drgreen/android_frameworks_base
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
205,470 Commits 1 Branch 0 Tags
Commit Graph

2156 Commits

Author SHA1 Message Date
Chris Craik
43b42d8bdb am 41e4d8b3: am 0c218e43: am 9ca926ad: am fdb87b76: am 127b478a: Merge "Remove all usage of fmin and fmax" into mnc-dev 
* commit '41e4d8b3ed044ed724bc340b77d3f9ca60105407':
  Remove all usage of fmin and fmax
 2015-07-08 03:15:47 +00:00
Chris Craik
127b478ac6 Merge "Remove all usage of fmin and fmax" into mnc-dev 2015-07-08 01:53:57 +00:00
Chris Craik
e6a15ee3d0 Remove all usage of fmin and fmax 
bug:22208220

Removes needless call, and upconversion to doubles in multiple places.

Change-Id: I1b949fa5f206446ac34de800154c0147d6bd8034
 2015-07-07 18:42:42 -07:00
Chris Craik
4a067fb999 am 9d080ff6: am ab614e54: am 30eaf590: am df134e87: am 54d1f043: Merge "Simplify matrix copy constructor/load" into mnc-dev 
* commit '9d080ff64162da9772974400c53566659ab28eec':
  Simplify matrix copy constructor/load
 2015-07-07 20:38:06 +00:00
Chris Craik
54d1f0439e Merge "Simplify matrix copy constructor/load" into mnc-dev 2015-07-07 19:09:43 +00:00
John Reck
d67db081a3 am 7772b03d: am 6ac3b54d: am 8e5d4761: am dfbae1f9: am 15b64099: Merge "Adjust for pipeline stalls" into mnc-dev 
* commit '7772b03d5e3a8bce01cac511197afc52c9e56f5f':
  Adjust for pipeline stalls
 2015-07-07 19:08:46 +00:00
Chris Craik
d2bfcc74fc Simplify matrix copy constructor/load 
bug:22208220

Change-Id: Id096eb8de8da0dd1687d6af9456b16ce2a1a589b
 2015-07-07 10:09:35 -07:00
John Reck
be3fba05e8 Adjust for pipeline stalls 
Bug: 20853441

Calculating duration that crosses the UI-RT
sync point will now subtract out the time spent waiting
in queue under the assumption that this time will be
accounted for in the previous frame's metrics

Change-Id: Ia8213f4410638840613f5ae439e98dfb77532a6a
 2015-07-06 16:12:46 -07:00
John Reck
6b6ba748f0 am e463f28e: am 1826da24: am f70a5a78: Merge "Set go-faster arguments" into mnc-dev 
* commit 'e463f28eae88d0433a7bb41f5534b86000578900':
  Set go-faster arguments
 2015-07-04 07:28:02 +00:00
John Reck
0606a04136 am 91d01e5d: am 0cec5293: am d7580333: Merge "Use std::sort instead of quickSortX" into mnc-dev 
* commit '91d01e5d8b9a50ad04f0d0b5d6e35885f5d878a3':
  Use std::sort instead of quickSortX
 2015-07-04 07:26:47 +00:00
Chris Craik
7b293b5556 am c3988e0a: am b4c02e8f: am d3ecb3e6: Merge "Switch from fminf/fmaxf to std::min/max" into mnc-dev 
* commit 'c3988e0aa337b74693feb0abb898eb87fbb61b58':
  Switch from fminf/fmaxf to std::min/max
 2015-07-03 20:35:50 +00:00
John Reck
f70a5a78bb Merge "Set go-faster arguments" into mnc-dev 2015-07-01 23:52:20 +00:00
John Reck
d758033368 Merge "Use std::sort instead of quickSortX" into mnc-dev 2015-07-01 20:55:23 +00:00
John Reck
5600d579b9 Set go-faster arguments 
Bug: 22208220

Increases performance of shadows by yet another
15% by using more aggressive math optimizations

Change-Id: Ie3fabeae3647632a19ff573bf6b2d7f7a7cef555
 2015-07-01 13:51:44 -07:00
Chris Craik
d3ecb3e6f0 Merge "Switch from fminf/fmaxf to std::min/max" into mnc-dev 2015-07-01 18:10:33 +00:00
John Reck
1e4209e387 Use std::sort instead of quickSortX 
Bug: 22208220

Libcxx has a really good sort algorithm, use that
instead. This speeds up shadow tesellation by about 10%.

Also less code

Change-Id: Iaad424187121d7644076f94ba8a3cf4c110da2f2
 2015-07-01 09:54:47 -07:00
Chris Craik
df72b63928 Switch from fminf/fmaxf to std::min/max 
bug:22208220

Shows considerable improvement in performance, especially in tight
loops.

