OSDN Git Service

android-x86/frameworks-base.git
10 years agoMerge "AArch64: Make eglGetDisplay(int) work for EGL_DEFAULT_DISPLAY"
Narayan Kamath [Thu, 27 Feb 2014 13:18:06 +0000 (13:18 +0000)]
Merge "AArch64: Make eglGetDisplay(int) work for EGL_DEFAULT_DISPLAY"

10 years agoMerge "Make getAssetInt throw unconditionally."
Narayan Kamath [Thu, 27 Feb 2014 13:13:08 +0000 (13:13 +0000)]
Merge "Make getAssetInt throw unconditionally."

10 years agoAArch64: Make eglGetDisplay(int) work for EGL_DEFAULT_DISPLAY
Ashok Bhat [Wed, 26 Feb 2014 10:00:48 +0000 (10:00 +0000)]
AArch64: Make eglGetDisplay(int) work for EGL_DEFAULT_DISPLAY

Note that files changed in this patch have been automatically
generated by running frameworks/native/opengl/tools/glgen/gen
script

This will allow eglGetDisplay(int) to work on both 32-bit and
64-bit systems when EGL_DEFAULT_DISPLAY is passed as a parameter.

Change-Id: Idf27d6e00d623d331bb2d4d7f85fa450e0db26c9
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
10 years agoMerge "Fix AOSP build - DO NOT MERGE"
Chris Craik [Wed, 26 Feb 2014 18:48:27 +0000 (18:48 +0000)]
Merge "Fix AOSP build - DO NOT MERGE"

10 years agoMerge "Fix graphics corruption caused by HWUI caches"
Chris Craik [Wed, 26 Feb 2014 18:47:59 +0000 (18:47 +0000)]
Merge "Fix graphics corruption caused by HWUI caches"

10 years agoFix AOSP build - DO NOT MERGE
Chris Craik [Wed, 26 Feb 2014 18:44:33 +0000 (10:44 -0800)]
Fix AOSP build - DO NOT MERGE

Change-Id: I42b420c2ce89ce364a2809b28b827964e6923fa1

10 years agoFix graphics corruption caused by HWUI caches
Sangkyu Lee [Thu, 9 Jan 2014 05:11:57 +0000 (14:11 +0900)]
Fix graphics corruption caused by HWUI caches

Some caches(PatchCache, TextureCache, PathCache) for HWUI
uses deferred removal for their cache entries even though
actual resource objects are immediately freed by
ResourceCache.
For this reason, the uniqueness of a resource address in
the caches is not guaranteed in specific cases.
(Because malloc() can return the same address when malloc()
and free() called very frequently.)

So it can be possible the cache have two cache entries for
two different resources but the same memory address.
(Of course one of the resources is already freed.)
It also can be possible mGarbage vector in PatchCache has
duplicated addresses and this can lead to duplicated free
blocks in the free block list and graphics corruption.
(Deferred removal was implmeneted based on an assumption of
unique resource addresses.)

So this patch makes sure resource objects are freed after
the resources are removed from the caches to guarantee
the uniqueness of a resource address and prevent graphics
corruption.

Change-Id: I040f033a4fc783d2c4bc04b113589657c36fb15b
Signed-off-by: Sangkyu Lee <sk82.lee@lge.com>
10 years agoMake getAssetInt throw unconditionally.
Narayan Kamath [Tue, 25 Feb 2014 15:48:07 +0000 (15:48 +0000)]
Make getAssetInt throw unconditionally.

All callers have been moved over to getNativeAsset, and
this method has been hidden from the public API internally.

Change-Id: I5f29b5ddb4449adadf426e49a4085c7320289a15

10 years agoMerge "Use reinterpret_cast when an integer is cast to a pointer"
Narayan Kamath [Tue, 25 Feb 2014 13:05:47 +0000 (13:05 +0000)]
Merge "Use reinterpret_cast when an integer is cast to a pointer"

10 years agoMerge "Use long for pointers in opengl/EGL classes"
Narayan Kamath [Tue, 25 Feb 2014 13:05:43 +0000 (13:05 +0000)]
Merge "Use long for pointers in opengl/EGL classes"

10 years agoUse long for pointers in opengl/EGL classes
Ashok Bhat [Mon, 24 Feb 2014 10:06:55 +0000 (10:06 +0000)]
Use long for pointers in opengl/EGL classes

Note that files changed in this patch have been automatically
generated by running frameworks/native/opengl/tools/glgen/gen script

This patch updates EGL classes in frameworks/base to support
64-bit platforms. Key changes in the EGL classes include

[x] EGLObjectHandle class - EGLObjectHandle class has two public
    methods (constructor and getHandle) that assume handles are
    32-bit. They have not been changed. Instead, two new hidden
    methods (EGLObjectHandle(long) and getNativeHandle) have been
    added.

