msarett [Wed, 25 Mar 2015 13:29:18 +0000 (06:29 -0700)]
Fixing memory leak in ico decoder
BUG=skia:
NOTREECHECKS=true
TBR=scroggo@google.com
NOTRY=true
Review URL: https://codereview.chromium.org/
1036873002
msarett [Wed, 25 Mar 2015 12:27:48 +0000 (05:27 -0700)]
Enabling ico decoding with use of png and bmp decoders
BUG=skia:3257
NOPRESUBMIT=true
Review URL: https://codereview.chromium.org/
1011343003
halcanary [Wed, 25 Mar 2015 12:06:30 +0000 (05:06 -0700)]
SkPDF refactor skpdfdocument
Most fields removed, made local to emitPDF function (since we never
emit twice).
Variables made into stack variables or auto-unrefed.
we hold devices, not pages.
addResourcesToCatalog inlined
unused setPage function removed.
deprecated getCountOfFontTypes function removed
private functions removed
BUG=skia:3585
Review URL: https://codereview.chromium.org/
1034583002
borenet [Wed, 25 Mar 2015 11:47:02 +0000 (04:47 -0700)]
Add copyright check to PRESUBMIT
BUG=skia:1724
Review URL: https://codereview.chromium.org/
1003363005
reed [Wed, 25 Mar 2015 02:24:34 +0000 (19:24 -0700)]
SkChopCubicAt2 using Sk2s -- 2x faster
BUG=skia:
TBR=
Review URL: https://codereview.chromium.org/
1036753002
halcanary [Tue, 24 Mar 2015 22:29:39 +0000 (15:29 -0700)]
SkPDF: unit test cleanup
BUG=skia:3585
Review URL: https://codereview.chromium.org/
1033833002
reed [Tue, 24 Mar 2015 20:55:33 +0000 (13:55 -0700)]
Revert of pathops version two (patchset #16 id:150001 of https://codereview.chromium.org/
1002693002/)
Reason for revert:
ASAN investigation
Original issue's description:
> pathops version two
>
> R=reed@google.com
>
> marked 'no commit' to attempt to get trybots to run
>
> TBR=reed@google.com
>
> Committed: https://skia.googlesource.com/skia/+/
ccec0f958ffc71a9986d236bc2eb335cb2111119
TBR=caryclark@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.chromium.org/
1029993002
caryclark [Tue, 24 Mar 2015 20:52:49 +0000 (13:52 -0700)]
Revert of remove obsolete tests (patchset #1 id:1 of https://codereview.chromium.org/
1030883002/)
Reason for revert:
working on asan fix
Original issue's description:
> remove obsolete tests
>
> remove tests that are breaking the build -- uninitialized memory
>
> TBR=reed
>
> Committed: https://skia.googlesource.com/skia/+/
e02c944e97ad6a3b8befb1f4f784919767941ba6
TBR=reed@android.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.chromium.org/
1036633002
caryclark [Tue, 24 Mar 2015 20:47:57 +0000 (13:47 -0700)]
Revert of fix destructor order to fix build (patchset #1 id:1 of https://codereview.chromium.org/
1033703002/)
Reason for revert:
working on asan fix
Original issue's description:
> fix destructor order to fix build
>
> TBR=reed
>
> Committed: https://skia.googlesource.com/skia/+/
c207f9b2e8d6fb5386197fa8a8d258d2c4603418
TBR=reed@android.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.chromium.org/
1033803002
tomhudson [Tue, 24 Mar 2015 20:47:41 +0000 (13:47 -0700)]
Revert of Enabling ico decoding with use of png and bmp decoders (patchset #10 id:280001 of https://codereview.chromium.org/
1011343003/)
Reason for revert:
Reverting on suspicion of massive bot failures - possible command line too long?
Original issue's description:
> Enabling ico decoding with use of png and bmp decoders
>
> BUG=skia:3257
>
> Committed: https://skia.googlesource.com/skia/+/
15bfd075d38e4422a477e22940d06a137f66cc97
TBR=scroggo@google.com,reed@google.com,djsollen@google.com,msarett@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:3257
Review URL: https://codereview.chromium.org/
1022843005
tomhudson [Tue, 24 Mar 2015 19:46:40 +0000 (12:46 -0700)]
Switch DM to use utils/android/SkAndroidSDKCanvas
Use (better) paint filter shared with nanobench, instead
of custom implementation.
