OSDN Git Service

android-x86/frameworks-native.git
10 years agoadd HEVC decoder and cleanup jb-x86
Michael Chen [Fri, 11 Oct 2013 12:36:37 +0000 (20:36 +0800)]
add HEVC decoder and cleanup

Change-Id: I16342228014320161a22302a3c14df7725556cd2

10 years agoAdd OMX_AUDIO_CodingAPE and OMX_AUDIO_CodingDTS
Chih-Wei Huang [Sun, 25 Aug 2013 15:13:56 +0000 (23:13 +0800)]
Add OMX_AUDIO_CodingAPE and OMX_AUDIO_CodingDTS

Copied from the stagefright-plugins project.

10 years agoopengl: Implement fast tls forwarders for x86
Chih-Wei Huang [Fri, 12 Jul 2013 02:42:55 +0000 (10:42 +0800)]
opengl: Implement fast tls forwarders for x86

Based on android-ia and rebased to JB-MR2.

10 years agoMerge branch 'android-ia' into jb-x86
Chih-Wei Huang [Fri, 2 Aug 2013 04:21:03 +0000 (12:21 +0800)]
Merge branch 'android-ia' into jb-x86

10 years agouse HAL_PIXEL_FORMAT_BGRA_8888 with hwc 1.1
Tapani Pälli [Thu, 16 May 2013 12:03:48 +0000 (15:03 +0300)]
use HAL_PIXEL_FORMAT_BGRA_8888 with hwc 1.1

this is the format supported by our framebuffer

Change-Id: Ic2906f237e1141a970115a44078922c037224040
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Acked-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
10 years agoFix BufferQueue drain timeout
Andy Ross [Thu, 2 May 2013 18:21:20 +0000 (11:21 -0700)]
Fix BufferQueue drain timeout

Commit 3df059d4 ("Fix hang under BufferQueue::disconnect") had a bad
value.  The unit is nanoseconds, not microseconds.  The mistaken 300us
timeout almost worked, but would result in spurious BufferQueue
failures in conditions (e.g. CTS runs) of heavy scheduler load.

Change-Id: I308ddba1f82b5d60be1318d9c37b75a7e8051b14
For: AXIA-2552
Signed-off-by: Andy Ross <andy.ross@windriver.com>
10 years agobinder: Fix service initialization race vs. service manager
Andy Ross [Tue, 19 Feb 2013 22:39:34 +0000 (14:39 -0800)]
binder: Fix service initialization race vs. service manager

The framework relies on the global IServiceManager running before a
BinderService object is instantiated.  But there was no detection of
the error condition when it was not (specifically: mediaserver can be
initialized before system_server), and in fact the default error
behavior (kill the proxy object) ensured that the process would never
be able to add services again.  Retry the addService() code, and allow
transient binder failures for the special case of the service manager
proxy.

Issue: AXIA-1706
Change-Id: Icac10bb0f47a2fe33ac9605a13633b83afa3ebff
Signed-off-by: Andy Ross <andy.ross@windriver.com>
10 years agoFix hang under BufferQueue::disconnect
Andy Ross [Tue, 18 Dec 2012 20:54:32 +0000 (12:54 -0800)]
Fix hang under BufferQueue::disconnect

disconnect() tries to drain the queue, but there are error cases in
the wild where the consumer has "lost" the BufferQueue object and
stopped consuming frames.  That causes a deadlock situation where
disconnect (itself called from under cleanup code,
e.g. eglDestroySurface()) never returns and hangs the process.

Fix this by detecting (with a timeout) the case of a queue that fails
to drain promptly and convert it to an existing error, basically
treating it the same as "abandoned".

Issue: AXIA-582
Change-Id: I313f7fb038e6f61d02a23c3b0d4b6951c54d9f20
Signed-off-by: Andy Ross <andy.ross@windriver.com>
10 years agoAdd EGLContext validation to fix bug ANDROID-1679
yi.xie@intel.com [Tue, 21 Feb 2012 02:11:08 +0000 (10:11 +0800)]
Add EGLContext validation to fix bug ANDROID-1679

Segmentation fault will occur when using an invalid EGLContext,
such as getting an EGLContext by "(EGLContext)3".

This validation process first uses msync system call
to check for unmapped memory pointer.

Moreover, Because existing properties(like dpy, config, version, etc.)
are hard to identify if this is an egl_context_t instance,
a magic property is introduced into the class egl_context_t,
which helps to see if it does represent an egl_context_t instance
for most basic accidental cases,
though not feasible to test for deliberate forgeries.

Once the pointer is confirmed valid,
invoke the isValid method of egl_context_t,
which compares magic value to the constant MAGIC to do validation.

