OSDN -- 开发和下载开源软件
简体中文翻译状态 translation status for zh

Android-x86
Fork
Donation

  • R/O
  • HTTP
  • SSH
  • HTTPS
Fork

frameworks-native: 提交列表

frameworks/native


jb-x86
RSS
Rev. 时间 作者
46d619d jb-x86 2013-11-04 23:47:25 Michael Chen

add HEVC decoder and cleanup

Change-Id: I16342228014320161a22302a3c14df7725556cd2
076a7da 2013-11-04 23:47:03 Chih-Wei Huang

Add OMX_AUDIO_CodingAPE and OMX_AUDIO_CodingDTS

Copied from the stagefright-plugins project.
b52bcd3 2013-08-02 13:23:45 Chih-Wei Huang

opengl: Implement fast tls forwarders for x86

Based on android-ia and rebased to JB-MR2.
dc2e201 2013-08-02 13:21:03 Chih-Wei Huang

Merge branch 'android-ia' into jb-x86
0cb8536 2013-08-02 13:20:35 Tapani Pälli

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>
6d50f12 2013-08-02 13:20:35 Andy Ross

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>
8bbfe07 2013-08-02 13:20:35 Andy Ross

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>
c9120c5 2013-08-02 13:20:35 Andy Ross

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>
f26c78f 2013-08-02 13:01:47 yi.xie@intel.com

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>
1e2af85 2013-08-02 13:01:47 Paul Drews

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>
17337ef 2013-08-02 13:01:46 Irina Tirdea

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>
644ddef 2013-08-02 13:01:46 Pauli Nieminen

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>
fc7e8db 2013-08-02 13:01:46 Pauli Nieminen

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>
26d8b29 2013-08-02 13:01:46 Pauli Nieminen

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>
7057f71 2013-08-02 13:01:46 Pauli Nieminen

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>
4594a12 2013-08-02 13:01:46 Daniel Charles

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>
48ca8e7 2013-07-25 18:38:51 Chih-Wei Huang

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.
52043fe 2013-07-25 18:38:51 Chih-Wei Huang

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.
b15a013 2013-07-25 18:38:51 Chih-Wei Huang

disables hardware acceleration if debug.egl.hw=0
70421f3 2013-07-16 00:19:12 The Android Automerger

merge in jb-mr2-zeroday-release history after reset to jb-mr2-dev
ac9a96d 2013-07-12 18:08:06 Mathias Agopian

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
176c0fd 2013-07-10 21:55:26 The Android Automerger

merge in jb-mr2-zeroday-release history after reset to jb-mr2-dev
c07b520 2013-07-05 04:08:16 Jesse Hall

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
e082dd6 2013-06-12 21:56:00 The Android Automerger

merge in jb-mr2-zeroday-release history after reset to jb-mr2-dev
ef07386 2013-06-12 10:03:01 Jeff Tinker

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
3042131 2013-05-19 21:55:29 The Android Automerger

merge in jb-mr2-release history after reset to jb-mr2-dev
23dacde 2013-05-18 05:20:53 Mathias Agopian

[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
ea74d3b 2013-05-17 10:15:57 Mathias Agopian

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
60b1606 2013-05-13 21:55:40 The Android Automerger

merge in jb-mr2-release history after reset to jb-mr2-dev
f743e3d 2013-05-10 02:13:22 Eino-Ville Talvala

Merge "Add support for HAL_PIXEL_FORMAT_YCbCr_420_888" into jb-mr2-dev
Show on old repository browser