R=djsollen@google.com
Review URL: https://codereview.chromium.org/
1036523002
msarett [Tue, 24 Mar 2015 19:24:27 +0000 (12:24 -0700)]
Enabling ico decoding with use of png and bmp decoders
BUG=skia:3257
Review URL: https://codereview.chromium.org/
1011343003
robertphillips [Tue, 24 Mar 2015 18:09:07 +0000 (11:09 -0700)]
Fix assert in debugger
Review URL: https://codereview.chromium.org/
1029903002
rmistry [Tue, 24 Mar 2015 17:58:37 +0000 (10:58 -0700)]
Increase timeout of ukwsj to get more consistent SKP captures
BUG=skia:3574
TBR=borenet
NOTRY=true
Review URL: https://codereview.chromium.org/
1038443002
halcanary [Tue, 24 Mar 2015 17:21:34 +0000 (10:21 -0700)]
SkPDF: Fix leak in SkPDFShader
TBR=mtklein@google.com
Review URL: https://codereview.chromium.org/
1006813010
reed [Tue, 24 Mar 2015 17:04:38 +0000 (10:04 -0700)]
enable color-matrix-floats again, watching arm64 bots
BUG=skia:
TBR=mtklein@google.com
Review URL: https://codereview.chromium.org/
1035493002
mtklein [Tue, 24 Mar 2015 16:47:11 +0000 (09:47 -0700)]
Go back to storeAligned / LoadAligned for SkPMFloat <->Sk4f.
This seems to fix the miscompilation bug on ARM64 / Release / GCC 4.9.
We switched this over originally for perf issues with NEON, but I can't see any now. Will keep an eye out.
BUG=skia:3570
Review URL: https://codereview.chromium.org/
1026403002
caryclark [Tue, 24 Mar 2015 16:16:25 +0000 (09:16 -0700)]
fix destructor order to fix build
TBR=reed
Review URL: https://codereview.chromium.org/
1033703002
caryclark [Tue, 24 Mar 2015 15:17:16 +0000 (08:17 -0700)]
remove obsolete tests
remove tests that are breaking the build -- uninitialized memory
TBR=reed
Review URL: https://codereview.chromium.org/
1030883002
halcanary [Tue, 24 Mar 2015 15:12:55 +0000 (08:12 -0700)]
SkPDF: clean up extra references
This cleans up code left behind from http://crrev.com/
870333002
Review URL: https://codereview.chromium.org/
1029263004
senorblanco [Tue, 24 Mar 2015 14:51:30 +0000 (07:51 -0700)]
Move SkMatrixImageFilter.h into src/core.
Callers in Blink have been updated to call
SkImageFilter::CreateMatrixFilter() instead.
BUG=skia:3568
Review URL: https://codereview.chromium.org/
1013753013
caryclark [Tue, 24 Mar 2015 14:28:17 +0000 (07:28 -0700)]
pathops version two
R=reed@google.com
marked 'no commit' to attempt to get trybots to run
TBR=reed@google.com
Review URL: https://codereview.chromium.org/
1002693002
reed [Tue, 24 Mar 2015 13:35:23 +0000 (06:35 -0700)]
Revert of Revert of remove colorfilter native-565 support. complicating w/ no real value. (patchset #1 id:1 of https://codereview.chromium.org/
1022673007/)
Reason for revert:
fix for chrome has landed...
Original issue's description:
> Revert of remove colorfilter native-565 support. complicating w/ no real value. (patchset #2 id:20001 of https://codereview.chromium.org/
1015533011/)
>
> Reason for revert:
> skia/ext/benchmarking_canvas.cc references HasFilter16 :(
>
> Original issue's description:
> > remove colorfilter native-565 support. complicating w/ no real value.
> >
> > BUG=skia:
> > TBR=
> >
> > Committed: https://skia.googlesource.com/skia/+/
2151353d161fe389cdc0db62cfe1932c7680f710
>
> TBR=reed@chromium.org
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/
92b47c49016749249ff8521e424c4373b4a74241
TBR=reed@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review URL: https://codereview.chromium.org/
1016103004
mtklein [Tue, 24 Mar 2015 13:16:30 +0000 (06:16 -0700)]
aacc + bbdd
SkMatrix::mapPts() using aacc/bbdd was always worse than using badc():
- On Intel, it was faster than exisiting swizzle, but badc() is 10% faster still (one pshufd instead of two).
- On ARM, existing swizzle < badc() < aacc()+bbdd(), even though aacc() then bbdd() is really a single vtrn instruction.
I will revert SkMatrix.cpp before submitting. Just thought you might like to look.
Will think more and try to gear up Instruments on ARM.
BUG=skia:
Review URL: https://codereview.chromium.org/
1012573003
reed [Tue, 24 Mar 2015 12:18:09 +0000 (05:18 -0700)]
Revert of remove colorfilter native-565 support. complicating w/ no real value. (patchset #2 id:20001 of https://codereview.chromium.org/
1015533011/)
Reason for revert:
skia/ext/benchmarking_canvas.cc references HasFilter16 :(
Original issue's description:
> remove colorfilter native-565 support. complicating w/ no real value.