[x] EG14 class - Two public methods eglGetDisplay and
    eglCreatePbufferFromClientBuffer assume that handles are 32-bit.
    They have been changed to throw unsupported operation exception
    on non 32-bit machines. Two new methods eglGetDisplay(long)
    and eglCreatePbufferFromClientBuffer(...long buffer..) have
    been added to support 64-bit handles.

Change-Id: I9e0f064e5b33700eb0baa2e1841a21f931f7a765
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
10 years agoUse reinterpret_cast when an integer is cast to a pointer
Ashok Bhat [Mon, 24 Feb 2014 09:49:07 +0000 (09:49 +0000)]
Use reinterpret_cast when an integer is cast to a pointer

Note that files changed in this patch have been generated
by running frameworks/native/opengl/tools/glgen/gen script.

Change-Id: Iff3ac2e804af41b9a6ec073fd65019c6806fe821
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
10 years agoMerge changes I6d4fdada,Ia34899a4
Greg Hackmann [Fri, 21 Feb 2014 21:01:00 +0000 (21:01 +0000)]
Merge changes I6d4fdada,Ia34899a4

* changes:
  open("/dev/rtc0") failure in AlarmManagerService.setTime() should be non-fatal
  Move time setting code from SystemClock to AlarmManagerService

10 years agoopen("/dev/rtc0") failure in AlarmManagerService.setTime() should be non-fatal
Greg Hackmann [Tue, 28 Jan 2014 00:30:09 +0000 (16:30 -0800)]
open("/dev/rtc0") failure in AlarmManagerService.setTime() should be non-fatal

Setting the time-of-day clock is still useful on systems where the RTC
device is not yet brought up or otherwise unavailable.  This matches the
in-kernel behavior of the Android alarm driver.

Change-Id: I6d4fdadab12e241ada7419425efd55bd13873c55
Signed-off-by: Greg Hackmann <ghackmann@google.com>
10 years agoMove time setting code from SystemClock to AlarmManagerService
Greg Hackmann [Thu, 20 Feb 2014 00:39:36 +0000 (16:39 -0800)]
Move time setting code from SystemClock to AlarmManagerService

On devices using /dev/rtc instead of /dev/alarm, updating the
time-of-day clock and RTC are separate syscalls.  Hence the clock and
RTC could be left in inconsistent states if two threads called
SystemClock.setCurrentTimeMillis() simultaneously.

By moving this code into AlarmManagerService, we can put a global lock
around AlarmManagerService.setTime() and prevent the race condition.

Note that access to SystemClock.setCurrentTimeMillis() is now gated by
android.permission.SET_TIME, where before it was gated by filesystem
permissions (i.e., could the process write to /dev/alarm or /dev/rtc).

Change-Id: Ia34899a4cde983656305fd2ef466dfe908ed23c8
Signed-off-by: Greg Hackmann <ghackmann@google.com>
10 years agoMerge "Fix documentation for Parcel.readException()."
Elliott Hughes [Fri, 21 Feb 2014 17:21:52 +0000 (17:21 +0000)]
Merge "Fix documentation for Parcel.readException()."

10 years agoMerge "Miscellaneous small documentation fixes."
Elliott Hughes [Fri, 21 Feb 2014 17:19:21 +0000 (17:19 +0000)]
Merge "Miscellaneous small documentation fixes."

10 years agoMerge "Remove unused JNITest class"
Narayan Kamath [Fri, 21 Feb 2014 15:02:30 +0000 (15:02 +0000)]
Merge "Remove unused JNITest class"

10 years agoRemove unused JNITest class
Ashok Bhat [Sat, 15 Feb 2014 11:28:48 +0000 (11:28 +0000)]
Remove unused JNITest class

JNITest class is no longer actively used. This patch
removes the class (java and jni) files.

JNI interfaces and calls are extensively tested in
the art unit tests (art/tests) and in cts (see
CtsJniTestCases).

