constructor warnings."
Chih-hung Hsieh [Thu, 1 Sep 2016 17:33:42 +0000 (17:33 +0000)]
Merge "Fix google-explicit-constructor warnings."

7 years agojava.security.SecureRandom: add getInstanceStrong() method
Sergio Giro [Tue, 30 Aug 2016 15:53:53 +0000 (16:53 +0100)]
java.security.SecureRandom: add getInstanceStrong() method

Port SecureRandom to jdk8u60.

The method getInstanceStrong returns a strong random number
generator. In Android this is assumed to be the one from
OpenSSLProvider.

This commit also contains cosmetic JavaDoc changes like
- Use of @code in place of <code> tags and use of a package-info.java in
place of a package.html file.
- Added comments.
- Uses of java.security.util.Debug (commented-out as that
debugging mechanism is not available in Android).
- Added braces in control flow blocks.

Test: make droid docs, vogar SecureRandomTest
Bug: 29631070
Change-Id: I9080fbd5e9292d7cb3eaa234f7c5849411336fac

7 years agojava.security.Provider: port to jdk8u60 (second part)
Sergio Giro [Thu, 1 Sep 2016 17:03:15 +0000 (18:03 +0100)]
java.security.Provider: port to jdk8u60 (second part)

- Add overridings for HashMap operations

Provider is a subclass of HashMap. When operations are on
Strings, they must be processed as part of the
algorithm->implementation mapping of the provider, and
so these methods must be overriden accordingly.

Newly overridden Methods added are:

compute, computeIfAbsent, computeIfPresent, getOrDefault,
merge, putIfAbsent, replace, replaceAll.

- Check that classes and constructors are public when
creating instances of services.

- Cosmetic changes like use of {@code} instead of <code>

Bug: 29631070
Test: make droid docs; vogar ProviderTest
Change-Id: I0238f4a48f31ddc50a3f18c27515023940d9ba02

7 years agoMerge "Track exposure of new j.l.reflect methods"
Neil Fuller [Thu, 1 Sep 2016 13:15:52 +0000 (13:15 +0000)]
Merge "Track exposure of new j.l.reflect methods"

7 years agojava.security.PKCS12Attribute: port class from jdk8u60
Sergio Giro [Wed, 31 Aug 2016 19:00:41 +0000 (20:00 +0100)]
java.security.PKCS12Attribute: port class from jdk8u60

This class provides facilities to deal with attributes in entries for
PKCS12 keystores.

Test: check-ojluni-files, make droid docs, vogar PKCS12AttributeTest
Bug: 29631070
Change-Id: I8d40fe126f189d503d6df7bfa45c4763bb7cb0d1