>
> BUG=skia:
> TBR=
>
> Committed: https://skia.googlesource.com/skia/+/
2151353d161fe389cdc0db62cfe1932c7680f710
TBR=reed@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review URL: https://codereview.chromium.org/
1022673007
reed [Tue, 24 Mar 2015 11:13:45 +0000 (04:13 -0700)]
Revert "Revert of impl colormatrix w/ floats (patchset #7 id:120001 of https://codereview.chromium.org/
967943002/)"
disabled for the momment -- need to understand arm64 glitches
This reverts commit
374a4c3b744e8c77455efa73b139950136f7b374.
BUG=skia:
TBR=
Review URL: https://codereview.chromium.org/
1028403002
reed [Tue, 24 Mar 2015 10:50:22 +0000 (03:50 -0700)]
remove colorfilter native-565 support. complicating w/ no real value.
BUG=skia:
TBR=
Review URL: https://codereview.chromium.org/
1015533011
reed [Tue, 24 Mar 2015 10:48:27 +0000 (03:48 -0700)]
Revert of correct reporting of kAlphaUnchanged_Flag in ModeFilter (patchset #1 id:1 of https://codereview.chromium.org/
1033583002/)
Reason for revert:
reverting experiment. need to rebaseilne layouttest
svg/css/text-gradient-shadow.svg
Original issue's description:
> correct reporting of kAlphaUnchanged_Flag in ModeFilter
>
> BUG=skia:
> TBR=
>
> Committed: https://skia.googlesource.com/skia/+/
afced66b508ac4cd3616e2f707db27b7d6f670e1
TBR=
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review URL: https://codereview.chromium.org/
1026313002
reed [Tue, 24 Mar 2015 03:05:04 +0000 (20:05 -0700)]
correct reporting of kAlphaUnchanged_Flag in ModeFilter
BUG=skia:
TBR=
Review URL: https://codereview.chromium.org/
1033583002
reed [Tue, 24 Mar 2015 01:52:57 +0000 (18:52 -0700)]
Revert of impl colormatrix w/ floats (patchset #7 id:120001 of https://codereview.chromium.org/
967943002/)
Reason for revert:
Arm64 seems to be having glitches :(
See gm:colormatrix on Arm64/TegraK1/Nexus9
last square should be all white, but it has stripes
Original issue's description:
> impl colormatrix w/ floats
>
> this needs to land first https://codereview.chromium.org/
1031713003
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/
7971def11be91ed08eae7107b372322d24e67544
TBR=mtklein@google.com,caryclark@google.com,reed@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review URL: https://codereview.chromium.org/
1030743003
reed [Mon, 23 Mar 2015 21:35:25 +0000 (14:35 -0700)]
impl colormatrix w/ floats
this needs to land first https://codereview.chromium.org/
1031713003
BUG=skia:
Review URL: https://codereview.chromium.org/
967943002
mtklein [Mon, 23 Mar 2015 20:19:03 +0000 (13:19 -0700)]
Start fresh on swizzles
This removes all the existing Sk4x swizzles and adds badc(), which is
both fast on all implementations and currently useful.
BUG=skia:
Review URL: https://codereview.chromium.org/
997353005
tomhudson [Mon, 23 Mar 2015 19:51:20 +0000 (12:51 -0700)]
Both DM and nanobench need this, so moving the duplicated code to
one common spot. (It's incomplete, and has had bugs, so it's not
like we can confidently write once, copy-paste, and not maintain
again.)
Because SkPathEffect::exposedInAndroidJavaAPI() only builds in the
Android Framework, we might want to make all this code Framework-only?
R=djsollen@google.com,mtklein@google.com
Review URL: https://codereview.chromium.org/
1021823003
mtklein [Mon, 23 Mar 2015 19:01:45 +0000 (12:01 -0700)]
Replace _mm_cvtps_epi32(x) with _mm_cvttps_epi32(_mm_add_ps(0.5f), x).
We don't have control over which way _mm_cvtps_epi32 rounds.
- This makes the SSE SkPMFloat rounding consistent with _neon and _none.
- Sk4f::cast<Sk4i>() is closer to (int)float's behavior. (Correct when >=0).
Add tests that would fail at head.
BUG=skia:
Review URL: https://codereview.chromium.org/
1029163002
bungeman [Mon, 23 Mar 2015 16:08:54 +0000 (09:08 -0700)]
Add font "index" to lmp parser.
When ttc index support was added, it was added and tested on Android
with the jb parser. This adds it to the lmp parser.
