Bug: skia:12524
Change-Id: I99ade4b9f4fd442d9ddaa51651faf47321c06a5b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/474296
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
This is redundant information now that we have SkVMSlotInfo.
Change-Id: Ia05b1eaa722023e719042c83255708aa9debed61
Bug: skia:12614
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473777
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
* Remove 3 iPad Pro 9.7"s killed in the March 2020 Skolo flood.
* Add skia-iphone-8-03, which was mysteriously missing.
Change-Id: Idf3bbc553958a1327b0909183f85ab43dced3429
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/474020
Reviewed-by: Leandro Lovisolo <lovisolo@google.com>
Commit-Queue: Erik Rose <erikrose@google.com>
This adds new validation rules that we were breaking.
Binding and DescriptorSet can't be applied to push constants, nor to
struct members.
Bug: skia:12670
Bug: chromium:1270328
Change-Id: I332f77717b08d9945c8e5b79c5bf649a8f5f2043
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/474056
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
If a FreeType font has glyphs which may need the current color this is
non-changing information, so calculate it once at and cache it. This
information is needed on evert time a scaler context rec is created and
adding the overhead of looking through the font data (especially the
extra indirections from a ttc) can take a bit of time.
Bug: b/206120675
Change-Id: I92f2083a6bbd673fc148347c4e04ed81a441c3ba
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/474059
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>
Previously, while BindBufferInfo had the right offset for binding the
buffers, the writer would always overwrite the contents at the
beginning. Never encountered it because the CommandBufferTest made its
VertexWriters directly, and DrawPass isn't end-to-end yet.
Bug: skia:12466
Change-Id: Ib874006d80a7286fcc3710cd38b3c6141ff928ef
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/474057
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Auto-Submit: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Update the check for situations using the direct glyph drawing
method. Mostly renaming.
Change-Id: I9e2ad8b669e9868b5fffd4864fb4551fec11f19c
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/474038
Commit-Queue: Herb Derby <herb@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Bug: skia:12559
Change-Id: I76b225c9ca81264a15869324007d774d210053b5
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473416
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
Bug: b/205516620
Internally, Skia does not have full GL support for protected backend
textures. But there are some clients who know that they need to create
one so that they can render a protected buffer into it. Plumb in support
for clients calling GrDirectContext::createBackendTexture.
GrGLDefines.h:
- Add macros for querying whether a context is protected
GrGLCaps:
- Check whether the GL_EXT_protected_textures extension is present and
the context supports protected.
GrGLGpu:
- Remove failure cases for GrProtected::kYes. Instead, pass the value
down to createTexture where we then set the parameter in GL.
Change-Id: I465442fc4b19ed32a65079a168124a067bcd66a6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/472802
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
Change-Id: Idcedaf97d9036fc892bbfc4fa7be468dd7a11db9
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473660
Commit-Queue: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: Brian Osman <brianosman@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
Previously, we would avoid emitting redundant trace_vars by checking to
see if the slot is being assigned to the exact same Var.ID. However,
this had the potential to eliminate useful trace_vars:
- At the start of execution, slots all contain 0. Code which explicitly
assigned a zero into a slot would not be shown in a trace. (So things
like `color = half4(0,1,0,1)` would only emit traces for color.ga.)
- A function call's parameter slots are reused every time it is called,
so calling a function twice would only emit traces for the parameters
that aren't the same Val.ID as the previous call.
- A VarDeclaration inside a loop reuses its slot each time through the
loop, even though conceptually it's a "new" variable.
We now track a slot's "written-to" status. At the start of execution,
no slots have been "written-to". These slots will always emit a
trace_var opcode (fixing the first issue). Also, issuing a function
call or declaring a variable will reset the "written-to" status of the
associated slots (fixing the second and third issues).
When the debugger is not in use, the written-to field is unused.
Change-Id: I482a86cb6e90d0f85dd2a161e984f212782a7b4d
Bug: skia:12614
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473776
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Bug: skia:12662
Change-Id: Ic18220668a4f87e7340a53b3f191887a7a016a04
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473141
Reviewed-by: Brian Osman <brianosman@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Bug: 1265199
Change-Id: I217d75377c33940cdbabba278ee1ac06e94e177f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/469899
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
We now canonicalize commutative operations by ordering their value IDs.
