OSDN Git Service

android-x86/frameworks-native.git
10 years agoInputDispatcher: Optimize count()
Jon McCaffrey [Tue, 18 Nov 2014 20:07:08 +0000 (12:07 -0800)]
InputDispatcher: Optimize count()

Count items as they are added and removed rather than iterating over the
entire list to count them.

Increases performance slightly, particularly when tracing is turned on,
which causes count to be queried more often, and when the number of
items in the queues grow large.  This can happen due to applications not
responding, for example.

Change-Id: I0f11f7edd46089612af910cdfabfeb3ee685d7d9

10 years agoMerge "Frameworks/native: Add pragmas to glgen headers"
Andreas Gampe [Wed, 12 Nov 2014 18:30:40 +0000 (18:30 +0000)]
Merge "Frameworks/native: Add pragmas to glgen headers"

10 years agoFrameworks/native: Add pragmas to glgen headers
Andreas Gampe [Tue, 11 Nov 2014 06:51:01 +0000 (22:51 -0800)]
Frameworks/native: Add pragmas to glgen headers

The code currently being generated by glgen is not the cleanest, but
we would like to get core/jni onto -Werror. Thus add pragmas turning
off warnings locally to the C headers.

Also fix signed-comparison in two functions.

Longer term TODO: Emit clean code.

Change-Id: Iee8582f8c0c1de076d64851d3b6ca467afd5bc43

10 years agoMerge "EGL: make x86/64 wrappers resistant to -fno-omit-frame-pointer flag"
Jesse Hall [Wed, 12 Nov 2014 17:21:26 +0000 (17:21 +0000)]
Merge "EGL: make x86/64 wrappers resistant to -fno-omit-frame-pointer flag"

10 years agoEGL: make x86/64 wrappers resistant to -fno-omit-frame-pointer flag
Michal Wajdeczko [Fri, 8 Aug 2014 15:26:25 +0000 (17:26 +0200)]
EGL: make x86/64 wrappers resistant to -fno-omit-frame-pointer flag

Wrapper functions were broken on builds that forced -fno-omit-frame-pointer flag.

Change-Id: I5a80f9587fb3db821b4156af56acda59a0b4579b
Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
Signed-off-by: Yong Yao <yong.yao@intel.com>
10 years agoMerge "Add missing include to fix libc++ build."
Dan Albert [Fri, 7 Nov 2014 18:04:49 +0000 (18:04 +0000)]
Merge "Add missing include to fix libc++ build."

10 years agoAdd missing include to fix libc++ build.
Dan Albert [Fri, 7 Nov 2014 17:33:44 +0000 (09:33 -0800)]
Add missing include to fix libc++ build.

This was previously committed, but I guess it got lost in the merge
queue.

Bug: 15193147
Change-Id: Ic0e5d90ab403fb833d6a3bb7c55d279c56534869

10 years agoam d12586b8: Merge "Move hwc tests to libc++."
Dan Albert [Wed, 5 Nov 2014 02:41:59 +0000 (02:41 +0000)]
am d12586b8: Merge "Move hwc tests to libc++."

* commit 'd12586b8593cf0cb894d02104a9794ed9dd85d16':
  Move hwc tests to libc++.

10 years agoMerge "Move hwc tests to libc++."
Dan Albert [Wed, 5 Nov 2014 01:38:09 +0000 (01:38 +0000)]
Merge "Move hwc tests to libc++."

10 years agoMove hwc tests to libc++.
Dan Albert [Wed, 5 Nov 2014 01:25:12 +0000 (17:25 -0800)]
Move hwc tests to libc++.

Since these tests were pulling in a static library linked against
stlport, the tests needed to be explicit about which STL they're
using. Might as well be libc++.

Bug: 15193147
Change-Id: I4e0fc4ea10f7982dd04e7baaf57139dd5d55543e

10 years agoam dbab409c: Merge "Avoid two warnings from clang."
Chih-Hung Hsieh [Tue, 4 Nov 2014 01:24:07 +0000 (01:24 +0000)]
am dbab409c: Merge "Avoid two warnings from clang."

* commit 'dbab409c49ca5dd6686912d0b3eeb19ef1c37a3e':
  Avoid two warnings from clang.

10 years agoMerge "Avoid two warnings from clang."
Chih-Hung Hsieh [Tue, 4 Nov 2014 01:14:16 +0000 (01:14 +0000)]
Merge "Avoid two warnings from clang."

10 years agoAvoid two warnings from clang.
Chih-Hung Hsieh [Fri, 24 Oct 2014 21:10:09 +0000 (14:10 -0700)]
Avoid two warnings from clang.

Clang compiler warns about undefined static functions like getReturnString
before all references are removed inside IF_LOG_COMMANDS.
To remove such warnings, we can leave those static functions defined
and the compiler will remove them at the end when they are not used.

Add a cast to long before converting from int to void*,
to avoid type cast warnings on 64 bit targets.