Review URL: https://codereview.chromium.org/
1023313002
reed [Mon, 23 Mar 2015 15:15:14 +0000 (08:15 -0700)]
remove meaningless matrix benches, add mapPts() and add new benches
mapPts definitely faster than mapPoints (identity and perspective same speed). Up to 3x for large values of N.
cloned from https://codereview.chromium.org/
1031443002/
BUG=skia:
Review URL: https://codereview.chromium.org/
1030653002
joshualitt [Mon, 23 Mar 2015 15:13:04 +0000 (08:13 -0700)]
Initial change to move text blob to GrTextContext
BUG=skia:
Review URL: https://codereview.chromium.org/
1019633002
tomhudson [Mon, 23 Mar 2015 14:46:13 +0000 (07:46 -0700)]
Get rid of excess cleverness in benchmark
RotatedRectBench was asking for its base layer size, which may
not be what it expects with odd canvas modes (particularly proxies).
Most benchmarks are not so sophisticated; they hard-wire their
size and just use that (expected) value.
R=mtklein@google.com,djsollen@google.com
BUG=skia:3566
Review URL: https://codereview.chromium.org/
1015013004
tomhudson [Mon, 23 Mar 2015 14:22:40 +0000 (07:22 -0700)]
Make SkCanvas::getBaseLayerSize() virtual
Subclasses of SkCanvas may need to override the behavior here - for
example, any proxy or deferred canvas may not know its own size and
need to delegate to another object.
We'll also work on reducing use of this function
(https://skbug.com/3569), but some of the current uses seem to be
semantically necessary.
R=reed@google.com
BUG=skia:3566
Review URL: https://codereview.chromium.org/
1022423002
kkinnunen [Mon, 23 Mar 2015 13:27:23 +0000 (06:27 -0700)]
Fix two typos in coding style
Fix two typos in coding style:
- Memory Managemt
- if (NULL == x) equated to if (x)
NOTRY=true
DOCS_PREVIEW= https://skia.org/?cl=
1032463002
Review URL: https://codereview.chromium.org/
1032463002
robertphillips [Mon, 23 Mar 2015 12:46:51 +0000 (05:46 -0700)]
Preserve texture compression when saving a revised skp from debugger
Without this CL, saving a revised skp out the debugger can greatly expand the memory it requires.
Review URL: https://codereview.chromium.org/
1020103005
kkinnunen [Mon, 23 Mar 2015 09:18:02 +0000 (02:18 -0700)]
Do not playback pending commands for full deferred canvas writePixels
Do not playback pending commands for full deferred canvas writePixels.
Changes the test to catch cases where discard is done without
a snapshot.
Review URL: https://codereview.chromium.org/
939103002
skia.buildbots [Sun, 22 Mar 2015 08:14:17 +0000 (01:14 -0700)]
Update SKP version
Automatic commit by the RecreateSKPs bot.
TBR=
Review URL: https://codereview.chromium.org/
1030453002
reed [Sat, 21 Mar 2015 02:41:10 +0000 (19:41 -0700)]
remove SK_SUPPORT_LEGACY_ARCTO_QUADS code
BUG=skia:
TBR=
Review URL: https://codereview.chromium.org/
1023103004
reed [Fri, 20 Mar 2015 22:25:25 +0000 (15:25 -0700)]
change default impl to facilitate transition in blink/chrome
BUG=skia:
TBR=
need deps roll to see if this works... since skia can no longer build/test in this legacy mode.
Review URL: https://codereview.chromium.org/
1026923002
reed [Fri, 20 Mar 2015 21:20:09 +0000 (14:20 -0700)]
remove SK_SUPPORT_LEGACY_PIXELREF_UNFLATTENABLE code
BUG=skia:
TBR=
Review URL: https://codereview.chromium.org/
1020403002
mtklein [Fri, 20 Mar 2015 20:48:42 +0000 (13:48 -0700)]
Some usability ideas around SkTHash.
- By default, use new SkGoodHash to hash keys, which is:
* for 4 byte values, use SkChecksum::Mix,
* for SkStrings, use SkChecksum::Murmur3 on the data,
* for other structs, shallow hash the struct with Murmur3.
- Expand SkChecksum::Murmur3 to support non-4-byte-aligned data.
- Add const foreach() methods.
- Have foreach() take a functor, which allows lambdas.
BUG=skia:
Review URL: https://codereview.chromium.org/
1021033002
reed [Fri, 20 Mar 2015 20:46:08 +0000 (13:46 -0700)]
cleanup using constructors
BUG=skia:
TBR=
Review URL: https://codereview.chromium.org/
1024193002
reed [Fri, 20 Mar 2015 20:23:43 +0000 (13:23 -0700)]
use Sk2s for conics
BUG=skia:
Review URL: https://codereview.chromium.org/
1025033002
mtklein [Fri, 20 Mar 2015 20:17:42 +0000 (13:17 -0700)]
Sk2x::invert() and Sk2x::approxInvert()
BUG=skia:
Review URL: https://codereview.chromium.org/
1024993002
jvanverth [Fri, 20 Mar 2015 19:35:07 +0000 (12:35 -0700)]
Return to old distance field glyph sizes.
Should fix some blurry or light glyphs on Windows.