Change-Id: I4bcf6584a3c145bfc55e73c9c73dcf6199290b3c
 2015-06-30 17:58:21 -07:00
Chris Craik
ba3964f48c am 6aa9b4ab: am cf225194: am a763bb83: Merge "Saturate alpha values when overlaying glyphs" into mnc-dev 
* commit '6aa9b4abf335d7ecf5c33cc48b41f9f9f8cbd8ab':
  Saturate alpha values when overlaying glyphs
 2015-06-30 21:54:52 +00:00
Chris Craik
de25a672b1 Saturate alpha values when overlaying glyphs 
bug:19062769

Glyphs were stored to the output buffer with the assumption of
non-overlap, which is incorrect for certain fonts/strings. Instead,
blend src into dst, so new glyphs don't clobber existing content.

Change-Id: I6e22037500e67d5348ee2a43d939416c23c4d1ea
 2015-06-30 13:24:18 -07:00
Chris Craik
8868e02e2d am 00562d69: am 5805765d: am f8d25ea8: Merge "Don\'t promote functors onto layers due to outline rect clipping" into mnc-dev 
* commit '00562d693bfcfa4409977becbbd0db2c8408e291':
  Don't promote functors onto layers due to outline rect clipping
 2015-06-26 21:04:48 +00:00
Chris Craik
f8d25ea8eb Merge "Don't promote functors onto layers due to outline rect clipping" into mnc-dev 2015-06-26 20:28:36 +00:00
Chris Craik
b60d3e7856 Don't promote functors onto layers due to outline rect clipping 
bug:21320333

Change-Id: Ie70982c88498395148dab6eaca861de66e6207ed
 2015-06-25 17:15:17 -07:00
Tenghui Zhu
4540789492 am a1b1effc: am 4b05c5a9: am 5daee1dc: Merge "Limit dotProduct value to 1.0f, so acosf would not return NaN." into mnc-dev 
* commit 'a1b1effc1e19d534b25578f5ed4986d3c65c8baa':
  Limit dotProduct value to 1.0f, so acosf would not return NaN.
 2015-06-25 22:20:32 +00:00
Lazar Trsic
1240752c44 Limit dotProduct value to 1.0f, so acosf would not return NaN. 
Cherry pick of b561f39d01c211425bfefaaa7b31ebe097e7ba79 from AOSP master.

Due to precision loss of float math, we sometimes get 1.000001f for
dotProduct. This causes NaN result from acosf() and floor() funcs.

At the moment, this does not cause any problems on ARM, as casting
NaN to int results in 0. On mips however (possibly on x86), such cast
gives INT_MAX, so crash occurs when trying to use the resulting value.

Change-Id: I8e0285a0306a65b8469d9f4885c19665066fc4c8
 2015-06-25 14:22:13 -07:00
Chris Craik
d739b5f39c am ee6d943b: am 46de9f56: am 6659ac77: Merge "Avoid over-damaging layer area for multidraws" into mnc-dev 
* commit 'ee6d943b9e13cde263ccb622dd529922dfa1f3b3':
  Avoid over-damaging layer area for multidraws
 2015-06-24 18:40:14 +00:00
Chris Craik
6659ac77ed Merge "Avoid over-damaging layer area for multidraws" into mnc-dev 2015-06-24 17:58:01 +00:00
Tom Hudson
02a26300fb Switch framework to new Skia enums 
Allow Skia to finish deprecating old path operation enumerations.

Change-Id: I61e4489e639043acb78d47994ab62a31f18bd878
 2015-06-24 16:56:43 +00:00
Chris Craik
3375f8ad30 Avoid over-damaging layer area for multidraws 
bug:19608961

Change-Id: Ib7495f155da62a352e9a010a362fd40ec0acad21
 2015-06-24 09:43:54 -07:00
Tom Hudson
7605fdd466 am d29c9a86: am 5e69713a: am a81a0c86: Merge "Sync canvas proxy CTM (b/21945972)" into mnc-dev 
* commit 'd29c9a86d2734db66533d6525f64caca57d63808':
  Sync canvas proxy CTM (b/21945972)
 2015-06-24 15:23:08 +00:00
Tom Hudson
a81a0c8647 Merge "Sync canvas proxy CTM (b/21945972)" into mnc-dev 2015-06-24 14:53:58 +00:00
Tom Hudson
90fb1f6732 Sync canvas proxy CTM (b/21945972) 
SkiaCanvasProxy was being created with an identity transform, ignoring
any transform that may have been applied in Java (or C++) to the
android graphics Canvas it was proxy for. This CL makes sure the
DisplayListCanvas transform is propagated to the Proxy every time
asSkCanvas() is called.

We could instead move the code to the SkiaCanvasProxy constructor
if we got rid of the cached proxy on DisplayListCanvas; nobody's
using the proxy heavily enough that that should be a performance
hit at this time.