Change-Id: If1b884134af57f703b51eb483c4bd0a85c78ec5a
Original-Change-Id: Iebebf4238498c667319c62469baa5d8538127b2d
Signed-off-by: Yi <yi.xie@intel.com>
Signed-off-by: Oliver McFadden <oliver.mcfadden@intel.com>
Signed-off-by: Russell Webb <russell.webb@intel.com>
10 years agoPre-test native window pointer for validity
Paul Drews [Wed, 18 Jan 2012 19:01:49 +0000 (11:01 -0800)]
Pre-test native window pointer for validity

This fixes a segfault in the oglconform tests when running
the test:

  adb shell /data/app/oglconform -v 4 -minFmt -es -test egl-basic negative.drawables.winSurfBadWin

According the the egl specification this should return
EGL_BAD_NATIVE_WINDOW.  The segfault was resulting from the
fact that the window handle is a (cast to opaque) pointer,
and there are not straightforward ways of pre-testing a
pointer for validity.  The fix implemented here is:

  () Take advantage of the msync() system-call logic to test
     pre-test basic pointer validity
  () Once confirmed that we can reference through the
     pointer, look for the "magic" value dropped in by the
     implementation to mark this type at runtime.

Change-Id: Id7d21c822c8a416e45caba25c63d7076d1e547a5
Original-Change-Id: I3c8df8ef5ece5c0fa71c47417766b5453c355f8b
Signed-off-by: Russell Webb <russell.webb@intel.com>
Signed-off-by: Paul Drews <paul.drews@intel.com>
10 years agoremove network location provider from default config
Irina Tirdea [Tue, 17 Jan 2012 14:50:57 +0000 (16:50 +0200)]
remove network location provider from default config

android.hardware.location.network is only available if
Network Location Provider is available. Not all configurations
that include gps have NLP, so this should be removed from
default configuration files. For devices that have NLP
android.hardware.location.network.xml should be installed.

Change-Id: I3a25c5687dbae62972fac1bd789a67c81e00d8b8
Original-Change-Id: Iab06f0fff3f109eee24186548f2624e0a97f2104
Signed-off-by: Irina Tirdea <irina.tirdea@intel.com>
10 years agoegl: test: Fix cast to bigger unsigned integer
Pauli Nieminen [Wed, 4 Jan 2012 19:37:42 +0000 (21:37 +0200)]
egl: test: Fix cast to bigger unsigned integer

ASSERT_EQ casts integers to BiggestInt type which makes signed char
result to negative values which don't match with excepted results. Fix
is to use unsigned char which doesn't have the casting problem.

Change-Id: I8e7f91dc81d1fe815d58406e0774f1bfd3c843a7
Original-Change-Id: I44752fff6950ee024e0814150e55bd8351ee423d
Signed-off-by: Pauli Nieminen <pauli.nieminen@linux.intel.com>
10 years agoegl: test: Check that eglGetConfigs checks for NULL parameter
Pauli Nieminen [Wed, 4 Jan 2012 19:18:12 +0000 (21:18 +0200)]
egl: test: Check that eglGetConfigs checks for NULL parameter

Passing NULL as the last paramter of eglGetConfigs is required to
generate EGL_BAD_PARAMETER. Add unit test to check that eglGetConfigs
functions correctly.

Change-Id: I22fb382ca84334237eafbe291e3f943d2045944d
Original-Change-Id: I9c998873c08fe60b2bcf6ae73da5954c79da3dda
Signed-off-by: Pauli Nieminen <pauli.nieminen@linux.intel.com>
Signed-off-by: Russell Webb <russell.webb@intel.com>
10 years agoegl: Add NULL check for eglGetconfigs num_config parameter
Pauli Nieminen [Wed, 4 Jan 2012 14:25:10 +0000 (16:25 +0200)]
egl: Add NULL check for eglGetconfigs num_config parameter

EGL 1.4 specification says in 3.4.1:
"An EGL_BAD_PARAMETER error is generated if num config is NULL."

We have to check for the condition and return the required error to
application.

Change-Id: Ib42709f0420161ce661536394d4c0779bc62be57
Original-Change-Id: I4fe670c4362f4ae494f91d69ae105e29ccf3c769
Signed-off-by: Pauli Nieminen <pauli.nieminen@linux.intel.com>
10 years agoFix build problem when including hal and egl to same c file
Pauli Nieminen [Tue, 15 Nov 2011 18:33:05 +0000 (20:33 +0200)]
Fix build problem when including hal and egl to same c file

If including hal and egl headers to same c file compilation files
because ANativeWindow type is defined in two headers. In HC egl didn't
include system/core/include/system/window.h which hid the type conflict.

Change-Id: Icaf7b29f913f3d7fa8f6d70bb196585d22e7d05c
Original-Change-Id: Icd37329d80fb4e8b9baf9015c9b00e43ebc4f76b
Signed-off-by: Pauli Nieminen <pauli.nieminen@linux.intel.com>
10 years agoegl_extensions: remove force of extension not supported
Daniel Charles [Fri, 13 Jul 2012 23:11:49 +0000 (16:11 -0700)]
egl_extensions: remove force of extension not supported

This problem was introduced by commit 48d438d05f14c.