Review URL: https://codereview.chromium.org/
1026613003
halcanary [Fri, 20 Mar 2015 19:31:52 +0000 (12:31 -0700)]
PDF: remove last use of SkPDFImage
Add a GM.
BUG=skia:255
Review URL: https://codereview.chromium.org/
950633003
mtklein [Fri, 20 Mar 2015 19:21:53 +0000 (12:21 -0700)]
Hack around skia:3570 for now.
BUG=skia:3570
Review URL: https://codereview.chromium.org/
1021353002
scroggo [Fri, 20 Mar 2015 18:55:55 +0000 (11:55 -0700)]
Add SkEncodedFormat, used by SkCodec.
Needed by Android to determine the format.
Review URL: https://codereview.chromium.org/
1018953003
mtklein [Fri, 20 Mar 2015 18:28:52 +0000 (11:28 -0700)]
Add divide to Sk2x, use native vdiv and vsqrt on ARM 64.
Tests pass on N7 + N9.
BUG=skia:
CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Mac10.7-Clang-Arm7-Debug-iOS-Trybot,Build-Ubuntu-GCC-Arm64-Release-Android-Trybot
Review URL: https://codereview.chromium.org/
1027753003
mtklein [Fri, 20 Mar 2015 18:20:12 +0000 (11:20 -0700)]
Allow negative error for Sk4f::sqrt() test.
BUG=skia:
Review URL: https://codereview.chromium.org/
1028743002
scroggo [Fri, 20 Mar 2015 18:20:02 +0000 (11:20 -0700)]
Make android framework depend on SkCodec.
This results in exporting include/codec so the framework can use it.
Review URL: https://codereview.chromium.org/
1020903002
joshualitt [Fri, 20 Mar 2015 17:30:14 +0000 (10:30 -0700)]
Let text contexts fall back directly to paths
BUG=skia:
Review URL: https://codereview.chromium.org/
1015173002
reed [Fri, 20 Mar 2015 17:03:36 +0000 (10:03 -0700)]
Revert of PDF: remove last use of SkPDFImage (patchset #5 id:120001 of https://codereview.chromium.org/
950633003/)
Reason for revert:
static void draw(SkCanvas* canvas,
const SkPaint& p,
const SkBitmap& src,
SkColorType colorType,
const char text[]) {
SkASSERT(src.colorType() == colorType);
canvas->drawBitmap(src, 0.0f, 0.0f);
canvas->drawText(text, strlen(text), 0.0f, 12.0f, p);
}
This assert is firing, at least on macs, where all images get decoded into 32bit at the moment.
Original issue's description:
> PDF: remove last use of SkPDFImage
>
> Add a GM.
>
> BUG=skia:255
>
> Committed: https://skia.googlesource.com/skia/+/
86ad8d643624a55b02e529100bbe4e2940115fa1
TBR=mtklein@google.com,halcanary@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:255
Review URL: https://codereview.chromium.org/
1024113002
mtklein [Fri, 20 Mar 2015 16:25:26 +0000 (09:25 -0700)]
Specialize Sk2d for ARM64
The implementation is nearly identical to Sk2f, with these changes:
- float32x2_t -> float64x2_t
- vfoo -> vfooq
- one extra Newton's method step in sqrt().
Also, generally fix NEON detection to be defined(SK_ARM_HAS_NEON).
SK_ARM_HAS_NEON is not being set on ARM64 bots right now (nor does the compiler
seem to set __ARM_NEON__), so this CL fixes everything up.
BUG=skia:
Committed: https://skia.googlesource.com/skia/+/
e57b5cab261a243dcbefa74c91c896c28959bf09
CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Mac10.7-Clang-Arm7-Debug-iOS-Trybot,Build-Ubuntu-GCC-Arm64-Release-Android-Trybot
Review URL: https://codereview.chromium.org/
1020963002
henrik.smiding [Fri, 20 Mar 2015 16:20:46 +0000 (09:20 -0700)]
Replace SSE optimization of Color32A_D565
Adds an SSE2 version of the Color32A_D565 function, to replace
the existing SSE4 version. Also does some minor cleanup.
Performance improvement in the following Skia benchmarks.
Measured on Atom Silvermont:
Xfermode_SrcOver - x3
luma_colorfilter_large - x4.6
luma_colorfilter_small - x2
tablebench - ~15%
chart_bw - ~10%
Measured on Corei7 Haswell:
luma_colorfilter_large running SSE2 - x2
luma_colorfilter_large running SSE4 - x2.3
Also improves performance in WPS Office application and 2D subtest of 0xbenchmark on Android.
Signed-off-by: Henrik Smiding <henrik.smiding@intel.com>
Review URL: https://codereview.chromium.org/
923523002
halcanary [Fri, 20 Mar 2015 16:10:56 +0000 (09:10 -0700)]
PDF: remove last use of SkPDFImage
Add a GM.