BUG: 17043248
Change-Id: Ica31101522279d2c8eacc90e2939767002fb827b

10 years agoam e42cba8f: Merge "Initialize ARect fields in default ctor"
Dan Stoza [Wed, 29 Oct 2014 23:04:05 +0000 (23:04 +0000)]
am e42cba8f: Merge "Initialize ARect fields in default ctor"

* commit 'e42cba8feb2bbfe385f162a06e41e2c2c0f40c99':
  Initialize ARect fields in default ctor

10 years agoMerge "Initialize ARect fields in default ctor"
Dan Stoza [Wed, 29 Oct 2014 22:58:47 +0000 (22:58 +0000)]
Merge "Initialize ARect fields in default ctor"

10 years agoInitialize ARect fields in default ctor
Dmitriy Ivanov [Tue, 28 Oct 2014 23:41:10 +0000 (16:41 -0700)]
Initialize ARect fields in default ctor

 So that we can use Rect() for update and such.

Bug: 18095940
Change-Id: I03ed87121d2d279f22013b75e95dd7254551ca1f

10 years agoam cb8a9fcc: Merge "Switch to libsigchain.a"
Dmitriy Ivanov [Tue, 21 Oct 2014 03:30:18 +0000 (03:30 +0000)]
am cb8a9fcc: Merge "Switch to libsigchain.a"

* commit 'cb8a9fcc56288378d05d80c55396db7981e3828d':
  Switch to libsigchain.a

10 years agoMerge "Switch to libsigchain.a"
Dmitriy Ivanov [Tue, 21 Oct 2014 03:14:18 +0000 (03:14 +0000)]
Merge "Switch to libsigchain.a"

10 years agoSwitch to libsigchain.a
Dmitriy Ivanov [Mon, 20 Oct 2014 21:17:26 +0000 (14:17 -0700)]
Switch to libsigchain.a

  Use LOCAL_WHOLE_STATIC_LIBRARIES to instruct static linker
  to export symbols from the static library.

Change-Id: I0e41cf173be96c52cdad72b1ef4a1bf819e65dd2

10 years agoam 649bafb7: Merge "bufferqueue: workaround: allow NULL fence with queueBuffer (DO...
Jesse Hall [Mon, 20 Oct 2014 14:13:05 +0000 (14:13 +0000)]
am 649bafb7: Merge "bufferqueue: workaround: allow NULL fence with queueBuffer (DO NOT MERGE)" into lmp-dev

* commit '649bafb7ba2c3204295bc3451ff7af7e28e714ed':
  bufferqueue: workaround: allow NULL fence with queueBuffer (DO NOT MERGE)

10 years agoMerge "bufferqueue: workaround: allow NULL fence with queueBuffer (DO NOT MERGE)...
Jesse Hall [Mon, 20 Oct 2014 13:59:24 +0000 (13:59 +0000)]
Merge "bufferqueue: workaround: allow NULL fence with queueBuffer (DO NOT MERGE)" into lmp-dev

10 years agoam 0b6acb22: Add version number to SensorService dump output.
Aravind Akella [Sun, 19 Oct 2014 21:05:16 +0000 (21:05 +0000)]
am 0b6acb22: Add version number to SensorService dump output.

* commit '0b6acb2231baed861488f2d181621272d0658d25':
  Add version number to SensorService dump output.

10 years agoAdd version number to SensorService dump output.
Aravind Akella [Sat, 18 Oct 2014 23:37:13 +0000 (16:37 -0700)]
Add version number to SensorService dump output.

Change-Id: I64f9482ade523ec3fafe14bff14db7196e32413f

10 years agobufferqueue: workaround: allow NULL fence with queueBuffer (DO NOT MERGE)
Jesse Hall [Sun, 19 Oct 2014 04:47:04 +0000 (21:47 -0700)]
bufferqueue: workaround: allow NULL fence with queueBuffer (DO NOT MERGE)

On one device there is a bug, not yet root-caused, that causes fence
fds to not make it across binder from producer to consumer in the
IGraphicBufferProducer::queueBuffer call. Rather than returning an
error, which the producer typically treats as a fatal error, this
change allows the buffer to be queued with no fence. This avoids an
application crash at the risk of (likely single-frame) visible
corruption.

Bug: 17946343
Change-Id: I9ca89f94098c455e1e90f5f58d5336c936b04a9c

10 years agoam c52e54c4: Merge "Some more test cleanup."
Dan Albert [Fri, 17 Oct 2014 17:53:05 +0000 (17:53 +0000)]
am c52e54c4: Merge "Some more test cleanup."

* commit 'c52e54c47184704613d446eb15d051ceb43d82b3':
  Some more test cleanup.

10 years agoMerge "Some more test cleanup."
Dan Albert [Fri, 17 Oct 2014 17:43:57 +0000 (17:43 +0000)]
Merge "Some more test cleanup."

10 years agoClean up unused parameter warnings.
Narayan Kamath [Fri, 17 Oct 2014 09:37:08 +0000 (10:37 +0100)]
Clean up unused parameter warnings.