When the extension GL_EXT_debug_marker is not supported by h/w
the dummy version is used already. Whenever the h/w is ready
to support the extension it will show up and be used by the
OpenGLRenderer.

Change-Id: Iddb58859b187d872a9a3a2986d30f24b8f83e76b
Signed-off-by: Daniel Charles <daniel.charles@intel.com>
10 years agoauto determine the density if not provided in surfaceflinger
Chih-Wei Huang [Tue, 15 Jan 2013 04:55:51 +0000 (12:55 +0800)]
auto determine the density if not provided in surfaceflinger

We hope to support tablet UI for different resolutions.
So adjust the density according to the resolution.

10 years agoenable console management
Chih-Wei Huang [Mon, 7 Jan 2013 08:39:35 +0000 (16:39 +0800)]
enable console management

The console management was removed since ICS. But it's useful for debugging.
Re-implement it based on the original class ConsoleManagerThread of class
DisplayHardwareBase.

10 years agodisables hardware acceleration if debug.egl.hw=0
Chih-Wei Huang [Sun, 6 Jan 2013 08:56:00 +0000 (16:56 +0800)]
disables hardware acceleration if debug.egl.hw=0

10 years agomerge in jb-mr2-zeroday-release history after reset to jb-mr2-dev
The Android Automerger [Mon, 15 Jul 2013 15:19:12 +0000 (08:19 -0700)]
merge in jb-mr2-zeroday-release history after reset to jb-mr2-dev

10 years agofix a dead-lock in sensorservice
Mathias Agopian [Fri, 12 Jul 2013 09:01:16 +0000 (02:01 -0700)]
fix a dead-lock in sensorservice

sensorservice would deadlock if for some reason
a sensor failed to enable.

simplifed the code a bit, and made it behave a little
closer to mr1.1 -- I couldn't convince myself that
some changes in how locks were used were correct.

Bug: 9794362
Change-Id: I6110f5dbb67e543f1c71d127de2299232badb36a

10 years agomerge in jb-mr2-zeroday-release history after reset to jb-mr2-dev
The Android Automerger [Wed, 10 Jul 2013 12:55:26 +0000 (05:55 -0700)]
merge in jb-mr2-zeroday-release history after reset to jb-mr2-dev

10 years agoFind non-extension GLES wrappers in eglGetProcAddress
Jesse Hall [Thu, 4 Jul 2013 19:08:16 +0000 (12:08 -0700)]
Find non-extension GLES wrappers in eglGetProcAddress

This allows apps to find OpenGL ES 3.0 functions using
eglGetProcAddress() instead of dlopen/dlsym.

Bug: 9681677
Change-Id: I7ce6e1636bc47d6b0bf20a4e46bd67235714d129

10 years agomerge in jb-mr2-zeroday-release history after reset to jb-mr2-dev
The Android Automerger [Wed, 12 Jun 2013 12:56:00 +0000 (05:56 -0700)]
merge in jb-mr2-zeroday-release history after reset to jb-mr2-dev

10 years agoPrevent IPCThreadState::joinThreadPool from spinning on unexpected error
Jeff Tinker [Tue, 11 Jun 2013 18:30:21 +0000 (11:30 -0700)]
Prevent IPCThreadState::joinThreadPool from spinning on unexpected error

Adds logging to help determine what is happening to the /dev/binder
fd and dump the process state when it happens.

bug: 8912673
Change-Id: I2aa0c66fc499e91e0bf9ee4ae20404bec35adc82

10 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Sun, 19 May 2013 12:55:29 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

10 years ago[DO NOT MERGE] fix a bug where surfaceflinger and system_server could deadlock
Mathias Agopian [Thu, 16 May 2013 02:14:52 +0000 (19:14 -0700)]
[DO NOT MERGE] fix a bug where surfaceflinger and system_server could deadlock

because surfaceflinger handles screenshot in a different
thread from the binder thread that requested it and because
the IGraphicBufferProducer is a synchronous interface
calling back into the system server; it is possible for
the latter to run out of binder threads (b/c it holds
a lock while calling into SF).

The solution is to make sure all calls on IGraphicBufferProducer
happen on the incoming binder thread. We achieve this by creating
a IGBP wrapper which is given to the screenshot code.