BUG=skia:255
Review URL: https://codereview.chromium.org/
950633003
senorblanco [Fri, 20 Mar 2015 15:54:32 +0000 (08:54 -0700)]
Add a new GM to show problems with shear/rotate CTM w/ image filters.
Review URL: https://codereview.chromium.org/
1028663002
mtklein [Fri, 20 Mar 2015 15:46:34 +0000 (08:46 -0700)]
Revert of Specialize Sk2d for ARM64 (patchset #3 id:40001 of https://codereview.chromium.org/
1020963002/)
Reason for revert:
https://uberchromegw.corp.google.com/i/client.skia.compile/builders/Build-Mac10.7-Clang-Arm7-Debug-iOS/builds/2441/steps/build%20most/logs/stdio
https://uberchromegw.corp.google.com/i/client.skia.compile/builders/Build-Mac10.7-Clang-Arm7-Release-iOS/builds/2424/steps/build%20most/logs/stdio
https://uberchromegw.corp.google.com/i/client.skia.compile/builders/Build-Ubuntu-GCC-Arm64-Release-Android/builds/8/steps/build%20most/logs/stdio
Original issue's description:
> Specialize Sk2d for ARM64
>
> The implementation is nearly identical to Sk2f, with these changes:
> - float32x2_t -> float64x2_t
> - vfoo -> vfooq
> - one extra Newton's method step in sqrt().
>
> Also, generally fix NEON detection to be defined(SK_ARM_HAS_NEON).
> SK_ARM_HAS_NEON is not being set on ARM64 bots right now (nor does the compiler
> seem to set __ARM_NEON__), so this CL fixes everything up.
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/
e57b5cab261a243dcbefa74c91c896c28959bf09
TBR=msarett@google.com,reed@google.com,mtklein@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review URL: https://codereview.chromium.org/
1028523003
mtklein [Fri, 20 Mar 2015 15:32:35 +0000 (08:32 -0700)]
Specialize Sk2d for ARM64
The implementation is nearly identical to Sk2f, with these changes:
- float32x2_t -> float64x2_t
- vfoo -> vfooq
- one extra Newton's method step in sqrt().
Also, generally fix NEON detection to be defined(SK_ARM_HAS_NEON).
SK_ARM_HAS_NEON is not being set on ARM64 bots right now (nor does the compiler
seem to set __ARM_NEON__), so this CL fixes everything up.
BUG=skia:
Review URL: https://codereview.chromium.org/
1020963002
Mike Klein [Fri, 20 Mar 2015 15:19:01 +0000 (11:19 -0400)]
Fix ambiguity in negate()s.
BUG=skia:
Review URL: https://codereview.chromium.org/
1027793002
mtklein [Fri, 20 Mar 2015 15:12:06 +0000 (08:12 -0700)]
Sk2x/Sk4x: add negate(), operator -()
BUG=skia:
Review URL: https://codereview.chromium.org/
1027783002
egdaniel [Fri, 20 Mar 2015 14:03:52 +0000 (07:03 -0700)]
Only use 256x256 tiles on hd2000 nanobench bots
Initial experiments did show that the 256 tile size fixed the hd2000 win7
nanobot failures. However it did not have any effect on other bots, so this
change is to move back to the larger tile size on all bots expect for the
hd2000.
BUG=skia:
Review URL: https://codereview.chromium.org/
1022083002
robertphillips [Fri, 20 Mar 2015 13:43:11 +0000 (06:43 -0700)]
Disable sharing of stencil buffers across differently sized rendertargets
Due to revertapalooza, stencil buffers shared across differently sized rendertargets aren't getting cleared correctly. This CL disables the sharing until the clearing issues can be remedied. Note that stencil buffers should still be shared between identically sized render targets and should still be lazily allocated.
Review URL: https://codereview.chromium.org/
1020203002
senorblanco [Fri, 20 Mar 2015 13:38:17 +0000 (06:38 -0700)]
Move SkMatrixImageFilter into core, and add a factory fn for it.
BUG=skia:
Review URL: https://codereview.chromium.org/
1011273003
senorblanco [Fri, 20 Mar 2015 13:38:00 +0000 (06:38 -0700)]
Remove deprecated flavour of SkDropShadowImageFilter::Create().
Review URL: https://codereview.chromium.org/
1026603002
mtklein [Fri, 20 Mar 2015 13:33:02 +0000 (06:33 -0700)]
Guard SIMD code with !defined(SKNX_NO_SIMD).
This should make it easy to compare performance of the non-SIMD Sk2x / Sk4x
code with our existing portable scalar code. I'm not adding this to SkPMFloat
only because we don't have an existing scalar baseline there to compare to.
