Commit Graph

49407 Commits

Author SHA1 Message Date
Jamie Gennis
d1700756ec SurfaceFlinger: Add EventControlThread
This change adds a new thread for calling HWComposer's eventControl
asynchronously.  The DispSync-based vsync approach ends up enabling and
disabling HWComposer's vsync callbacks at arbitrary times, and some HWComposer
implementations do not have these calls optimized.

Bug: 11175503
Change-Id: I719be82bd200b391c61d40863b991c7b59acdfd6
2013-10-15 14:31:41 -07:00
Jesse Hall
247423751f Merge "Disable hardware vsync when blanking the screen" into klp-dev 2013-10-15 19:53:11 +00:00
Aravind Akella
487c7815d7 am 641a1868: Merge "SensorService Flush API bug fix." into klp-dev
* commit '641a18688bc6b208bbeaffa6083d3cf5f6790e4b':
  SensorService Flush API bug fix.
2013-10-15 10:05:09 -07:00
Aravind Akella
641a18688b Merge "SensorService Flush API bug fix." into klp-dev 2013-10-15 17:01:55 +00:00
The Android Automerger
3bd1b65ffb merge in klp-release history after reset to klp-dev 2013-10-15 06:00:54 -07:00
Jesse Hall
a04fd39cb8 am 8e26b28b: Filter out vsync events from HWC with duplicate timestamps
* commit '8e26b28be62e7c0e093900540db5e5cdcf52a0ff':
  Filter out vsync events from HWC with duplicate timestamps
2013-10-14 20:34:58 -07:00
Aravind Akella
c551eac5b1 SensorService Flush API bug fix.
SensorService is dead locking itself when flush is called on older devices which don't support batching. mConnectionLock is acquired twice.

Change-Id: I5c25585bfb2b396df4b05826a9cba1da7997a3ee
2013-10-14 17:09:19 -07:00
Jesse Hall
948fe0ce74 Disable hardware vsync when blanking the screen
Bug: 11220224
Change-Id: I99d0a42e1a6bb0aaf89706f6d100e9ef2a5deaa4
2013-10-14 16:25:01 -07:00
Jesse Hall
8e26b28be6 Filter out vsync events from HWC with duplicate timestamps
Bug: 11220224
Change-Id: I4efe0b66ea8969bf0ec3c4fcb325d354c8a0c315
2013-10-14 15:34:00 -07:00
The Android Automerger
4fec2900a3 merge in klp-release history after reset to klp-dev 2013-10-13 19:20:31 -07:00
Aravind Akella
055558b257 am 2d344979: Merge "Adding PackageManager feature strings for stepcounter and stepdetector." into klp-dev
* commit '2d344979eb200b7fb8ba4d811a22c80b1d77b564':
  Adding PackageManager feature strings for stepcounter and stepdetector.
2013-10-13 15:51:59 -07:00
Aravind Akella
2d344979eb Merge "Adding PackageManager feature strings for stepcounter and stepdetector." into klp-dev 2013-10-13 22:47:47 +00:00
Aravind Akella
d7a5342828 Adding PackageManager feature strings for stepcounter and stepdetector.
Change-Id: I75ec21900fdc3ea80b61ec12c851a0c5ce9c44e8
2013-10-13 14:53:45 -07:00
Jesse Hall
44b09f8b55 am 44e45cfe: Merge "DispSync: don\'t compensate for wakeup latency" into klp-dev
* commit '44e45cfeb37a04c206730b12ff001525ebe42a7b':
  DispSync: don't compensate for wakeup latency
2013-10-13 13:18:58 -07:00
Jesse Hall
44e45cfeb3 Merge "DispSync: don't compensate for wakeup latency" into klp-dev 2013-10-13 20:13:59 +00:00
The Android Automerger
26ef413c59 merge in klp-release history after reset to klp-dev 2013-10-13 06:02:02 -07:00
Andy McFadden
af4ae923c3 am 6410c8c4: Merge "Use black for blackout layer" into klp-dev
* commit '6410c8c4097f1a3246aa3d03b3291d47f982bc4b':
  Use black for blackout layer
2013-10-11 14:47:25 -07:00
Andy McFadden
6410c8c409 Merge "Use black for blackout layer" into klp-dev 2013-10-11 21:39:49 +00:00
Aravind Akella
4c4cd9c8cc Change API from flush(handle) to flush(). Call flush on all active sensors in the given SensorEventConnection.
Change-Id: I4ef2bec80406c517903ab9782dc9eaf3fa8b7f36
2013-10-11 12:42:13 -07:00
Andy McFadden
c6f2169559 Use black for blackout layer
SurfaceFlinger was rendering dark purple for secure content, which
showed up when we took a screen shot for the orientation change
animation.  Use black instead.