Bug: 8734824
Change-Id: Ife2441c7322e51ecfb20e0df03dacf6bce49578e

10 years agomake the warning timout of Fence::waitForever() implicit and longer
Mathias Agopian [Fri, 17 May 2013 01:03:22 +0000 (18:03 -0700)]
make the warning timout of Fence::waitForever() implicit and longer

- timeout is now 3 seconds instead of 1
- simplifies the API a bit
- allows us to change/tweak this timeout globaly

Bug: 8988871

Change-Id: I8d3c6ec43a372f602fb3f29856710339f86c0ec9

10 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Mon, 13 May 2013 12:55:40 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

10 years agoMerge "Add support for HAL_PIXEL_FORMAT_YCbCr_420_888" into jb-mr2-dev
Eino-Ville Talvala [Thu, 9 May 2013 17:13:22 +0000 (17:13 +0000)]
Merge "Add support for HAL_PIXEL_FORMAT_YCbCr_420_888" into jb-mr2-dev

10 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Thu, 9 May 2013 12:55:26 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

10 years agoMerge "Remove experimental HWC virtual display support" into jb-mr2-dev
Jesse Hall [Wed, 8 May 2013 20:53:42 +0000 (20:53 +0000)]
Merge "Remove experimental HWC virtual display support" into jb-mr2-dev

10 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Wed, 8 May 2013 12:55:32 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

10 years agoMerge "atrace: add the 'dalvik' trace category" into jb-mr2-dev
Jamie Gennis [Wed, 8 May 2013 01:21:17 +0000 (01:21 +0000)]
Merge "atrace: add the 'dalvik' trace category" into jb-mr2-dev

10 years agoatrace: add the 'dalvik' trace category
Jamie Gennis [Tue, 7 May 2013 22:20:39 +0000 (15:20 -0700)]
atrace: add the 'dalvik' trace category

Bug: 8856374
Change-Id: Ie7759328333d21d8123ce9d4f1a177cae5ede51e

10 years agoCorrect native API documentation related to MediaDrm offline use case
Jeff Tinker [Tue, 7 May 2013 21:07:10 +0000 (14:07 -0700)]
Correct native API documentation related to MediaDrm offline use case

These doc updates were made previously to the java API but some
did not make it into the native API.

Change-Id: I9b7bfcd77bc593685254baca1c148deb7e1f91c7
related-to-bug: 8621588

10 years agoRemove experimental HWC virtual display support
Jesse Hall [Tue, 7 May 2013 18:58:47 +0000 (11:58 -0700)]
Remove experimental HWC virtual display support

Bug: 8384764
Change-Id: I97b52ed83ad85466bd91cb9291308994048568a1

10 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Tue, 7 May 2013 12:55:26 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

10 years agoAdd support for HAL_PIXEL_FORMAT_YCbCr_420_888
Eino-Ville Talvala [Sun, 5 May 2013 01:07:43 +0000 (18:07 -0700)]
Add support for HAL_PIXEL_FORMAT_YCbCr_420_888

- Add fields to CpuConsumer::LockedBuffer for new information
- New lock methods for GraphicBuffer and GraphicBufferMapper for
  the format

Bug: 8734880
Change-Id: If31f82c62d64b6942cf4cc6e5715585c03273f12

10 years agoAdd Dalvik heap definition for 7" xhdpi devices.
Dianne Hackborn [Mon, 6 May 2013 22:06:20 +0000 (15:06 -0700)]
Add Dalvik heap definition for 7" xhdpi devices.

Change-Id: I6a8e406ee98d3061d498bb2fbd1b97aa9e79b4b5

10 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Mon, 6 May 2013 12:55:30 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

10 years agoMerge "Report actual disk space used" into jb-mr2-dev
Marco Nelissen [Sun, 5 May 2013 01:17:48 +0000 (01:17 +0000)]
Merge "Report actual disk space used" into jb-mr2-dev

10 years agoMerge "Fix issue #8768456: Settings > App Info under reports..." into jb-mr2-dev
Dianne Hackborn [Thu, 2 May 2013 22:11:17 +0000 (22:11 +0000)]
Merge "Fix issue #8768456: Settings > App Info under reports..." into jb-mr2-dev

10 years agoGenerate new EGLExt class
Jesse Hall [Tue, 30 Apr 2013 20:45:14 +0000 (13:45 -0700)]
Generate new EGLExt class

Initially populated with EGL_ANDROID_presentation_time (moved from the
EGL14 class) and the ES-relevant parts of EGL_KHR_create_context.

Bug: 8678160
Change-Id: Ifed2ee3da264ca701ae1f4b309a0758f7fcc3acc

10 years agoFix issue #8768456: Settings > App Info under reports...
Dianne Hackborn [Thu, 2 May 2013 01:55:10 +0000 (18:55 -0700)]
Fix issue #8768456: Settings > App Info under reports...

...app storage size for apps w/ .so files

The lib directories are tagged with the apk install number,
so must be explicitly passed down to installd.