Manual application of change 692018430f115343e36b0680acc659f2a731a5d7.

Change-Id: I100f838d70e591a13abb198dc420000ef114a429

10 years agoSome more test cleanup.
Dan Albert [Fri, 17 Oct 2014 03:46:05 +0000 (20:46 -0700)]
Some more test cleanup.

Needed to build successfully when libc++ is the default STL.

Also fix a missing include. Not sure how this was getting by before.
Probably some transitive include.

Change-Id: Ie96b79e96b4e154d786fbf2d5ae5cf5892608837

10 years agoam e9034b45: Merge "Use LOCAL_CLANG := false for GLES wrappers for now"
Chih-hung Hsieh [Thu, 16 Oct 2014 03:06:25 +0000 (03:06 +0000)]
am e9034b45: Merge "Use LOCAL_CLANG := false for GLES wrappers for now"

* commit 'e9034b4502ae6e693ad52fe2e435066dc969a3ab':
  Use LOCAL_CLANG := false for GLES wrappers for now

10 years agoMerge "Use LOCAL_CLANG := false for GLES wrappers for now"
Chih-hung Hsieh [Thu, 16 Oct 2014 02:52:34 +0000 (02:52 +0000)]
Merge "Use LOCAL_CLANG := false for GLES wrappers for now"

10 years agoam 9cc84950: Merge "Don\'t check if this == NULL."
Dan Albert [Wed, 15 Oct 2014 22:52:43 +0000 (22:52 +0000)]
am 9cc84950: Merge "Don\'t check if this == NULL."

* commit '9cc849500df0bba42d3ed19b90c38ff1dffd2262':
  Don't check if this == NULL.

10 years agoMerge "Don't check if this == NULL."
Dan Albert [Wed, 15 Oct 2014 22:45:07 +0000 (22:45 +0000)]
Merge "Don't check if this == NULL."

10 years agoUse LOCAL_CLANG := false for GLES wrappers for now
Bernhard Rosenkraenzer [Mon, 13 Oct 2014 11:56:57 +0000 (13:56 +0200)]
Use LOCAL_CLANG := false for GLES wrappers for now

GLES wrappers built with clang currently result in crashes in libhwui,
UI doesn't come up.

Force gcc for the wrappers until this is investigated and fixed
properly.

Bug: 18003438
Change-Id: I24a17ec019409930883da8040fe2fb05b0fea0c5
Signed-off-by: Bernhard Rosenkraenzer <Bernhard.Rosenkranzer@linaro.org>
10 years agoam 60fd3fee: Migrate CA certificates to all users
Robin Lee [Wed, 15 Oct 2014 19:26:36 +0000 (19:26 +0000)]
am 60fd3fee: Migrate CA certificates to all users

* commit '60fd3feecab4336d964ca8e31c7c3220e1afb558':
  Migrate CA certificates to all users

10 years agoMigrate CA certificates to all users
Robin Lee [Tue, 7 Oct 2014 15:55:02 +0000 (16:55 +0100)]
Migrate CA certificates to all users

Copies the /data/misc/keychain/cacert-* directories to all users on
the device, whereas previously they were simply copied to user 0.

This is a shallow copy so anything that wasn't supposed to be there
will disappear.

Bug: 17811821
Change-Id: Iae5909ab8d5efdb83c9c8fdf0e10ab7060d022cc

10 years agoam 26a28149: Merge "Link sigchain to surfaceflinger main executable"
Dmitriy Ivanov [Tue, 14 Oct 2014 18:21:44 +0000 (18:21 +0000)]
am 26a28149: Merge "Link sigchain to surfaceflinger main executable"

* commit '26a2814963de75f14dfbe7cf3d45e993727891cf':
  Link sigchain to surfaceflinger main executable

10 years agoMerge "Link sigchain to surfaceflinger main executable"
Dmitriy Ivanov [Tue, 14 Oct 2014 17:37:53 +0000 (17:37 +0000)]
Merge "Link sigchain to surfaceflinger main executable"

10 years agoam 3454f123: media: add kMetadataBufferTypeGraphicBuffer
Lajos Molnar [Tue, 14 Oct 2014 08:27:07 +0000 (08:27 +0000)]
am 3454f123: media: add kMetadataBufferTypeGraphicBuffer

* commit '3454f123d0a10bd0ce0760828996aa26c80a8fd4':
  media: add kMetadataBufferTypeGraphicBuffer

10 years agomedia: add kMetadataBufferTypeGraphicBuffer
Lajos Molnar [Tue, 14 Oct 2014 05:56:09 +0000 (22:56 -0700)]
media: add kMetadataBufferTypeGraphicBuffer

Bug: 17935149
Change-Id: I1c26d1e83d8fa0a9ccdb25f6f3b19a86b1dc6f37

10 years agoam ffb49774: Improve ANR diagnostics.
Jeff Brown [Sun, 12 Oct 2014 16:01:36 +0000 (16:01 +0000)]
am ffb49774: Improve ANR diagnostics.