Change-Id: I62f7c72deb5d206fa3f545ae39a9cb9011110d0a
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
10 years agoMerge "Fixes for 64bit in libhwui"
Narayan Kamath [Fri, 21 Feb 2014 14:23:57 +0000 (14:23 +0000)]
Merge "Fixes for 64bit in libhwui"

10 years agoFixes for 64bit in libhwui
Kévin PETIT [Thu, 13 Feb 2014 11:03:40 +0000 (11:03 +0000)]
Fixes for 64bit in libhwui

Using reinterpret_cast allows conversions between pointer types
and integers which don't have the same size. The corresponding
flags is -Wint-to-pointer-cast.

Change-Id: I8ff0c79d235fa6d07b8d9305edc185d946d5133d
Signed-off-by: Kévin PETIT <kevin.petit@arm.com>
10 years agoMerge "Fix several bad function definitions."
Narayan Kamath [Thu, 20 Feb 2014 12:55:27 +0000 (12:55 +0000)]
Merge "Fix several bad function definitions."

10 years agoMerge "AArch64: Make AssetAtlasService 64-bit compatible"
Narayan Kamath [Thu, 20 Feb 2014 12:55:20 +0000 (12:55 +0000)]
Merge "AArch64: Make AssetAtlasService 64-bit compatible"

10 years agoFix several bad function definitions.
Narayan Kamath [Thu, 6 Feb 2014 11:50:31 +0000 (11:50 +0000)]
Fix several bad function definitions.

We claim these functions want jlong as input (8 bytes wide)
but the definitions use pointer types or jints
(4 bytes wide for 32 bit).

bug: 12890271

(cherry picked from 017546e65280a3389b6167f5fc3b5d1c3001154a)

Change-Id: Iede02be90e6a3c9a2db1fe04dcd0db99abe37097

10 years agoAArch64: Make AssetAtlasService 64-bit compatible
Ashok Bhat [Mon, 27 Jan 2014 16:00:23 +0000 (16:00 +0000)]
AArch64: Make AssetAtlasService 64-bit compatible

Changes in this patch include

[x] Long(64-bit) is used to store native pointers in
    AssetAtlasService and related classes as they can be 64-bit.

[x] Some minor changes have been done to conform with
    standard JNI practice (e.g. use of jint instead of int
    in JNI function prototypes)

Change-Id: Ib4c77c134e3ad5b21732e20cde9a54a0b16bdab1
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
10 years agoMerge "libandroidfw: fix build error for 64-bit"
Narayan Kamath [Thu, 20 Feb 2014 11:50:59 +0000 (11:50 +0000)]
Merge "libandroidfw: fix build error for 64-bit"

10 years agolibandroidfw: fix build error for 64-bit
Patrik Bannura [Wed, 12 Feb 2014 12:38:54 +0000 (13:38 +0100)]
libandroidfw: fix build error for 64-bit

Fixed casting int to pointer in logging code.
Previously ints where printed as pointers and
the %p format specifier was used for convenience
to get hex output. This change uses %x and also
gets rid of the casts.

Change-Id: Ia2539769d245f50c0f3884119f80682defe69d61

10 years agoMerge "A background started service is removed from mStartingBackground when timeout...
Craig Mautner [Wed, 19 Feb 2014 15:36:00 +0000 (15:36 +0000)]
Merge "A background started service is removed from mStartingBackground when timeout. Fix is to make sure mStartingBackground is updated to remove one."

10 years agoMerge "Remove java.io.tmpdir assignment in AndroidRuntime."
Narayan Kamath [Wed, 19 Feb 2014 14:24:47 +0000 (14:24 +0000)]
Merge "Remove java.io.tmpdir assignment in AndroidRuntime."

10 years agoMerge "Fixed a possible infinite loop in JsonReader#skipValue()"
Calin Juravle [Wed, 19 Feb 2014 11:15:00 +0000 (11:15 +0000)]
Merge "Fixed a possible infinite loop in JsonReader#skipValue()"

10 years agoA background started service is removed from mStartingBackground when timeout.
Junu Kim [Wed, 19 Feb 2014 07:25:21 +0000 (16:25 +0900)]
A background started service is removed from mStartingBackground when timeout.
Fix is to make sure mStartingBackground is updated to remove one.