Change-Id: Iae8815afd6ba964f5b2ed86a0d04a91827391ed6

10 years agoReport actual disk space used
Marco Nelissen [Wed, 1 May 2013 17:10:59 +0000 (10:10 -0700)]
Report actual disk space used

Report size based on blocks used, not the "nominal" size.
b/8744465

Change-Id: Idbb72f47af6d475b51305a14be291bf01acb14bb

10 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Wed, 1 May 2013 12:55:47 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

10 years agoMerge "Fix glGetUniformIndices JNI prototype" into jb-mr2-dev
Jesse Hall [Wed, 1 May 2013 02:30:47 +0000 (02:30 +0000)]
Merge "Fix glGetUniformIndices JNI prototype" into jb-mr2-dev

10 years agoMerge "Recommend different time source" into jb-mr2-dev
Andy McFadden [Wed, 1 May 2013 00:43:55 +0000 (00:43 +0000)]
Merge "Recommend different time source" into jb-mr2-dev

10 years agoFix glGetUniformIndices JNI prototype
Jesse Hall [Mon, 29 Apr 2013 22:58:38 +0000 (15:58 -0700)]
Fix glGetUniformIndices JNI prototype

Bug: 8657863
Change-Id: I130d1b54d0cbd1f8613d3108482526d05d73d9e4

10 years agoRecommend different time source
Andy McFadden [Fri, 26 Apr 2013 18:17:57 +0000 (11:17 -0700)]
Recommend different time source

We were recommending SystemClock#uptimeMillis() from Java code, but
millisecond resolution is a little coarse.  We don't have an
appropriate call in SystemClock, so use System.nanoTime() instead.

Bug 8733179

Change-Id: Ifb85a03eabf7ff168966adcb5ed5914bbf3884b2

10 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Mon, 29 Apr 2013 12:55:47 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

11 years agoEGL: expose the EGL_KHR_create_context extension
Jamie Gennis [Fri, 26 Apr 2013 01:48:41 +0000 (18:48 -0700)]
EGL: expose the EGL_KHR_create_context extension

Bug: 8616866
Change-Id: Ic60f9f21397a1760c552b6f141d5177d1cbaf1f8

11 years agoMerge "libgui_test: fix a couple minor test issues" into jb-mr2-dev
Jamie Gennis [Thu, 25 Apr 2013 19:39:10 +0000 (19:39 +0000)]
Merge "libgui_test: fix a couple minor test issues" into jb-mr2-dev

11 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Wed, 24 Apr 2013 12:55:32 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

11 years agolibgui_test: fix a couple minor test issues
Jamie Gennis [Tue, 23 Apr 2013 22:00:45 +0000 (15:00 -0700)]
libgui_test: fix a couple minor test issues

This change fixes how the MultiTextureConsumerTest.EGLImageTargetWorks checks
for pixels.  It removes the call to eglSwapBuffers so that the test does not
rely on EGL swap-preserve behavior, and it makes the test use checkPixel()
rather than doing glReadPixels itself.

Bug: 8349336
Change-Id: I0f446a0083eebd07af6fd208762878b4e367725a

11 years agoMerge "Add hardware feature android.hardware.bluetooth.le xml file" into jb-mr2-dev
Matthew Xie [Tue, 23 Apr 2013 22:47:36 +0000 (22:47 +0000)]
Merge "Add hardware feature android.hardware.bluetooth.le xml file" into jb-mr2-dev

11 years agoAdd hardware feature android.hardware.bluetooth.le xml file
Matthew Xie [Thu, 18 Apr 2013 22:32:04 +0000 (15:32 -0700)]
Add hardware feature android.hardware.bluetooth.le xml file

The feature is used to differ platforms that support bluetooth low
engery radio
bug 8369273

Change-Id: I8714b6b03a4e5692d4573ea7c156b002b7ee42dd

11 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Tue, 23 Apr 2013 12:55:29 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

11 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Mon, 22 Apr 2013 12:55:39 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

11 years agofix uninitialized variables in GLConsumer
Mathias Agopian [Mon, 22 Apr 2013 09:24:02 +0000 (11:24 +0200)]
fix uninitialized variables in GLConsumer

mDefaultWidth, mDefaultHeight and mCurrentScallingMode are now
initialized to the same default value that BufferQueue uses.

Change-Id: I0d4da2022b06419d12745716d8ddbd48c8869953

11 years agoallow dumpsys SurfaceFlinger in user builds from the shell user
Mathias Agopian [Thu, 18 Apr 2013 23:41:04 +0000 (16:41 -0700)]
allow dumpsys SurfaceFlinger in user builds from the shell user

Bug: 8659013
Change-Id: I1de43bb5808cb8c1103cfc73dca59ff70fc13f91

11 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Thu, 18 Apr 2013 12:55:41 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

11 years agoMerge "Fix activation issue with auto disabled sensors." into jb-mr2-dev
Jaikumar Ganesh [Thu, 18 Apr 2013 04:56:29 +0000 (04:56 +0000)]
Merge "Fix activation issue with auto disabled sensors." into jb-mr2-dev

11 years agoFix activation issue with auto disabled sensors.
Jaikumar Ganesh [Tue, 16 Apr 2013 22:52:23 +0000 (15:52 -0700)]
Fix activation issue with auto disabled sensors.