* commit 'ffb497743831ae4857b674629b58ea3c46d01431':
  Improve ANR diagnostics.

10 years agoImprove ANR diagnostics.
Jeff Brown [Sat, 11 Oct 2014 02:01:34 +0000 (19:01 -0700)]
Improve ANR diagnostics.

Print more details about the exact reason that an ANR has occurred.
Also start checking that the window actually has a registered
input connection that is not in a broken state.  These windows
are supposed to be cleaned up by the window manager promptly
as if the app had crashed but the pattern of ANRs we are observing
suggests that broken windows might be sticking around longer than
they should.

Bug: 17721767
Change-Id: Ie2803a3fa9642381ecadc198fec15e1b70d93c20

10 years agoLink sigchain to surfaceflinger main executable
Dmitriy Ivanov [Tue, 7 Oct 2014 20:40:41 +0000 (13:40 -0700)]
Link sigchain to surfaceflinger main executable

 And export necessary symbols to preempt calls from libart.so

Bug: 15345057
Bug: 15426766
Change-Id: If0b7f59a59824e30aa0c33dad76c7a44932180de

10 years agoam 9ea09757: Fix broken error check in Parcel::readBlob
Narayan Kamath [Thu, 9 Oct 2014 12:50:12 +0000 (12:50 +0000)]
am 9ea09757: Fix broken error check in Parcel::readBlob

* commit '9ea09757fa63ad2a9bea9c846b51c1210d69c79f':
  Fix broken error check in Parcel::readBlob

10 years agoFix broken error check in Parcel::readBlob
Narayan Kamath [Wed, 8 Oct 2014 16:35:45 +0000 (17:35 +0100)]
Fix broken error check in Parcel::readBlob

mmap returns MAP_FAILED (which is -1) and not NULL on
failure.

Diagnosed by cferris.

bug: 17909809

Change-Id: I609788ebf94742ef88af002d2d3f3bc9b9e520ac

10 years agoam 7830ef3d: Change ordering of memory allocation and calling Thread::run().
Aravind Akella [Wed, 8 Oct 2014 00:03:48 +0000 (00:03 +0000)]
am 7830ef3d: Change ordering of memory allocation and calling Thread::run().

* commit '7830ef3dd0ff3749d974c2dd85a8fa59dc47aeca':
  Change ordering of memory allocation and calling Thread::run().

10 years agoChange ordering of memory allocation and calling Thread::run().
Aravind Akella [Tue, 7 Oct 2014 21:13:12 +0000 (14:13 -0700)]
Change ordering of memory allocation and calling Thread::run().

In some cases this is causing a crash as device.poll is called with
NULL.

Bug: 17896339
Change-Id: Id431599f2c661338c355c7081b6602f8449a9198

10 years agoam 78639c22: Merge "Parcel: extra validation/debug code for writeDupFileDescriptor...
Jesse Hall [Mon, 6 Oct 2014 22:45:57 +0000 (22:45 +0000)]
am 78639c22: Merge "Parcel: extra validation/debug code for writeDupFileDescriptor" into lmp-dev

* commit '78639c22acfa5e56d87fd6fe8340d083536b29fc':
  Parcel: extra validation/debug code for writeDupFileDescriptor

10 years agoMerge "Parcel: extra validation/debug code for writeDupFileDescriptor" into lmp-dev
Jesse Hall [Mon, 6 Oct 2014 22:36:53 +0000 (22:36 +0000)]
Merge "Parcel: extra validation/debug code for writeDupFileDescriptor" into lmp-dev

10 years agoParcel: extra validation/debug code for writeDupFileDescriptor
Jesse Hall [Mon, 6 Oct 2014 16:49:45 +0000 (09:49 -0700)]
Parcel: extra validation/debug code for writeDupFileDescriptor

Temporary extra debug validation for b/17477219: a Parcel recipient is
getting a positive but invalid fd unexpectedly. Trying to track down
where it's coming from.

Debug code for bug: 17477219
Change-Id: Idb1e71621025a3928c7adc88fd44790e1abd2a01

10 years agoam a7eb4b74: Merge "Fix sockfd leakage in SensorService." into lmp-dev
Aravind Akella [Fri, 3 Oct 2014 21:50:52 +0000 (21:50 +0000)]
am a7eb4b74: Merge "Fix sockfd leakage in SensorService." into lmp-dev

* commit 'a7eb4b74f74423dafd88adf94a7984ae927c3a23':
  Fix sockfd leakage in SensorService.