We'll have to keep our wits about us: I just tried your new benchmarks, and
Clang's autovectorizer produced almost as good SSE as we did with intrinsics for
geo_evalquadat1 and geo_evalquadtangentat1, but not for geo_chopquadat1,
which went serial.
BUG=skia:
Review URL: https://codereview.chromium.org/
1026723003
reed [Fri, 20 Mar 2015 13:32:52 +0000 (06:32 -0700)]
disable LCD for layers w/ filters
BUG=468311
see also skiabug.com/3567
Review URL: https://codereview.chromium.org/
1002603003
reed [Fri, 20 Mar 2015 13:01:08 +0000 (06:01 -0700)]
use Sk2s for EvalQuadTangent and ChopQuadAt
cloned from https://codereview.chromium.org/
1026633002/
BUG=skia:
Review URL: https://codereview.chromium.org/
1024873003
mtklein [Fri, 20 Mar 2015 13:00:57 +0000 (06:00 -0700)]
operator overloads for Sk4x, use them all where possible
BUG=skia:
NOTRY=true
Review URL: https://codereview.chromium.org/
1024633003
reed [Thu, 19 Mar 2015 23:10:54 +0000 (16:10 -0700)]
add SkMatrix::decomposeScale
BUG=skia:
Review URL: https://codereview.chromium.org/
1006173005
mtklein [Thu, 19 Mar 2015 21:28:31 +0000 (14:28 -0700)]
Try out operator overloads for Sk2x.
BUG=skia:
Review URL: https://codereview.chromium.org/
1024473005
bungeman [Thu, 19 Mar 2015 20:57:36 +0000 (13:57 -0700)]
Support BGR with DirectWrite.
BUG=chromium:468509
Review URL: https://codereview.chromium.org/
1018863003
senorblanco [Thu, 19 Mar 2015 20:39:19 +0000 (13:39 -0700)]
Remove now-unused uniqueID param from SkDropShadowImageFilter::Create().
Also make both input & crop rect default NULL, as in other filter Create functions.
BUG=skia:
Review URL: https://codereview.chromium.org/
1016343002
egdaniel [Thu, 19 Mar 2015 20:09:17 +0000 (13:09 -0700)]
Blacklist certain win7 gms that are never finishing on dm
BUG=skia:
Review URL: https://codereview.chromium.org/
1013483004
mtklein [Thu, 19 Mar 2015 19:08:29 +0000 (12:08 -0700)]
Sk2x for NEON
Also decreases the precision of Sk4f::rsqrt() for speed, keeping Sk4f::sqrt() the same:
instead of doing two estimation steps in rsqrt(), do one there and one more in sqrt().
Tests pass on my Nexus 7. float64x2_t is still a TODO for when I get a hold of a Nexus 9.
BUG=skia:
Review URL: https://codereview.chromium.org/
1018423003
mtklein [Thu, 19 Mar 2015 19:03:29 +0000 (12:03 -0700)]
Port to real Sk2f.
The bench improves from 39 to 30, about half from porting to Sk2f, half from
x.add(x) instead of x.multiply(two).
Remove Sk4f Load2/store2 now that we have Sk2f.
BUG=skia:
Review URL: https://codereview.chromium.org/
1019773004
reed [Thu, 19 Mar 2015 18:37:18 +0000 (11:37 -0700)]
remove unneeded SK_SUPPORT_LEGACY_MIPMAP_EFFECTIVE_SCALE guard
BUG=skia:
TBR=
Review URL: https://codereview.chromium.org/
1019343002
jvanverth [Thu, 19 Mar 2015 18:34:13 +0000 (11:34 -0700)]
Add minimum size for using distance fields.
Since we're not scaling down as far, we can increase the base
sizes a little.
Also darken debug text colors to get better value comparison
with black.
BUG=skia:3541
Review URL: https://codereview.chromium.org/
1001183003
egdaniel [Thu, 19 Mar 2015 18:00:09 +0000 (11:00 -0700)]
Move skp nanobench tile size back to 256x256 to see if it fixes nanobench crashes
Going back to old nanobench tile size to see if the increase to tile is what has been
causing recent nanobench crashes. The crashes seem very nondeterministic and hard to
debug manually.
256x256 is too small of a tile to give accurate gpu results but if this fixes we can try some compromise in the middle
BUG=skia:
Review URL: https://codereview.chromium.org/
1022823003
mtklein [Thu, 19 Mar 2015 17:44:27 +0000 (10:44 -0700)]
Sk2x
This adds an API, an SSE impl, a portable impl, and some tests for Sk2f/Sk2d/Sk2s.
BUG=skia:
Review URL: https://codereview.chromium.org/
1025463002
bungeman [Thu, 19 Mar 2015 17:43:57 +0000 (10:43 -0700)]
Use typographic font metrics when the font requests.