Change-Id: I0e42beb550d33e6e400349b85bbb89848e18d520

10 years agoFixed a possible infinite loop in JsonReader#skipValue()
Calin Juravle [Tue, 18 Feb 2014 19:08:39 +0000 (19:08 +0000)]
Fixed a possible infinite loop in JsonReader#skipValue()

Bug: https://code.google.com/p/android/issues/detail?id=65846

Change-Id: Idaa0e75fc5612550c4b9ccb9afa5535ea5b189df

10 years agoMerge "Structure device not freed in error case"
Jeff Sharkey [Tue, 18 Feb 2014 17:12:14 +0000 (17:12 +0000)]
Merge "Structure device not freed in error case"

10 years agoMerge "Tethering: Skip link local addresses when enabling NAT"
Jeff Sharkey [Tue, 18 Feb 2014 17:11:09 +0000 (17:11 +0000)]
Merge "Tethering: Skip link local addresses when enabling NAT"

10 years agoMerge "Fix, MountService now only sends one onShutDownComplete"
Jeff Sharkey [Tue, 18 Feb 2014 17:10:02 +0000 (17:10 +0000)]
Merge "Fix, MountService now only sends one onShutDownComplete"

10 years agoMerge "Move Json{Reader,Writer} test to cts."
Narayan Kamath [Mon, 17 Feb 2014 18:20:33 +0000 (18:20 +0000)]
Merge "Move Json{Reader,Writer} test to cts."

10 years agoMove Json{Reader,Writer} test to cts.
Narayan Kamath [Mon, 17 Feb 2014 17:51:43 +0000 (17:51 +0000)]
Move Json{Reader,Writer} test to cts.

These tests aren't really run automatically (or perhaps
even manually) in their present location, and increasing
CTS coverage is never a bad thing.

Change-Id: I81cce1211021011dca551fde62e725e177ca4df0

10 years agoMerge "Extended locales in AAPT / AssetManager."
Narayan Kamath [Mon, 17 Feb 2014 11:00:29 +0000 (11:00 +0000)]
Merge "Extended locales in AAPT / AssetManager."

10 years agoMerge "AssetManager support for 3 letter lang/country codes."
Narayan Kamath [Mon, 17 Feb 2014 11:00:14 +0000 (11:00 +0000)]
Merge "AssetManager support for 3 letter lang/country codes."

10 years agoMerge "Fix destruction issues relating to AllocationAdapter."
Tim Murray [Fri, 14 Feb 2014 23:50:08 +0000 (23:50 +0000)]
Merge "Fix destruction issues relating to AllocationAdapter."

10 years agoFix destruction issues relating to AllocationAdapter.
Tim Murray [Wed, 12 Feb 2014 19:16:17 +0000 (11:16 -0800)]
Fix destruction issues relating to AllocationAdapter.

bug 12971201

Change-Id: I3d9f66f527a35837ac866a695bdcc41d908a2562

10 years agoMerge "Check the return value of listFiles on the ifw directory"
Ben Gruver [Fri, 14 Feb 2014 21:15:49 +0000 (21:15 +0000)]
Merge "Check the return value of listFiles on the ifw directory"

10 years agoCheck the return value of listFiles on the ifw directory
Ben Gruver [Tue, 1 Oct 2013 00:03:34 +0000 (17:03 -0700)]
Check the return value of listFiles on the ifw directory

listFiles can return null if the directory doesn't exist

Bug: 10497143
Change-Id: Ie4cbdf18a259a9641669a4e6b9bf49b4879acc67

10 years agoMerge "Add hidden minor version ID for support lib workarounds."
Tim Murray [Fri, 14 Feb 2014 20:47:33 +0000 (20:47 +0000)]
Merge "Add hidden minor version ID for support lib workarounds."

10 years agoRemove java.io.tmpdir assignment in AndroidRuntime.
Narayan Kamath [Fri, 14 Feb 2014 17:14:09 +0000 (17:14 +0000)]
Remove java.io.tmpdir assignment in AndroidRuntime.

This value is overridden by the framework anyway
(see ActivityThread.handleBindApplication). Besides,
it doesn't seem like a great idea to let tools clutter
/sdcard/ with temp files.

Change-Id: I26c710cbea7397f89e6103e54a73044a87da95b1

10 years agoExtended locales in AAPT / AssetManager.
Narayan Kamath [Tue, 21 Jan 2014 15:32:36 +0000 (15:32 +0000)]
Extended locales in AAPT / AssetManager.

Support 3 letter language codes, script codes &
variants. The bulk of the changes are related to
the implementation of command line filtering of
locales etc. The previous code assumed that the
value of each "axis" (locale, density, size etc.)
could be represented by a 4 byte type. This is
no longer the case.

This change introduces a new class, AaptLocaleValue
which holds a (normalized) locale parsed from a
directory name or a filter string. This class takes
responsibility for parsing locales as well as
writing them to ResTable_config structures, which is
their representation in the resource table.