Bug 11157921

Change-Id: I3895e8168891d49dc5b84eed599bcd0a303bb70a
2013-10-11 11:16:03 -07:00
Michael Wright
a8b5982c32 Eliminate latency when resampling is disabled
Since the latency was introduced to avoid egregious errors during
resampling, remove it when we aren't resampling.

Change-Id: Ia8e9a9afb1e16212fec81a8c9c661c598e1b0ebf
2013-10-10 22:13:38 -07:00
Dave Burke
06c964d9b3 am 9c0dfa09: Merge "Eliminate latency when resampling is disabled" into klp-dev
* commit '9c0dfa09731a8366a7cdcd7171dde64b73af0d5c':
  Eliminate latency when resampling is disabled
2013-10-10 19:39:05 -07:00
Dave Burke
9c0dfa0973 Merge "Eliminate latency when resampling is disabled" into klp-dev 2013-10-11 02:33:15 +00:00
Aravind Akella
1f1e576840 am 5d6aa951: Merge "Change API from flush(handle) to flush(). Call flush on all active sensors in the given SensorEventConnection." into klp-dev
* commit '5d6aa95129a13aae37a60b3e4246a0592dab396d':
  Change API from flush(handle) to flush(). Call flush on all active sensors in the given SensorEventConnection.
2013-10-10 19:21:05 -07:00
Aravind Akella
5d6aa95129 Merge "Change API from flush(handle) to flush(). Call flush on all active sensors in the given SensorEventConnection." into klp-dev 2013-10-11 02:05:41 +00:00
Michael Wright
ad526f8505 Eliminate latency when resampling is disabled
Since the latency was introduced to avoid egregious errors during
resampling, remove it when we aren't resampling.

Change-Id: Ia8e9a9afb1e16212fec81a8c9c661c598e1b0ebf
2013-10-10 18:54:12 -07:00
Jesse Hall
8c76d2a16e am 5c65efe2: Merge "Wake up disp sync thread when setting display period" into klp-dev
* commit '5c65efe2b1d632226f02751e6b1d3cfd7c8ecb26':
  Wake up disp sync thread when setting display period
2013-10-10 16:05:37 -07:00
Jesse Hall
5c65efe2b1 Merge "Wake up disp sync thread when setting display period" into klp-dev 2013-10-10 22:24:32 +00:00
Jesse Hall
72f69d9f23 Wake up disp sync thread when setting display period
Bug: 10624956
Change-Id: I272c15f319521bfbd1bb70451366c7dd84d60bf9
2013-10-10 14:48:22 -07:00
Andy McFadden
0e9c51ec78 Fix blank / partial screenshots
The screen capture code wasn't waiting for the render to finish,
so sometimes you'd see an empty or partial image.

Bug 11131777

Change-Id: Ic64087322ce3bb15bb5f4fb1eb07579880fe6197
2013-10-10 12:16:58 -07:00
Andy McFadden
c00ab7b5aa am 2d8d120d: Fix blank / partial screenshots
* commit '2d8d120dc175c131a75e93fb1de61a9f432ddce9':
  Fix blank / partial screenshots
2013-10-10 08:23:01 -07:00
Jamie Gennis
0d5c60edf1 DispSync: don't compensate for wakeup latency
This change removes the wakeup latency compensation from the software-generated
vsync events.  Choreographer can't handle timestamps in the future, so don't
aim for early wake-ups with the expectation that the actual wake-up will be
late.

Bug: 11153576
2013-10-09 17:56:42 -07:00
Aravind Akella
701166d9f6 Change API from flush(handle) to flush(). Call flush on all active sensors in the given SensorEventConnection.
Change-Id: I4ef2bec80406c517903ab9782dc9eaf3fa8b7f36
2013-10-09 17:26:01 -07:00
Andy McFadden
2d8d120dc1 Fix blank / partial screenshots
The screen capture code wasn't waiting for the render to finish,
so sometimes you'd see an empty or partial image.