BUG:21945972
R=djsollen@google.com

Change-Id: I99ed1563802a2449bb9939cb67976cd60dd8611c
 2015-06-24 10:43:16 -04:00
Chris Craik
e984230f6b am 0664eb56: am 2c62ad43: am c686757a: Merge "Unify clipping return value behavior - true if not empty" into mnc-dev 
* commit '0664eb56abcd4cb5b5e2cd96a1983ea734742b70':
  Unify clipping return value behavior - true if not empty
 2015-06-24 00:55:51 +00:00
Chris Craik
c686757a18 Merge "Unify clipping return value behavior - true if not empty" into mnc-dev 2015-06-24 00:12:51 +00:00
Chris Craik
5ec6a2878b Unify clipping return value behavior - true if not empty 
bug:22035484

Change-Id: I23ecca38554a60c590181290900deb0e2c48cea1
 2015-06-23 15:42:43 -07:00
Chris Craik
05298735d3 am cbd4c6fa: am 3ac8261f: am 0d99d43e: Merge "Dump GL errors and crash on failure to allocate texture for layer" into mnc-dev 
* commit 'cbd4c6fa1828da9f16956e897310155ed5c1aff6':
  Dump GL errors and crash on failure to allocate texture for layer
 2015-06-23 20:55:19 +00:00
Chris Craik
0d99d43e81 Merge "Dump GL errors and crash on failure to allocate texture for layer" into mnc-dev 2015-06-23 18:01:02 +00:00
Chris Craik
5686bae289 Dump GL errors and crash on failure to allocate texture for layer 
bug:22029728

Change-Id: Ib6f778975225b36567b12e30967955640ec185b3
 2015-06-23 10:34:54 -07:00
Chris Craik
83be135f48 am 3f3910a9: am d902c813: am 15d8a2bc: Merge "Avoid flushing DisplayListCanvas state for noop calls" into mnc-dev 
* commit '3f3910a91a15a63d7700235dc04191f0d7649105':
  Avoid flushing DisplayListCanvas state for noop calls
 2015-06-23 03:04:03 +00:00
Chris Craik
772687d242 Avoid flushing DisplayListCanvas state for noop calls 
bug:22006795

Change-Id: I2eceee69772b08f5319ea882be429a5b36860b7d
 2015-06-22 14:51:18 -07:00
John Reck
ba2609cc76 am f9fd28e7: am de3799ca: am f24bed80: Merge "Fix null deref" into mnc-dev 
* commit 'f9fd28e7948c245786c40d808642eaf868e94518':
  Fix null deref
 2015-06-16 22:04:29 +00:00
John Reck
3f725f0a5a Fix null deref 
Bug: 21857172

willHaveFunctor was assuming that either mStagingDisplayList
or mDisplayList would be non-null but this isn't true. Both
can be null in some scenarios (notably after a trimMemory)

Change-Id: Ia2915b84cfc907ec9c7cbec54ed6dc26283ec998
 2015-06-16 10:29:31 -07:00
Chris Craik
d7607cf520 am a15300e6: am f58af2e9: am 4c9dafbf: Merge "Put WebViews with on a HW layer if stencil/shader clipping is needed" into mnc-dev 
* commit 'a15300e6c91f2cfb09b72e8c707bc694bd961c34':
  Put WebViews with on a HW layer if stencil/shader clipping is needed
 2015-06-12 23:24:13 +00:00
Chris Craik
4c9dafbfec Merge "Put WebViews with on a HW layer if stencil/shader clipping is needed" into mnc-dev 2015-06-12 22:46:21 +00:00
Chris Craik
2385fe4675 am 95d1c25d: am f0088128: am eaef7b6c: Merge "Fix roundrect clipping / blending ordering" into mnc-dev 
* commit '95d1c25d03fbde301ea4db28140c6a42e248206d':
  Fix roundrect clipping / blending ordering
 2015-06-12 22:01:09 +00:00
Chris Craik
eaef7b6cb1 Merge "Fix roundrect clipping / blending ordering" into mnc-dev 2015-06-12 21:23:23 +00:00
Chris Craik
a766cb2bce Put WebViews with on a HW layer if stencil/shader clipping is needed 
bug:17322378
Change-Id: I0de574bf116b30e2ad4194366e19d47d49708902
 2015-06-12 13:34:58 -07:00
Chris Craik
b1f990de40 Fix roundrect clipping / blending ordering 
bug:21803918

Also, don't consider rr clipped ops opaque, and put stage
checking behind a debug flag.

Change-Id: I9304ae571337a7cf835524682d1a590a21d30fa4
 2015-06-12 13:20:37 -07:00
Chris Craik
161fa43942 am 623b8b1d: am 95dc7db4: am 816870bf: Merge "Revert to save-layer based alpha if layer would be too large" into mnc-dev 
* commit '623b8b1d652614f23d697260c90fc8d4c4ffef10':
  Revert to save-layer based alpha if layer would be too large
 2015-06-12 18:50:43 +00:00
Chris Craik
4e9d9b2a59 Revert to save-layer based alpha if layer would be too large 
bug:21011574
Change-Id: I323f2398188ffc2bbed6a98578d25d7e5f6c337d
 2015-06-12 11:07:23 -07:00
John Reck
ecfa74e256 am 6fb3b3c9: am 8693fc6f: am 475c545f: Merge "Upload bitmaps in SkShader* safely" into mnc-dev 
* commit '6fb3b3c97221057d5758df7abc69362719adfc29':
  Upload bitmaps in SkShader* safely
 2015-06-09 23:23:27 +00:00