Auto disabled sensors get auto disabled after trigger.
An activation after this wasn't working because the
state was not being reset.

b/8609561

Change-Id: If72c9f27345e91671d7ad0a7a066f6dc3d255b78

11 years agoMerge "BufferItemConsumer: add functions to set default buffer format/size" into...
Igor Murashkin [Wed, 17 Apr 2013 19:39:52 +0000 (19:39 +0000)]
Merge "BufferItemConsumer: add functions to set default buffer format/size" into jb-mr2-dev

11 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Wed, 17 Apr 2013 12:55:32 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

11 years agoAdd a test that checks multi-texturing works with GL_TEXTURE_EXTERNAL_OES
Mathias Agopian [Wed, 17 Apr 2013 06:32:38 +0000 (23:32 -0700)]
Add a test that checks multi-texturing works with GL_TEXTURE_EXTERNAL_OES

Bug: 8395618
Change-Id: I160d29699593b11058533ca753e275d5deeb3b54

11 years agofix "traceGpuCompletion" debug feature
Mathias Agopian [Tue, 16 Apr 2013 23:39:46 +0000 (16:39 -0700)]
fix "traceGpuCompletion" debug feature

eglCreateSyncKHR requires a call to glFlush which we were not doing.
fixed by moving the code above eglSwapBuffers(), which both
fixes the problem and gives us a slightly better idea of when
the GPU is done drawing.

Change-Id: Ic826ef1fe25a6247742c3d49d0cb69f4031e3593

11 years agoMerge changes I66085c30,Ic9b3be9f,Icc6b23ab,I359a9511,I11a7394e,Ic8bcd03e,I40834d6f...
Jesse Hall [Tue, 16 Apr 2013 21:29:39 +0000 (21:29 +0000)]
Merge changes I66085c30,Ic9b3be9f,Icc6b23ab,I359a9511,I11a7394e,Ic8bcd03e,I40834d6f,I02a1548a,I638a36b0,I814befe2,Iaaef7d53,I349575e5,I0b5acc78,I6fad8803 into jb-mr2-dev

* changes:
  Subclass GLES30 from GLES20, @Deprecate GL_STENCIL_INDEX
  Special-case glGetActiveUniformBlockName
  Special-case glGetStringi
  Special-case glMapBufferRange
  Special-case glGetBufferPointerv
  Special-case glGetUniformIndices
  Special-case glGetTransformFeedbackVarying
  Special-case glTransformFeedbackVaryings
  Add buffer object versions of several functions
  Add *int64 and GLsync types and related functions
  Add ES3 functions and constants, difficult ones commented out
  Support "const GLChar*" and "const GLenum*" types
  Minor changes to ES3 functions inherited from ES2
  Generate GLES30 class, just a clone of GLES20 for now

11 years agoBufferItemConsumer: add functions to set default buffer format/size
Igor Murashkin [Tue, 16 Apr 2013 18:24:40 +0000 (11:24 -0700)]
BufferItemConsumer: add functions to set default buffer format/size

Bug: 8629088
Change-Id: I06f2e26a70d6dbcfcd70b08b4461e8e401f6e83f

11 years agoSubclass GLES30 from GLES20, @Deprecate GL_STENCIL_INDEX
Jesse Hall [Mon, 15 Apr 2013 23:35:08 +0000 (16:35 -0700)]
Subclass GLES30 from GLES20, @Deprecate GL_STENCIL_INDEX

Bug: 8566953
Change-Id: I66085c300a8658a4018796645257edb73be94719

11 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Mon, 15 Apr 2013 12:55:54 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

11 years agoMerge "Make gen more readable, use UNIX-style line endings" into jb-mr2-dev
Jesse Hall [Sat, 13 Apr 2013 21:18:25 +0000 (21:18 +0000)]
Merge "Make gen more readable, use UNIX-style line endings" into jb-mr2-dev

11 years agoMerge "make sure to update the "current state" when we "fake" the transaction" into...
Mathias Agopian [Sat, 13 Apr 2013 00:27:31 +0000 (00:27 +0000)]
Merge "make sure to update the "current state" when we "fake" the transaction" into jb-mr2-dev

11 years agoMerge "Sensor changes." into jb-mr2-dev
Jaikumar Ganesh [Sat, 13 Apr 2013 00:22:38 +0000 (00:22 +0000)]
Merge "Sensor changes." into jb-mr2-dev

11 years agoAdd new resources trace.
Dianne Hackborn [Fri, 12 Apr 2013 21:52:35 +0000 (14:52 -0700)]
Add new resources trace.

Change-Id: Iab8c1ef51f491e818fb4bab0ba7a042f0d9bec9a

11 years agomake sure to update the "current state" when we "fake" the transaction
Mathias Agopian [Fri, 12 Apr 2013 23:58:11 +0000 (16:58 -0700)]
make sure to update the "current state" when we "fake" the transaction

this mimics the code we used for the size. basically we need to
also update the "current state" because it gets copied to the
"drawing state" when a transaction occurs, and it would
"undo" our change.