10 years agoMerge "Fix sockfd leakage in SensorService." into lmp-dev
Aravind Akella [Fri, 3 Oct 2014 21:45:36 +0000 (21:45 +0000)]
Merge "Fix sockfd leakage in SensorService." into lmp-dev

10 years agoam 6747be9f: Generate the SurfaceFlinger shader cache on initialization
Riley Andrews [Fri, 3 Oct 2014 17:26:52 +0000 (17:26 +0000)]
am 6747be9f: Generate the SurfaceFlinger shader cache on initialization

* commit '6747be9fab9981ec24729cede5e0fe779e8b8914':
  Generate the SurfaceFlinger shader cache on initialization

10 years agoGenerate the SurfaceFlinger shader cache on initialization
Riley Andrews [Mon, 29 Sep 2014 20:29:40 +0000 (13:29 -0700)]
Generate the SurfaceFlinger shader cache on initialization

Blobcache is not yet enabled for surfaceflinger (as it should be).
As a temporary workaround, generate all needed shaders during
surfaceflinger initialization instead of doing the compilation
on-demand during ui transitions.

Change-Id: I14455b20a3f85f177d85c9c8b76d8ccc35379b39

10 years agoFix sockfd leakage in SensorService.
Aravind Akella [Mon, 29 Sep 2014 00:52:41 +0000 (17:52 -0700)]
Fix sockfd leakage in SensorService.

i) Call removeFd() only if the fd in the BitTube has been
previously added to the Looper. Use a flag to determine whether the fd
has been previously added or not.
ii) Increment mPendingFlushEventsToSend after holding a connectionLock.
iii) Store the number of acks that are pending in SensorEventQueue
 and send them all at once.

Bug: 17472228
Change-Id: I1ec834fea1112a9cfbd9cddd2198438793698502

10 years agoam f05ef4e5: Merge "Surface: cancel the dequeued buffer when requestBuffer fails...
Jesse Hall [Thu, 2 Oct 2014 23:41:45 +0000 (23:41 +0000)]
am f05ef4e5: Merge "Surface: cancel the dequeued buffer when requestBuffer fails" into lmp-dev

* commit 'f05ef4e542c3041379025e16497ff23710de5dbf':
  Surface: cancel the dequeued buffer when requestBuffer fails

10 years agoMerge "Surface: cancel the dequeued buffer when requestBuffer fails" into lmp-dev
Jesse Hall [Thu, 2 Oct 2014 23:11:08 +0000 (23:11 +0000)]
Merge "Surface: cancel the dequeued buffer when requestBuffer fails" into lmp-dev

10 years agoam 0ba908e7: Merge "add OMX_VIDEO_AVCLevel52 constant" into lmp-dev
Lajos Molnar [Thu, 2 Oct 2014 22:49:19 +0000 (22:49 +0000)]
am 0ba908e7: Merge "add OMX_VIDEO_AVCLevel52 constant" into lmp-dev

* commit '0ba908e7a55ce2641f8886db2a059ccaa3f7a22a':
  add OMX_VIDEO_AVCLevel52 constant

10 years agoMerge "add OMX_VIDEO_AVCLevel52 constant" into lmp-dev
Lajos Molnar [Thu, 2 Oct 2014 22:44:30 +0000 (22:44 +0000)]
Merge "add OMX_VIDEO_AVCLevel52 constant" into lmp-dev

10 years agoam 36462d9d: Merge "Add more logging for dup(fd) failure" into lmp-dev
Michael Lentine [Thu, 2 Oct 2014 19:18:26 +0000 (19:18 +0000)]
am 36462d9d: Merge "Add more logging for dup(fd) failure" into lmp-dev

* commit '36462d9de00105ae46ee3dbb0f644fa6af599f97':
  Add more logging for dup(fd) failure

10 years agoMerge "Add more logging for dup(fd) failure" into lmp-dev
Michael Lentine [Thu, 2 Oct 2014 19:01:09 +0000 (19:01 +0000)]
Merge "Add more logging for dup(fd) failure" into lmp-dev

10 years agoSurface: cancel the dequeued buffer when requestBuffer fails
Jesse Hall [Thu, 2 Oct 2014 18:09:03 +0000 (11:09 -0700)]
Surface: cancel the dequeued buffer when requestBuffer fails

Partial fix for bug: 17477219
Change-Id: Ibf5a9e26e02c4be8854925a77a70f5c9c7dcf6f2

10 years agoAdd more logging for dup(fd) failure
Michael Lentine [Thu, 2 Oct 2014 16:11:04 +0000 (09:11 -0700)]
Add more logging for dup(fd) failure

Bug: 17477219
Change-Id: Ide0ae16d777c9af783023c705c18a93c00999147

10 years agoadd OMX_VIDEO_AVCLevel52 constant
Lajos Molnar [Thu, 2 Oct 2014 04:49:18 +0000 (21:49 -0700)]
add OMX_VIDEO_AVCLevel52 constant

Bug: 17676461
Change-Id: I120041e9b2ffe2a232a2419bcb5fe88cb49961cb

10 years agoam a91968fd: Merge "Accept socket connection before tracing." into lmp-dev
Christopher Ferris [Thu, 2 Oct 2014 01:16:58 +0000 (01:16 +0000)]
am a91968fd: Merge "Accept socket connection before tracing." into lmp-dev

* commit 'a91968fd099378d1673c7443be16adcb97ee40af':
  Accept socket connection before tracing.