This includes minor changes at the java / JNI level
for AssetManager. We now call locale.toLanguageTag()
to give the native layer a well formed BCP-47 tag.
I've removed some duplicated parsing code in
AssetManager.cpp and replaced them with functions on
ResTable_config. The native getLocales function has
been changed to return well formed BCP-47 locales as
well, so that the corresponding java function can use
Locale.forLanguageTag to construct a Locale object
out of it.

Finally, this change introduces default and copy
constructors for ResTable_config to prevent having
to memset() the associated memory to 0 on every
stack allocation.

(cherry-picked from commit 91447d88f2bdf9c2bf8d1a53570efef6172fba74)

Change-Id: I1b43086860661012f949fb8e5deb7df44519b854

10 years agoAssetManager support for 3 letter lang/country codes.
Narayan Kamath [Mon, 20 Jan 2014 13:57:11 +0000 (13:57 +0000)]
AssetManager support for 3 letter lang/country codes.

- 3 letter codes are packed into the existing 32 bit locale
  field in ResTable_config
- We introduce new fields for script / variant information.

Note that we define a "match" between two ResTable_config
structures to be purely on the basis of their language &
country (disregarding the script and the variant). However,
configs with scripts and variants are considered to be more
specific than those without.

(cherry picked from commit 378c6775a62d9c461cde51f06c1b14bb014c78fd)

Change-Id: I7dce82a3fe2412834252723f458826ae41535a78

10 years agoMerge "AArch64: Use long for pointers in SurfaceSession class"
Narayan Kamath [Fri, 14 Feb 2014 13:10:43 +0000 (13:10 +0000)]
Merge "AArch64: Use long for pointers in SurfaceSession class"

10 years agoMerge "frameworks/base: move idmap from frameworks/native"
Dianne Hackborn [Fri, 14 Feb 2014 01:45:12 +0000 (01:45 +0000)]
Merge "frameworks/base: move idmap from frameworks/native"

10 years agoMerge "Telephony: Handle DATA_CALL_LIST in DCC."
Wink Saville [Thu, 13 Feb 2014 23:37:03 +0000 (23:37 +0000)]
Merge "Telephony: Handle DATA_CALL_LIST in DCC."

10 years agoMerge "Telephony API extension v2"
Wink Saville [Thu, 13 Feb 2014 04:26:33 +0000 (04:26 +0000)]
Merge "Telephony API extension v2"

10 years agoMerge "Convert all selinux_android_restorecon and _setfilecon calls to new API."
Nick Kralevich [Wed, 12 Feb 2014 23:09:37 +0000 (23:09 +0000)]
Merge "Convert all selinux_android_restorecon and _setfilecon calls to new API."

10 years agoMerge "Finish fixing Zygote descriptor leakage problem"
Dave Platt [Wed, 12 Feb 2014 23:03:48 +0000 (23:03 +0000)]
Merge "Finish fixing Zygote descriptor leakage problem"

10 years agoMerge "Fix build."
Narayan Kamath [Wed, 12 Feb 2014 17:26:57 +0000 (17:26 +0000)]
Merge "Fix build."

10 years agoMerge "Improve Parcel's handling of non-primitive arrays"
Paul Duffin [Wed, 12 Feb 2014 10:27:13 +0000 (10:27 +0000)]
Merge "Improve Parcel's handling of non-primitive arrays"

10 years agoFix build.
Dianne Hackborn [Tue, 11 Feb 2014 21:56:21 +0000 (13:56 -0800)]
Fix build.

At least part of what is broken.  Other stuff still seems to be.

(cherry picked from 32bb5fae353b5bb6275e75952e89c514c7369cee)

Change-Id: I280376645166fe772ff7a31ff2e8494dce2adbf8

10 years agoMerge "Use size_t* in AssetManager::createIdmap"
Narayan Kamath [Wed, 12 Feb 2014 09:32:16 +0000 (09:32 +0000)]
Merge "Use size_t* in AssetManager::createIdmap"

10 years agoframeworks/base: move idmap from frameworks/native
Colin Cross [Wed, 12 Feb 2014 02:04:44 +0000 (18:04 -0800)]
frameworks/base: move idmap from frameworks/native

idmap depends on libandroidfw, so it should go in frameworks/base.

Change-Id: I3c1db3baa355f53d6b523d60f4377e63eff00c30

10 years agoUse size_t* in AssetManager::createIdmap
Colin Cross [Wed, 12 Feb 2014 02:02:06 +0000 (18:02 -0800)]
Use size_t* in AssetManager::createIdmap

ResTable::createIdmap takes a size_t*, and the idmap command is
passing in a size_t*, make AssetManager::createIdmap take a
size_t*.  Should fix the mac build.