Bug: 8511430
Bug: 8581533
Change-Id: I08c02abbf21b7f168f7124cd14ee717d7d3d502c

11 years agoSensor changes.
Jaikumar Ganesh [Mon, 8 Apr 2013 23:43:12 +0000 (16:43 -0700)]
Sensor changes.

1. Some sensors can wake up the AP. Add wakelocks.
2. Handle backward compatibility for rotation vector
   heading accuracy.
3. Cleanup auto disabled sensors.
4. Fix race condition between enable and dispatch.

Change-Id: I39dddf12e208d83cd288201986ee994312555820

11 years agoMerge "DO NOT MERGE Make sure surfaces always have latest orientation info" into...
Chet Haase [Fri, 12 Apr 2013 14:11:53 +0000 (14:11 +0000)]
Merge "DO NOT MERGE Make sure surfaces always have latest orientation info" into jb-mr2-dev

11 years agoImprove screenshot performance on some devices (DO NOT MERGE)
Mathias Agopian [Thu, 11 Apr 2013 05:55:41 +0000 (22:55 -0700)]
Improve screenshot performance on some devices (DO NOT MERGE)

this affects devices that need a glReadPixels(). We use
a FBO instead of a GlConsumer as an intermediate render target, this
saves 2 calls to eglMakeCurrent().

On Galaxy Nexus this allows us to go from ~135ms to ~35ms for
recent's screenshots.

Bug: 8582615
Change-Id: I6b25291ecc235f1927579bbb2db3c731e985c6e8

11 years agoMerge "Improve screenshot performance on some devices (DO NOT MERGE)" into jb-mr2-dev
Mathias Agopian [Fri, 12 Apr 2013 03:18:17 +0000 (03:18 +0000)]
Merge "Improve screenshot performance on some devices (DO NOT MERGE)" into jb-mr2-dev

11 years agoImprove screenshot performance on some devices (DO NOT MERGE)
Mathias Agopian [Thu, 11 Apr 2013 05:55:41 +0000 (22:55 -0700)]
Improve screenshot performance on some devices (DO NOT MERGE)

this affects devices that need a glReadPixels(). We use
a FBO instead of a GlConsumer as an intermediate render target, this
saves 2 calls to eglMakeCurrent().

On Galaxy Nexus this allows us to go from ~135ms to ~35ms for
recent's screenshots.

Bug: 8582615
Change-Id: I6b25291ecc235f1927579bbb2db3c731e985c6e8

11 years agoDO NOT MERGE Make sure surfaces always have latest orientation info
Chet Haase [Thu, 11 Apr 2013 22:24:55 +0000 (15:24 -0700)]
DO NOT MERGE Make sure surfaces always have latest orientation info

When the screen is turned off, the current stack is set to -1. This causes
logic in iSurfaceFlinger's handleTransactionLocked() function to fail to
match the current stack, and the latest orientation is not set into the
layer. This causes BufferQueue, later, to potentially set an obsolete
transformHint on a created surface (such as in the case with ImageWallpaper's
Egl surface, in the bug below).

The fix is to note this situation and use a default value for the DisplayDevice,
which should have the current orientation information.

Issue #8508397 ImageWallpaper sometimes rendered in wrong orientation causing a ~30-40% drop in graphics performance

Change-Id: Ibae15d73b289a8343c67f4f6bb77fdf11dd95ee7

11 years agoMerge "Updated comments" into jb-mr2-dev
Andy McFadden [Thu, 11 Apr 2013 17:15:52 +0000 (17:15 +0000)]
Merge "Updated comments" into jb-mr2-dev

11 years agomerge in jb-mr2-release history after reset to jb-mr2-dev
The Android Automerger [Thu, 11 Apr 2013 12:55:37 +0000 (05:55 -0700)]
merge in jb-mr2-release history after reset to jb-mr2-dev

11 years agoSpecial-case glGetActiveUniformBlockName
Jesse Hall [Thu, 11 Apr 2013 04:49:55 +0000 (21:49 -0700)]
Special-case glGetActiveUniformBlockName

Bug: 8566953
Change-Id: Ic9b3be9fa0136495633ee3b1c38583cfa315942a

11 years agoSpecial-case glGetStringi
Jesse Hall [Wed, 10 Apr 2013 23:27:00 +0000 (16:27 -0700)]
Special-case glGetStringi

Modeled on the existing glGetString special-case.

Bug: 8566953
Change-Id: Icc6b23ab53b00510368900ad99c92028253dc797

11 years agoSpecial-case glMapBufferRange
Jesse Hall [Wed, 10 Apr 2013 23:13:35 +0000 (16:13 -0700)]
Special-case glMapBufferRange

Updating the generator to handle GLvoid*/java.nio.Buffer return values
probably wouldn't be too hard, but this is the only function that
needs it currently.