10 years agoMerge "Accept socket connection before tracing." into lmp-dev
Christopher Ferris [Thu, 2 Oct 2014 01:11:56 +0000 (01:11 +0000)]
Merge "Accept socket connection before tracing." into lmp-dev

10 years agoAccept socket connection before tracing.
Christopher Ferris [Thu, 2 Oct 2014 00:35:01 +0000 (17:35 -0700)]
Accept socket connection before tracing.

Sometimes dumping threads takes a long time and bugreport times
out. This change will cause us to accept the bugreport socket connection
before dumping threads and should avoid the failed to connect to dumpstate
service problems we've seen.

Bug: 17758374
Change-Id: I80afa0353cf1c340873f481a8d1d7faffff54120

10 years agoam db57cfbd: Merge "When eglMakeCurrent fails we need to fix the egl state." into...
Michael Lentine [Wed, 1 Oct 2014 22:50:32 +0000 (22:50 +0000)]
am db57cfbd: Merge "When eglMakeCurrent fails we need to fix the egl state." into lmp-dev

* commit 'db57cfbd6f9d5795846ef237fd297cb81e429679':
  When eglMakeCurrent fails we need to fix the egl state.

10 years agoMerge "When eglMakeCurrent fails we need to fix the egl state." into lmp-dev
Michael Lentine [Wed, 1 Oct 2014 22:44:45 +0000 (22:44 +0000)]
Merge "When eglMakeCurrent fails we need to fix the egl state." into lmp-dev

10 years agoam dd230cb8: Merge "add a parameter for PCM limiter enable" into lmp-dev
Chong Zhang [Wed, 1 Oct 2014 22:38:37 +0000 (22:38 +0000)]
am dd230cb8: Merge "add a parameter for PCM limiter enable" into lmp-dev

* commit 'dd230cb8d2a43ea453a507b12c612a221a3780e7':
  add a parameter for PCM limiter enable
  Use set_sched_policy to put dexopt operations in SP_BACKGROUND cgroup
  Revert "Modify EGL to disconnect the window when the surface gets destroyed."
  crash if getNativeBuffer() called on NULL GraphicBuffer

10 years agoam 5bf786d8: surfaceflinger: recompute visibility after sideband stream change
Jesse Hall [Wed, 1 Oct 2014 22:38:18 +0000 (22:38 +0000)]
am 5bf786d8: surfaceflinger: recompute visibility after sideband stream change

* commit '5bf786d8f00ff8eee64ebf330dac55e33a0f97b1':
  surfaceflinger: recompute visibility after sideband stream change

10 years agoWhen eglMakeCurrent fails we need to fix the egl state.
Michael Lentine [Wed, 1 Oct 2014 18:17:28 +0000 (11:17 -0700)]
When eglMakeCurrent fails we need to fix the egl state.

Bug: 16676660

Change-Id: Ie7bee9c78378b9e9206060444319e6ee35e1ab74

10 years agoMerge "add a parameter for PCM limiter enable" into lmp-dev
Chong Zhang [Wed, 1 Oct 2014 21:37:00 +0000 (21:37 +0000)]
Merge "add a parameter for PCM limiter enable" into lmp-dev

10 years agoadd a parameter for PCM limiter enable
Chong Zhang [Wed, 1 Oct 2014 04:50:08 +0000 (21:50 -0700)]
add a parameter for PCM limiter enable

Bug: 17470065
Change-Id: I327115fc16e7afe2bf5870e123d63f9887161206

10 years agosurfaceflinger: recompute visibility after sideband stream change
Jesse Hall [Tue, 30 Sep 2014 17:35:11 +0000 (10:35 -0700)]
surfaceflinger: recompute visibility after sideband stream change

We normally recompute layer visibility when a layer gets its first
buffer; before then it's treated as invisible. Sideband layers never
get a buffer (as far as SurfaceFlinger knows), so never became
visible. Now we also recompute visibility when a layer gets a new
sideband stream.

Bug: 17752511
Change-Id: I84e150f196eb2eb7bcd2616248e5e3fa73624809

10 years agoam cde4b13a: Merge "Allow more than one dex2oat extra option"
Narayan Kamath [Wed, 1 Oct 2014 11:15:12 +0000 (11:15 +0000)]
am cde4b13a: Merge "Allow more than one dex2oat extra option"

* commit 'cde4b13a12a0bc8a7bd031f5d5c1b2b2bdd13335':
  Allow more than one dex2oat extra option

10 years agoMerge "Allow more than one dex2oat extra option"
Narayan Kamath [Wed, 1 Oct 2014 11:06:03 +0000 (11:06 +0000)]
Merge "Allow more than one dex2oat extra option"

10 years agoam 5ee2c9d2: Work on issue #17656716: Unhandled exception in Window Manager
Dianne Hackborn [Tue, 30 Sep 2014 20:23:02 +0000 (20:23 +0000)]
am 5ee2c9d2: Work on issue #17656716: Unhandled exception in Window Manager

* commit '5ee2c9d20c0fbecf6b4a482eb8d8bfdf85d0d424':
  Work on issue #17656716: Unhandled exception in Window Manager

10 years agoWork on issue #17656716: Unhandled exception in Window Manager
Dianne Hackborn [Tue, 30 Sep 2014 18:30:03 +0000 (11:30 -0700)]
Work on issue #17656716: Unhandled exception in Window Manager

Try to clean up the code paths coming in and out of binder IPCs to
plug any places where we could disrupt the gather flag of a thread,
causing it to keep gathering stack crawls (which is the thing that
is causing our strict mode data to become so large).