The lower-numbered value ID is always placed first into a commutative
instruction. In other words, this instruction:
bit_and result, v7, v5
Would be silently converted to this:
bit_and result, v5, v7
This will allow these two logically-equivalent instructions to be
deduplicated:
bit_and result, v7, v5
bit_and result, v5, v7
Of course, deduplicating these ops can unlock additional free CSE/DCE.
The affected instructions are listed in http://review.skia.org/473238
Change-Id: Ib9beb79d6b72d7903184aaa9a53e8e5a02ae126d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473239
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Make the code more consistent with:
* drawMatrix - CTM from canvas
* drawOrigin - x, y from drawTextBlob
* positionMatrix drawMatrix adjusted by drawOrigin
Change-Id: Ia22baa2d229f041b77dc6cde063e25f185b8b64a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473237
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Herb Derby <herb@google.com>
This is a reland of 688cb15faa
Original change's description:
> Remove layer-coverage-tracking experiment
>
> This effectively reverts reviews.skia.org/122000
>
> Bug: skia:10987
> Change-Id: I989241110f17c0e3c2a896aea4bc2bc4cc8c910f
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/472801
> Reviewed-by: Brian Salomon <bsalomon@google.com>
> Commit-Queue: Brian Osman <brianosman@google.com>
Bug: skia:10987
Change-Id: I16493df8bd7942261d14c01747d0fdc91bbe5467
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473143
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This will hopefully let us pre-package certain binaries (e.g. DM,
fuzz) with more sensible defaults and not make the developer
type out all the settings.
For CanvasKit, which specifies its own build flags, I think I'll
need to make another transition setup, which would go in something
like modules/canvaskit/ck_binary_with_flags.bzl or something.
Some sausage-case-names were converted to snake_case_names as per
go/build-style#target-naming
The example this is based off is worth a look through before
diving into this:
https://github.com/bazelbuild/examples/tree/main/rules/starlark_configurations/cc_binary_selectable_copts
Change-Id: Ia919d47f4d1aa25cf294af7918e36d38838c179e
Bug: skia:12541
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/472688
Reviewed-by: Ben Wagner <bungeman@google.com>
Reviewed-by: Leandro Lovisolo <lovisolo@google.com>
Bug: skia:12657
Change-Id: I432dc004ac2c8a1f9b0fd0874b8690dcb7e16dbf
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473146
Reviewed-by: Eric Boren <borenet@google.com>
Commit-Queue: Ravi Mistry <rmistry@google.com>
To pick up canary.go task driver change in buildbot repo
Updated with:
$ go get go.skia.org/infra@74751ee0f1
$ go mod download
$ make -C infra/bots train
Bug: skia:12657
Change-Id: I3f42dd6a8041d2c84903e736e76029c198c811f7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473144
Reviewed-by: Eric Boren <borenet@google.com>
This verifies that compiled SkSL code generates a reasonable trace.
Change-Id: Ia81a694460eaac261dfb9287b64f0afec013f76c
Bug: skia:12614
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473145
Commit-Queue: John Stiles <johnstiles@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
This provides a callback mechanism for the interpreter when a trace
opcode is encountered. The callbacks are only invoked when the trace
mask is enabled.
Change-Id: I55db22e18106ae09e4ab0a503533d830282c772c
Bug: skia:12614
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473139
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Exceptionally large control points might require us to chop the path
very deeply. This CL converts the recursive chopping to loop with a
stack of control points on the heap, in order to avoid the risk of
stack overflow. It also adds a bail condition to avoid getting stuck
in endless recursion due to fp32 precision issues.
Bug: chromium:1266446
Change-Id: I005be4bc29de51d3c89f04b5d6c553a921a92aa3
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473197
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Since we outset the viewport by stroke width for pre-chopping,
astronomically wide strokes can result in an astronomical viewport
size, and therefore an exponential explosion chops and memory usage.