Change-Id: Idc16dedfe2aa7367c75f89a937a8242d494e8f8e

10 years agoMerge "Runtime resource overlay, iteration 2, test cases"
Dianne Hackborn [Tue, 11 Feb 2014 21:30:08 +0000 (21:30 +0000)]
Merge "Runtime resource overlay, iteration 2, test cases"

10 years agoMerge "Runtime resource overlay, iteration 2"
Dianne Hackborn [Tue, 11 Feb 2014 21:29:58 +0000 (21:29 +0000)]
Merge "Runtime resource overlay, iteration 2"

10 years agoMerge "New command line tool 'idmap'"
Dianne Hackborn [Tue, 11 Feb 2014 20:30:07 +0000 (20:30 +0000)]
Merge "New command line tool 'idmap'"

10 years agoImprove Parcel's handling of non-primitive arrays
Paul Duffin [Mon, 3 Feb 2014 15:02:17 +0000 (15:02 +0000)]
Improve Parcel's handling of non-primitive arrays

Treat arrays other than actual Object arrays (i.e. those whose component type
is actually Object) and the primitive arrays already handled specially by
Parcel as Serializable's.

Issue: 64583
Change-Id: I3ff797f3262e77e4de27f35709bceee5410d1ed4

10 years agoMerge "Use long to store pointers in GLES_JNI/EGL classes"
Narayan Kamath [Tue, 11 Feb 2014 15:39:04 +0000 (15:39 +0000)]
Merge "Use long to store pointers in GLES_JNI/EGL classes"

10 years agoMerge "Fix in MediaMuxer's JNI code"
Narayan Kamath [Tue, 11 Feb 2014 14:33:38 +0000 (14:33 +0000)]
Merge "Fix in MediaMuxer's JNI code"

10 years agoMerge "Use long to store pointers in ConsumerIrService"
Narayan Kamath [Tue, 11 Feb 2014 14:29:09 +0000 (14:29 +0000)]
Merge "Use long to store pointers in ConsumerIrService"

10 years agoMerge "Use long for pointers in EmojiFactory and PdfDocument"
Narayan Kamath [Tue, 11 Feb 2014 14:07:16 +0000 (14:07 +0000)]
Merge "Use long for pointers in EmojiFactory and PdfDocument"

10 years agoMerge "Fix TinyHashMap to use generic hash_type instead of hash_t"
Narayan Kamath [Tue, 11 Feb 2014 13:40:37 +0000 (13:40 +0000)]
Merge "Fix TinyHashMap to use generic hash_type instead of hash_t"

10 years agoMerge "Generic object based data copy."
Tim Murray [Mon, 10 Feb 2014 23:27:05 +0000 (23:27 +0000)]
Merge "Generic object based data copy."

10 years agoGeneric object based data copy.
Jason Sams [Tue, 26 Nov 2013 02:28:33 +0000 (18:28 -0800)]
Generic object based data copy.

This change consolidates the copy paths to Object based.
The runtime now uses reflection to identify the type of
array present.  This adds support for long/double and reduces
the amount of code present.  We could also support arrays of vectors
or objects in the future with this mechanism.

Change-Id: I2297c1c01fbe6a64c375d6368f25d7db781ea788

10 years agoMerge "frameworks/base: fix errors inside ALOGV"
Colin Cross [Mon, 10 Feb 2014 22:02:50 +0000 (22:02 +0000)]
Merge "frameworks/base: fix errors inside ALOGV"

10 years agoFix, MountService now only sends one onShutDownComplete
Rickard Helldin [Fri, 7 Feb 2014 08:35:04 +0000 (09:35 +0100)]
Fix, MountService now only sends one onShutDownComplete

The event onShutDownComplete are sent only when all
volumes are shutdown. (Not one for every volume.)

Change-Id: I6af521ee8285ca581efac7d3c5f013dfbff3ee30

10 years agoMerge "AArch64: Use long for pointers in RS Java/JNI code"
Narayan Kamath [Mon, 10 Feb 2014 13:33:26 +0000 (13:33 +0000)]
Merge "AArch64: Use long for pointers in RS Java/JNI code"

10 years agoframeworks/base: fix errors inside ALOGV
Colin Cross [Fri, 7 Feb 2014 04:17:48 +0000 (20:17 -0800)]
frameworks/base: fix errors inside ALOGV

Fix errors exposed by adding compile-time checking to disabled ALOGVs.