We now take care of saving and restoring this state in the core
IPC code path, not at the Java layer.

Change-Id: I73d564778da127bdce00f304225930e7f2318293

10 years agoam deb71b28: Merge "Fix a possible SensorService deadlock." into lmp-dev
Aravind Akella [Mon, 29 Sep 2014 00:56:17 +0000 (00:56 +0000)]
am deb71b28: Merge "Fix a possible SensorService deadlock." into lmp-dev

* commit 'deb71b2812702318900c36b7bcfa9759525ea7cc':
  Fix a possible SensorService deadlock.

10 years agoam bacc28ef: surfaceflinger: Use landscape resolution for default dpi
Jesse Hall [Mon, 29 Sep 2014 00:56:04 +0000 (00:56 +0000)]
am bacc28ef: surfaceflinger: Use landscape resolution for default dpi

* commit 'bacc28ef1df329f4dc21bae44b09a6c5018af908':
  surfaceflinger: Use landscape resolution for default dpi

10 years agoMerge "Fix a possible SensorService deadlock." into lmp-dev
Aravind Akella [Mon, 29 Sep 2014 00:49:24 +0000 (00:49 +0000)]
Merge "Fix a possible SensorService deadlock." into lmp-dev

10 years agosurfaceflinger: Use landscape resolution for default dpi
Jesse Hall [Sun, 28 Sep 2014 19:14:12 +0000 (12:14 -0700)]
surfaceflinger: Use landscape resolution for default dpi

When HWC doesn't provide DPI values for a display, we pick a default
DPI based on resolution. The intent was that 1080p and higher displays
would get XHIGH density, and lower resolutions would get TV density.
In KK (and possibly forever) we had a bug that we'd always use TV
density. That was fixed in L, but that fix exposed a pre-existing bug
that we always used the display's height in its native orientation,
rather than in landscape orientation. So an 800x1280 tablet like N7v1
started getting XHIGH density instead of the intended TV density.

Bug: 17461633
Change-Id: Ia57fa49e61f36bdda63ce283ef62c9953297222c

10 years agoFix a possible SensorService deadlock.
Aravind Akella [Thu, 25 Sep 2014 05:12:58 +0000 (22:12 -0700)]
Fix a possible SensorService deadlock.

 If the destructor of SensorEventConnection gets called when the main
 sendEvents loop of SensorService is executing it may result in a deadlock.
 The loop promotes each connection to a strong_pointer, calls sendEvents
 and cleans up the connection if necessary. It is possible that the sp's
 destructor may delete SensorEventConnection which will call the dtor
 ~SensorEventConnection(). This dtor again needs SensorService mLock to
 execute which may result in a deadlock.

Bug: 17617897
Change-Id: I76c244dbe85fadb591c0bd1a9a5eb01d93f56505

10 years agoam 71e351d9: Merge "Revert "Fix sockfd leakage in SensorService. Remove sockfd from...
Aravind Akella [Sat, 27 Sep 2014 02:46:32 +0000 (02:46 +0000)]
am 71e351d9: Merge "Revert "Fix sockfd leakage in SensorService. Remove sockfd from Looper  when the connection is removed from mActiveConnections."" into lmp-dev

* commit '71e351d96b551ccdbc39b52a0c66da86cae83701':
  Revert "Fix sockfd leakage in SensorService. Remove sockfd from Looper  when the connection is removed from mActiveConnections."

10 years agoMerge "Revert "Fix sockfd leakage in SensorService. Remove sockfd from Looper when...
Aravind Akella [Sat, 27 Sep 2014 02:09:29 +0000 (02:09 +0000)]
Merge "Revert "Fix sockfd leakage in SensorService. Remove sockfd from Looper  when the connection is removed from mActiveConnections."" into lmp-dev

10 years agoRevert "Fix sockfd leakage in SensorService. Remove sockfd from Looper when the...
Aravind Akella [Sat, 27 Sep 2014 02:06:32 +0000 (02:06 +0000)]
Revert "Fix sockfd leakage in SensorService. Remove sockfd from Looper  when the connection is removed from mActiveConnections."

This reverts commit 5105960d49bfa86efeeea8f87c87aa96b5771c37.