It is also simply inefficient to tessellate these strokes due to the
number of radial edges required. We're better off just converting them
to a path after a certain point
Bug: chromium:1266446
Change-Id: I23ea39b0bd64f22d4e293a881992e3669afbe530
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473196
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
This is a first step in making GrShaderCaps more general so we can use
it in Graphite.
Change-Id: I7fc874e5df8dba2cef05421840a4019e5499ab59
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473000
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>
If the device coordinates are constant across every lane, debug traces
can't generate reasonable results. If nothing actually uses the device
coordinates, the extra opcodes should DCE away.
Change-Id: Ic08a490a692e9c0bdd94f351193bc590d6540ded
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473136
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Trace opcodes will only be implemented in the interpreter. If we
encounter one while JITting a program, returning false will cause
execution to fall back to the interpreter automatically.
Change-Id: I94fce376cb7336c16072a174604de025f1b4e38a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473137
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Reviewed-by: Herb Derby <herb@google.com>
This was buried inside SkVM, but can be useful in other contexts as
well.
Change-Id: If5b7dbca24b5c3fb8d4a4f136421c9ddb2f005d8
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/473138
Auto-Submit: John Stiles <johnstiles@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Reviewed-by: Herb Derby <herb@google.com>
This effectively reverts reviews.skia.org/122000
Bug: skia:10987
Change-Id: I989241110f17c0e3c2a896aea4bc2bc4cc8c910f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/472801
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This associates the input SkSL source with our debug info. It is
serialized out to .trace files. This will actually let us display the
code associated with `trace_line`.
Bug: skia:12614
Change-Id: I6c10af6fa6dd97ed8411f3460e103f442722b7f6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/472999
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
The SkVMDebugInfo now includes an skvm::Coord that the caller is
responsible for filling in before converting the program. If it is set,
the value indicates the device coordinates that should be traced. If it
is unset, debug traces will not be emitted at all.
Within the SkVMCodeGenerator, we now have a new traceMask() call which
combines the current execution mask with the trace mask. Tracing opcodes
now pass the result of traceMask() instead of mask(). This will limit
trace data to the selected pixel, instead of tracing the entire draw.
Bug: skia:12614
Change-Id: I1f8ce7d18a31aa60a139a4a7335c2a285d6aee60
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/472798
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
We no longer indicate functions by their line number, which can be
ambiguous. The debug info now includes a list of function names which we
can refer to by index, and the `trace_call` opcode references functions
by their index in this list.
Change-Id: I4bdf2a6439637a07b116831fd2981f342c19ea4a
Bug: skia:12614
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/472796
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
We were inconsistent about special treatment for kAlpha_8 vs. other
alpha-only color types (other code paths were already adjusted).
Hopefully, we can remove this special behavior entirely, but as long
as we have this code, let's be somewhat principled.
Change-Id: I373a3c5e8c917d911b55b265631ca964c780a31d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/472996
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This makes behavior (more) consistent with the GPU. In particular, the
alpha_image GM now looks roughly identical (it was very different
before).
Bug: skia:9692
Change-Id: I405375760ffe90577f863fa2d8eb9d2fc0f6edf8
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/472799
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Bug: skia:12633
Change-Id: I3e4542e560a2c876b4b35bf340bdb949c4922c65
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/472556
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
Draws that take a primitive with colors (vertices, patch, atlas) allow
the user to specify how the primitive color is blended with SkShader.
At the public level the blend is specified as SkBlendMode. This change
moves the implementation at SkDevice level over to SkBlender.
Future changes to the existing APIs or future APIs could allow clients
to specify SkBlenders instead of SkBlendMode. THis would would allow for
programmable blending between shader and primitive color.
SkShader_Blend now recognizes if it was constructed with a Mode blender
and extracts the mode enum.
SkBlender_Base::asFragmentProcessor no longer assumes that the dstFP's
input color should be args.fDestColor. Instead the caller
must configure that if desired. (We don't want this for drawAtlas, etc).
Change-Id: I44085da4ec9f4873428f38d629565f7f42c6be91
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/470458
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
This fails exactly as it should, but we had no test for it.
Change-Id: I0aa3307c444f2c9bc3512ff43b784a56a7c09856
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/472449
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