FreeType always returns the 'hhea' font metrics for ascent and descent,
and ignores the 'OS/2'::fsSelection::UseTypoMetrics bit. (It also
ignores the VDMX table, which makes this change correct.) This change
uses the typographic font metrics when the font requests their use.
Review URL: https://codereview.chromium.org/
1020643002
mtklein [Thu, 19 Mar 2015 17:19:48 +0000 (10:19 -0700)]
Reorg Sk4x to match the pattern of SkPMFloat.
No real changes here, just moving files around:
- move impl files into src/opts
- rename _portable _none
BUG=skia:
Review URL: https://codereview.chromium.org/
1021713004
reed [Thu, 19 Mar 2015 17:18:47 +0000 (10:18 -0700)]
alt SkEvalQuadAt that returns its answer, using Sk2f
BUG=skia:
Review URL: https://codereview.chromium.org/
1011493003
reed [Thu, 19 Mar 2015 16:49:09 +0000 (09:49 -0700)]
remove unused clip parameter to SkEdge::setClip
BUG=skia:
Review URL: https://codereview.chromium.org/
1019183002
reed [Thu, 19 Mar 2015 15:31:14 +0000 (08:31 -0700)]
guarded change to SkImageGenerator to make getInfo() const
BUG=skia:
Review URL: https://codereview.chromium.org/
1017293002
rmistry [Thu, 19 Mar 2015 14:56:12 +0000 (07:56 -0700)]
Whitespace change to test the CQ
BUG=skia:
TBR=
Review URL: https://codereview.chromium.org/
1019313002
jvanverth [Thu, 19 Mar 2015 13:08:31 +0000 (06:08 -0700)]
Ensure that we use different glyph entries for regular and df text.
Currently if we switch between regular text and df text while using
the same GrContext, they may use the same entry in the Ganesh font cache,
which is incorrect. This change ensures that they will have different entries.
Committed: https://skia.googlesource.com/skia/+/
8dc58edd71c11f232860724dfa3b566895478034
Review URL: https://codereview.chromium.org/
1020593003
scroggo [Thu, 19 Mar 2015 13:03:39 +0000 (06:03 -0700)]
Run CodecSrc DM.
Rather than making SkCodec an option instead of SkImageDecoder,
create a separate CodecSrc. This allows us to compare the two.
For both CodecSrc and ImageSrc, do not decode to a gpu backend.
BUG=skia:3475
Review URL: https://codereview.chromium.org/
978823002
reed [Thu, 19 Mar 2015 11:10:42 +0000 (04:10 -0700)]
Revert of replace SkFixedDiv impl with native 64bit math (patchset #2 id:20001 of https://codereview.chromium.org/
1022543003/)
Reason for revert:
http://build.chromium.org/p/tryserver.blink/builders/linux_blink_rel/builds/53096
layouttests failures
Original issue's description:
> replace SkFixedDiv impl with native 64bit math
>
> BUG=skia:
> TBR=
>
> Committed: https://skia.googlesource.com/skia/+/
7c44ca926bf42b3b2e56131f250c0fd58f87ac71
TBR=
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review URL: https://codereview.chromium.org/
1018523008
reed [Thu, 19 Mar 2015 02:32:47 +0000 (19:32 -0700)]
remove unused SkXRay functions
BUG=skia:
TBR=
Review URL: https://codereview.chromium.org/
1016263002
reed [Thu, 19 Mar 2015 02:04:43 +0000 (19:04 -0700)]
replace SkFixedDiv impl with native 64bit math
BUG=skia:
TBR=
Review URL: https://codereview.chromium.org/
1022543003
jvanverth [Wed, 18 Mar 2015 22:14:05 +0000 (15:14 -0700)]
Revert of Ensure that we use different glyph entries for regular and df text. (patchset #2 id:20001 of https://codereview.chromium.org/
1020593003/)
Reason for revert:
Bad glyphs in dftext GM.
Original issue's description:
> Ensure that we use different glyph entries for regular and df text.
>
> Currently if we switch between regular text and df text while using
> the same GrContext, they may use the same entry in the Ganesh font cache,
> which is incorrect. This change ensures that they will have different entries.
>
> Committed: https://skia.googlesource.com/skia/+/
8dc58edd71c11f232860724dfa3b566895478034
TBR=joshualitt@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.chromium.org/
1011403005
jvanverth [Wed, 18 Mar 2015 21:46:38 +0000 (14:46 -0700)]
Ensure that we use different glyph entries for regular and df text.
Currently if we switch between regular text and df text while using
the same GrContext, they may use the same entry in the Ganesh font cache,
which is incorrect. This change ensures that they will have different entries.
Review URL: https://codereview.chromium.org/
1020593003
mtklein [Wed, 18 Mar 2015 21:06:55 +0000 (14:06 -0700)]
temporarily stifle Math test on TSAN bot
BUG=skia:3562
Review URL: https://codereview.chromium.org/
1017283002