Change-Id: I4d39ded2e50ef8a8e0aac27e438876203385a1f0

10 years agoam 4b20d402: Merge "Only mark repeated keys as PASS_TO_USER if the original key was...
Michael Wright [Fri, 26 Sep 2014 01:47:01 +0000 (01:47 +0000)]
am 4b20d402: Merge "Only mark repeated keys as PASS_TO_USER if the original key was." into lmp-dev

* commit '4b20d40218a25adcad21cc135a3103b4e04940f3':
  Only mark repeated keys as PASS_TO_USER if the original key was.

10 years agoam 728a8da1: Merge "surfaceflinger: Check sideband stream changed to trigger sideband...
Jesse Hall [Fri, 26 Sep 2014 01:47:00 +0000 (01:47 +0000)]
am 728a8da1: Merge "surfaceflinger: Check sideband stream changed to trigger sideband Layer update." into lmp-dev

* commit '728a8da14638e718ede11a7a894705524fc9d171':
  surfaceflinger: Check sideband stream changed to trigger sideband Layer update.

10 years agoMerge "Only mark repeated keys as PASS_TO_USER if the original key was." into lmp-dev
Michael Wright [Fri, 26 Sep 2014 01:38:56 +0000 (01:38 +0000)]
Merge "Only mark repeated keys as PASS_TO_USER if the original key was." into lmp-dev

10 years agoOnly mark repeated keys as PASS_TO_USER if the original key was.
Michael Wright [Wed, 24 Sep 2014 20:26:59 +0000 (13:26 -0700)]
Only mark repeated keys as PASS_TO_USER if the original key was.

Bug: 17629695
Change-Id: Ic569aba4149c6cc321c2cf8aede7c5e8a1196234

10 years agoMerge "surfaceflinger: Check sideband stream changed to trigger sideband Layer update...
Jesse Hall [Wed, 24 Sep 2014 22:07:01 +0000 (22:07 +0000)]
Merge "surfaceflinger: Check sideband stream changed to trigger sideband Layer update." into lmp-dev

10 years agoam 5105960d: Fix sockfd leakage in SensorService. Remove sockfd from Looper when...
Aravind Akella [Wed, 24 Sep 2014 03:48:09 +0000 (03:48 +0000)]
am 5105960d: Fix sockfd leakage in SensorService. Remove sockfd from Looper when the connection is removed from mActiveConnections.

* commit '5105960d49bfa86efeeea8f87c87aa96b5771c37':
  Fix sockfd leakage in SensorService. Remove sockfd from Looper when the connection is removed from mActiveConnections.

10 years agosurfaceflinger: Check sideband stream changed to trigger sideband Layer update.
chenhg [Tue, 23 Sep 2014 02:52:25 +0000 (10:52 +0800)]
surfaceflinger: Check sideband stream changed to trigger sideband Layer update.

BUG: 17382907
Change-Id: I65fd037bd25238ae68a3976468ae7fd7ce49c562

10 years agoFix sockfd leakage in SensorService. Remove sockfd from Looper
Aravind Akella [Tue, 23 Sep 2014 23:42:49 +0000 (16:42 -0700)]
Fix sockfd leakage in SensorService. Remove sockfd from Looper
when the connection is removed from mActiveConnections.

Bug: 17472228
Change-Id: Iaabe0968bd2104ef961ff5e8659bf91712e2e1db

10 years agoam 12c4bdad: Add rate-limited logging for unimplemented es functions.
Michael Lentine [Tue, 23 Sep 2014 19:16:54 +0000 (19:16 +0000)]
am 12c4bdad: Add rate-limited logging for unimplemented es functions.

* commit '12c4bdad7b2372ab64dcd2abbdbe47e709fb8b90':
  Add rate-limited logging for unimplemented es functions.

10 years agoam 6ee97e74: Merge "Fix a null dereference."
Dan Albert [Tue, 23 Sep 2014 17:53:39 +0000 (17:53 +0000)]
am 6ee97e74: Merge "Fix a null dereference."

* commit '6ee97e74d2c972dec2aa6a2f231b718eae54898f':
  Fix a null dereference.

10 years agoMerge "Fix a null dereference."
Dan Albert [Tue, 23 Sep 2014 17:41:11 +0000 (17:41 +0000)]
Merge "Fix a null dereference."

10 years agoFix a null dereference.
Dan Albert [Tue, 23 Sep 2014 16:51:51 +0000 (09:51 -0700)]
Fix a null dereference.

Change-Id: I7224c3368ff414620062bc08b80a6a94d55d7931

10 years agoAdd rate-limited logging for unimplemented es functions.
Michael Lentine [Thu, 11 Sep 2014 19:24:13 +0000 (12:24 -0700)]
Add rate-limited logging for unimplemented es functions.

Bug: 17322013

Change-Id: I7aa5ab2f76cf344cfb75802edbfd44e8e7526ff1

10 years agomerge in lmp-release history after reset to d94854030c76fc4ee4126ca1707a76fbd1f29126
The Android Automerger [Tue, 23 Sep 2014 13:16:24 +0000 (06:16 -0700)]
merge in lmp-release history after reset to d94854030c76fc4ee4126ca1707a76fbd1f29126