Change-Id: I9beeb5d5df498c63d6bed49734ee80a8e0afcefc

10 years agoMerge "Use canonical path for /vendor/app"
Christopher Tate [Fri, 7 Feb 2014 23:30:23 +0000 (23:30 +0000)]
Merge "Use canonical path for /vendor/app"

10 years agoMerge "Example controller app for system ui demo mode."
John Spurlock [Fri, 7 Feb 2014 20:58:36 +0000 (20:58 +0000)]
Merge "Example controller app for system ui demo mode."

10 years agoExample controller app for system ui demo mode.
John Spurlock [Fri, 7 Feb 2014 19:47:47 +0000 (14:47 -0500)]
Example controller app for system ui demo mode.

Demo mode is protected behind a system setting.  To enable:
$ adb shell settings put global sysui_demo_allowed 1

Change-Id: I2248e27a253bf9eac176e8be9153a2b147cefa40

10 years agoFix TinyHashMap to use generic hash_type instead of hash_t
Ashok Bhat [Fri, 7 Feb 2014 16:31:41 +0000 (16:31 +0000)]
Fix TinyHashMap to use generic hash_type instead of hash_t

TinyHashMap used hash_t(key) to generate hashcode. This
would not work for 64-bit pointers as hash_t is declared as
an uint32_t.

Replaced the hash_t(key) call to more generic android::hash_type(key).
This function is a template function declared in TypeHelpers.h and
has a version available for all data types including pointers.

Change-Id: I612cf18b49ca7c30b63f9d6938df68fed7d80d08
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
10 years agoMerge "Fix a typo."
Jean-Luc Brouillet [Fri, 7 Feb 2014 19:08:36 +0000 (19:08 +0000)]
Merge "Fix a typo."

10 years agoFix a typo.
Jean-Luc Brouillet [Fri, 7 Feb 2014 19:04:26 +0000 (11:04 -0800)]
Fix a typo.

Change-Id: I770fe97ab0635de84d15393aca6bca57552f2f10

10 years agoUse long to store pointers in ConsumerIrService
Ashok Bhat [Fri, 7 Feb 2014 12:26:17 +0000 (12:26 +0000)]
Use long to store pointers in ConsumerIrService

Change-Id: I4e6562b7dc09f87ad80b1e862b614a370d37feb5
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
10 years agoFix in MediaMuxer's JNI code
Ashok Bhat [Fri, 7 Feb 2014 12:22:50 +0000 (12:22 +0000)]
Fix in MediaMuxer's JNI code

android_media_MediaMuxer_native_setup should return jlong. It was
incorrectly returning jint.

Change-Id: I31a475af0818f4d5dcbb39380b42612188cf3853
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
10 years agoMerge "Make sure to turn off led after pulse()"
Dan Sandler [Fri, 7 Feb 2014 15:59:08 +0000 (15:59 +0000)]
Merge "Make sure to turn off led after pulse()"

10 years agoConvert all selinux_android_restorecon and _setfilecon calls to new API.
Stephen Smalley [Fri, 7 Feb 2014 14:16:12 +0000 (09:16 -0500)]
Convert all selinux_android_restorecon and _setfilecon calls to new API.

libselinux selinux_android_restorecon API is changing to the more
general interface with flags and dropping the older variants.

Also get rid of the old, no longer used selinux_android_setfilecon API
and rename selinux_android_setfilecon2 to it as it is the only API in use.

Change-Id: I1e71ec398ccdc24cac4ec76f1b858d0f680f4925
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
10 years agoAdd hidden minor version ID for support lib workarounds.
Tim Murray [Fri, 7 Feb 2014 00:39:38 +0000 (16:39 -0800)]
Add hidden minor version ID for support lib workarounds.

Change-Id: I7940e32c78caaab351fb28d92e5fadbea6292b27

10 years agoFinish fixing Zygote descriptor leakage problem
Dave Platt [Thu, 6 Feb 2014 01:06:42 +0000 (17:06 -0800)]
Finish fixing Zygote descriptor leakage problem

In order to prevent Zygote descriptors from leaking into the child
environment, they should be closed by the forked-off child process
before the child switches to the application UID.  These changes close
the descriptors via dup2(), substituting a descriptor open to
/dev/null in their place; this allows the Zygote Java code to close
the FileDescriptor objects cleanly.

This is a multi-project change: dalvik, art, libcore, frameworks/base,
and external/sepolicy are affected.  The CLs need to be approved
together, lest the build break or the software fail to boot.