Bug: 8566953
Change-Id: I359a951136ab479ab576ce2e5a2881b937b7e8c9

11 years agoSpecial-case glGetBufferPointerv
Jesse Hall [Wed, 10 Apr 2013 22:48:58 +0000 (15:48 -0700)]
Special-case glGetBufferPointerv

Bug: 8566953
Change-Id: I11a7394eb46f229fd3ae716e77487a45bac6ed53

11 years agoSpecial-case glGetUniformIndices
Jesse Hall [Wed, 10 Apr 2013 16:26:48 +0000 (09:26 -0700)]
Special-case glGetUniformIndices

Bug: 8566953
Change-Id: Ic8bcd03e8d41a81f48d603f67ce2046a4afa1561

11 years agoSpecial-case glGetTransformFeedbackVarying
Jesse Hall [Wed, 10 Apr 2013 07:50:32 +0000 (00:50 -0700)]
Special-case glGetTransformFeedbackVarying

Just a search-and-replace on the glGetActiveAttrib special-case, since
they have the same signature and parameter handling.

Bug: 8566953
Change-Id: I40834d6f32bf9ecb39718df29036aa2d1bbfa07a

11 years agoSpecial-case glTransformFeedbackVaryings
Jesse Hall [Wed, 10 Apr 2013 07:34:07 +0000 (00:34 -0700)]
Special-case glTransformFeedbackVaryings

Bug: 8566953
Change-Id: I02a1548aebc12bd3599903029bfd2e4ccea53211

11 years agoAdd buffer object versions of several functions
Jesse Hall [Wed, 10 Apr 2013 08:17:34 +0000 (01:17 -0700)]
Add buffer object versions of several functions

Some of these are new ES3 functions, some are existing ES2 functions
that can now use the new pixel pack/unpack buffer bindings.

glDrawElementsInstanced needs a special case since the pointer/offset
arg isn't the last one like the generator assumes.

Bug: 8566953
Change-Id: I638a36b0a31aefcb5bfee6f4d049348223045103

11 years agoAdd *int64 and GLsync types and related functions
Jesse Hall [Wed, 10 Apr 2013 08:01:00 +0000 (01:01 -0700)]
Add *int64 and GLsync types and related functions

Return values are declared with the C return type, but the JNI
function returns the JNI return type. In the case of GLsync/jlong as
in glFenceSync(), this causes a compile error. So the generator now
explicitly casts the return value to the JNI return type.

Bug: 8566953
Change-Id: I814befe2e4cce745434cbc4e1c8639fc3ce8aeae

11 years agoAdd ES3 functions and constants, difficult ones commented out
Jesse Hall [Fri, 29 Mar 2013 20:06:32 +0000 (13:06 -0700)]
Add ES3 functions and constants, difficult ones commented out

This change adds ES3 functions to GLES30.spec, disabling any that
require non-trivial changes for the generator to handle.

Steps taken to add these:
- Copy ES3 function declarations from gl3.h
- Remove GL_APICALL, GL_APIENTRY, and semicolon from each declaration
- Add whitespace around parens and *s as required by the parser
- Comment out functions that the generator doesn't understand or that
  it generates bad Java interfaces for (by inspection).

Bug: 8566953
Change-Id: Iaaef7d53e24f9a576759dbba72cd206bae1c1276

11 years agoMerge "fix another bug where screenshots could end-up all black" into jb-mr2-dev
Mathias Agopian [Thu, 11 Apr 2013 04:43:48 +0000 (04:43 +0000)]
Merge "fix another bug where screenshots could end-up all black" into jb-mr2-dev

11 years agofix another bug where screenshots could end-up all black
Mathias Agopian [Wed, 10 Apr 2013 23:27:17 +0000 (16:27 -0700)]
fix another bug where screenshots could end-up all black

SF transactions were always handled on VSYNC which allowed
the screenshot to sneak-in between closing the transaction
and vsync (before it's latched), resulting in a screenshot
with the previous state.

we now always force transactions to happen immediately
before screenhots.

Bug: 7552304
Change-Id: I0afc86b7e8366173daff5b9988bbb4d2a0f43860

11 years agoMake gen more readable, use UNIX-style line endings
Jesse Hall [Fri, 29 Mar 2013 17:49:58 +0000 (10:49 -0700)]
Make gen more readable, use UNIX-style line endings

Change-Id: I6f4a1afa19f84d89278e9447f14a124476106fef

11 years agoSupport "const GLChar*" and "const GLenum*" types
Jesse Hall [Fri, 29 Mar 2013 22:20:45 +0000 (15:20 -0700)]
Support "const GLChar*" and "const GLenum*" types

Bug: 8566953
Change-Id: I349575e592e8b0d93d14237230b69bec4e04cf3e