Bug 11131777

Change-Id: Ic64087322ce3bb15bb5f4fb1eb07579880fe6197
2013-10-09 16:38:02 -07:00
The Android Automerger
e0df32c9a7 merge in klp-release history after reset to klp-dev 2013-10-09 06:01:16 -07:00
Ed Heyl
6589a42c0d am 98c197d8: merge in KQS81M
* commit '98c197d826f95a82de41e16350f58a70a7ad23c6':
  GLConsumer: Add query for current frame number.
  Disable GLES 2.x rendering
  Initialize mIndent
2013-10-08 15:57:56 -07:00
Ed Heyl
98c197d826 merge in KQS81M 2013-10-08 15:40:38 -07:00
Jamie Gennis
790b6d092c am faf77cce: SurfaceFlinger: SW-based vsync events
* commit 'faf77cce9d9ec0238d6999b3bd0d40c71ff403c5':
  SurfaceFlinger: SW-based vsync events
2013-10-08 15:17:12 -07:00
Jamie Gennis
faf77cce9d SurfaceFlinger: SW-based vsync events
This change adds the DispSync class, which models the hardware vsync event
times to allow vsync event callbacks to be done at an arbitrary phase offset
from the hardware vsync.  This can be used to reduce the minimum latency from
Choreographer wake-up to on-screen image presentation.

Bug: 10624956
Change-Id: I8c7a54ceacaa4d709726ed97b0dcae4093a7bdcf
2013-10-07 17:59:53 -07:00
The Android Automerger
38234af6bd merge in klp-release history after reset to klp-dev 2013-10-06 06:01:27 -07:00
Alex Ray
859d1e490b am d457ec52: Add feature android.hardware.consumerir xml file
* commit 'd457ec52d7023e9a6112f313cd2bf94293162214':
  Add feature android.hardware.consumerir xml file
2013-10-04 14:18:35 -07:00
Marco Nelissen
e71575bed4 am 48d3741b: Merge "Allow blaming a single uid for work done" into klp-dev
* commit '48d3741ba5f4b3bb411ee5c083a11f4f52f9e907':
  Allow blaming a single uid for work done
2013-10-03 15:40:30 -07:00
Lajos Molnar
04e6ab6e27 am b1f94ef3: Merge "media/hardware/HardwareAPI.h: Added PrepareForAdaptivePlaybackParams" into klp-dev
* commit 'b1f94ef3f97bd20b928cf980791526861650eab9':
  media/hardware/HardwareAPI.h: Added PrepareForAdaptivePlaybackParams
2013-10-03 15:40:30 -07:00
Alex Ray
d457ec52d7 Add feature android.hardware.consumerir xml file
Bug: 10676690
Change-Id: I2cc62b9c97a71588c43ff96ca38d470976bde9b0
2013-10-03 13:09:37 -07:00
Marco Nelissen
48d3741ba5 Merge "Allow blaming a single uid for work done" into klp-dev 2013-10-03 16:34:51 +00:00
Lajos Molnar
b1f94ef3f9 Merge "media/hardware/HardwareAPI.h: Added PrepareForAdaptivePlaybackParams" into klp-dev 2013-10-03 16:10:25 +00:00
Marco Nelissen
9a7706bb17 Allow blaming a single uid for work done
This lets audioflinger assign blame for wakelocks to a single uid.
Currently this is only used for recording wakelocks.
b/10985160

Change-Id: Idc9adb22c29030010ac690d4cb95e7e7ea6b91e6
2013-10-03 08:49:07 -07:00
The Android Automerger
25274b00b0 merge in klp-release history after reset to klp-dev 2013-10-03 06:02:17 -07:00
Jesse Hall
062019b8ce Set the outbuf acquire fence after we actually have it.
In GLES-only mode, we don't have the outbuf acquire fence until after
GLES composition is done for the frame. We were setting the fence in
HWC's state immediately after dequeueing the buffer from the consumer,
before GLES had started. This fence got passed through HWC and on to
the consumer, so the consumer was reading the buffer before GLES was
done writing to it.

Now we update HWC's state just before set(), when we know we have the
right fence.

Bug: 11000763
Change-Id: Iea9db4c69634c352dc2d600f0bdb6bef2a432636
2013-10-02 14:44:42 -07:00
Mathias Agopian
d5350dea05 only clear FB when asked for the opaque layer
a layer need to be considered NOT opaque if it has a
plane-alpha.

Bug: 10846930
Change-Id: Ibd8981b63ede4560c7096bacc4cff46a7eb2a8bb
2013-10-02 14:44:42 -07:00