Round 2: indent change

Bug: 12114500
Change-Id: I090402136a8a8b7d6aad6eb153026e85d7cf6ad3

10 years agoAArch64: Use long for pointers in SurfaceSession class
Ashok Bhat [Wed, 8 Jan 2014 15:34:15 +0000 (15:34 +0000)]
AArch64: Use long for pointers in SurfaceSession class

For storing pointers, long is used in SurfaceSession class,
as native pointers can be 64-bit.

Change-Id: I1535a488dfffaddd20e3beac57e558035d715cbf
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
10 years agoMerge "Fix for positioning of glyphs within a bitmap"
Chris Craik [Wed, 5 Feb 2014 21:36:42 +0000 (21:36 +0000)]
Merge "Fix for positioning of glyphs within a bitmap"

10 years agoAArch64: Use long for pointers in RS Java/JNI code
Ashok Bhat [Tue, 4 Feb 2014 14:57:58 +0000 (14:57 +0000)]
AArch64: Use long for pointers in RS Java/JNI code

Changes include
[x] Some JNI functions, with return type jlong, casts
    pointer to jint before returning it. This has been fixed.

[x] Minor JNI function prototype changes where
    formal paramter type has been changed to a JNI
    type (int to jint for example).

[x] long is used for ScriptC, Sampler, Font, ProgramStore
    handles as they can be 64-bit.

[x] A new hidden constructor ScriptC(long, RenderScript)
    has been added. This should eventually replace public
    API method ScriptC(int, RenderScript).

[x] Font and FileA3D use getNativeAsset instead of getAssetInt
    to get Asset Handles. The getAssetInt method will be
    deprecated in favor of getNativeAsset, as the former does
    not support 64-bit.

[x] rsnPathCreate method accepts loop as an int. This should
    be long as the underlying RS function assumes this to be
    a pointer.

Change-Id: I919d857e5933febe63966049da83de9f9adee6f5
Signed-off-by: Ashok Bhat <ashok.bhat@arm.com>
10 years agoMerge "Add message to RS docs about application vs. library packaging."
Stephen Hines [Wed, 5 Feb 2014 01:52:31 +0000 (01:52 +0000)]
Merge "Add message to RS docs about application vs. library packaging."

10 years agoMerge "DO NOT MERGE: Mark new util type functions as @hide in AOSP."
Tim Murray [Mon, 3 Feb 2014 22:37:12 +0000 (22:37 +0000)]
Merge "DO NOT MERGE: Mark new util type functions as @hide in AOSP."

10 years agoMerge "Check that bound allocations are 1D"
Tim Murray [Mon, 3 Feb 2014 22:37:11 +0000 (22:37 +0000)]
Merge "Check that bound allocations are 1D"

10 years agoMerge "Fix debug context."
Tim Murray [Mon, 3 Feb 2014 22:37:06 +0000 (22:37 +0000)]
Merge "Fix debug context."

10 years agoMerge "Enable asynchronous destruction of BaseObjs."
Tim Murray [Mon, 3 Feb 2014 22:37:00 +0000 (22:37 +0000)]
Merge "Enable asynchronous destruction of BaseObjs."

10 years agoMerge "Fix most logging macros for context pointers."
Tim Murray [Mon, 3 Feb 2014 22:36:54 +0000 (22:36 +0000)]
Merge "Fix most logging macros for context pointers."

10 years agoMerge "Fix getBytesSize for YUV allocations."
Tim Murray [Mon, 3 Feb 2014 22:36:53 +0000 (22:36 +0000)]
Merge "Fix getBytesSize for YUV allocations."

10 years agoMerge "Move RenderScript from graphics/ to new fw/base subdirectory rs."
Tim Murray [Mon, 3 Feb 2014 22:36:50 +0000 (22:36 +0000)]
Merge "Move RenderScript from graphics/ to new fw/base subdirectory rs."

10 years agoMerge "Convert Java/JNI to 64-bit, part 2."
Tim Murray [Mon, 3 Feb 2014 22:36:45 +0000 (22:36 +0000)]
Merge "Convert Java/JNI to 64-bit, part 2."

10 years agoMerge changes If93a5d78,I1957f7ac
Tim Murray [Mon, 3 Feb 2014 22:36:37 +0000 (22:36 +0000)]
Merge changes If93a5d78,I1957f7ac

* changes:
  Convert Java/JNI to 64-bit, part 1.
  Add long/double read support.

10 years agoMerge "Use Object + type in place of specific array"
Tim Murray [Mon, 3 Feb 2014 22:36:36 +0000 (22:36 +0000)]
Merge "Use Object + type in place of specific array"