You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

596 lines
22 KiB

/*
* Copyright 2011 Google Inc.
*
* Use of this source code is governed by a BSD-style license that can be
* found in the LICENSE file.
*/
#include "include/core/SkBitmap.h"
#include "include/core/SkCanvas.h"
#include "include/core/SkSurface.h"
#include "include/gpu/GrBackendSurface.h"
#include "include/gpu/GrDirectContext.h"
#include "include/private/SkColorData.h"
#include "include/private/SkImageInfoPriv.h"
#include "src/core/SkMathPriv.h"
#include "src/gpu/GrDirectContextPriv.h"
#include "src/gpu/GrGpu.h"
#include "src/gpu/GrProxyProvider.h"
#include "tests/Test.h"
#include "tools/gpu/BackendSurfaceFactory.h"
#include <initializer_list>
static const int DEV_W = 100, DEV_H = 100;
static const SkIRect DEV_RECT = SkIRect::MakeWH(DEV_W, DEV_H);
static const U8CPU DEV_PAD = 0xee;
static SkPMColor get_canvas_color(int x, int y) {
SkASSERT(x >= 0 && x < DEV_W);
SkASSERT(y >= 0 && y < DEV_H);
U8CPU r = x;
U8CPU g = y;
U8CPU b = 0xc;
U8CPU a = 0x0;
switch ((x+y) % 5) {
case 0:
a = 0xff;
break;
case 1:
a = 0x80;
break;
case 2:
a = 0xCC;
break;
case 3:
a = 0x00;
break;
case 4:
a = 0x01;
break;
}
return SkPremultiplyARGBInline(a, r, g, b);
}
// assumes any premu/.unpremul has been applied
static uint32_t pack_color_type(SkColorType ct, U8CPU a, U8CPU r, U8CPU g, U8CPU b) {
uint32_t r32;
uint8_t* result = reinterpret_cast<uint8_t*>(&r32);
switch (ct) {
case kBGRA_8888_SkColorType:
result[0] = b;
result[1] = g;
result[2] = r;
result[3] = a;
break;
case kRGBA_8888_SkColorType: // fallthrough
case kRGB_888x_SkColorType:
result[0] = r;
result[1] = g;
result[2] = b;
result[3] = a;
break;
default:
SkASSERT(0);
return 0;
}
return r32;
}
static uint32_t get_bitmap_color(int x, int y, int w, SkColorType ct, SkAlphaType at) {
int n = y * w + x;
U8CPU b = n & 0xff;
U8CPU g = (n >> 8) & 0xff;
U8CPU r = (n >> 16) & 0xff;
U8CPU a = 0;
switch ((x+y) % 5) {
case 4:
a = 0xff;
break;
case 3:
a = 0x80;
break;
case 2:
a = 0xCC;
break;
case 1:
a = 0x01;
break;
case 0:
a = 0x00;
break;
}
if (kPremul_SkAlphaType == at) {
r = SkMulDiv255Ceiling(r, a);
g = SkMulDiv255Ceiling(g, a);
b = SkMulDiv255Ceiling(b, a);
}
return pack_color_type(ct, a, r, g , b);
}
static void fill_surface(SkSurface* surface) {
SkBitmap bmp;
bmp.allocN32Pixels(DEV_W, DEV_H);
for (int y = 0; y < DEV_H; ++y) {
for (int x = 0; x < DEV_W; ++x) {
*bmp.getAddr32(x, y) = get_canvas_color(x, y);
}
}
surface->writePixels(bmp, 0, 0);
}
/**
* Lucky for us, alpha is always in the same spot (SK_A32_SHIFT), for both RGBA and BGRA.
* Thus this routine doesn't need to know the exact colortype
*/
static uint32_t premul(uint32_t color) {
unsigned a = SkGetPackedA32(color);
// these next three are not necessarily r,g,b in that order, but they are r,g,b in some order.
unsigned c0 = SkGetPackedR32(color);
unsigned c1 = SkGetPackedG32(color);
unsigned c2 = SkGetPackedB32(color);
c0 = SkMulDiv255Ceiling(c0, a);
c1 = SkMulDiv255Ceiling(c1, a);
c2 = SkMulDiv255Ceiling(c2, a);
return SkPackARGB32NoCheck(a, c0, c1, c2);
}
static SkPMColor convert_to_PMColor(SkColorType ct, SkAlphaType at, uint32_t color) {
if (kUnpremul_SkAlphaType == at) {
color = premul(color);
}
switch (ct) {
case kRGBA_8888_SkColorType: // fallthrough
case kRGB_888x_SkColorType:
color = SkSwizzle_RGBA_to_PMColor(color);
break;
case kBGRA_8888_SkColorType:
color = SkSwizzle_BGRA_to_PMColor(color);
break;
default:
SkASSERT(0);
break;
}
return color;
}
static bool check_pixel(SkPMColor a, SkPMColor b, bool didPremulConversion) {
if (!didPremulConversion) {
return a == b;
}
int32_t aA = static_cast<int32_t>(SkGetPackedA32(a));
int32_t aR = static_cast<int32_t>(SkGetPackedR32(a));
int32_t aG = static_cast<int32_t>(SkGetPackedG32(a));
int32_t aB = SkGetPackedB32(a);
int32_t bA = static_cast<int32_t>(SkGetPackedA32(b));
int32_t bR = static_cast<int32_t>(SkGetPackedR32(b));
int32_t bG = static_cast<int32_t>(SkGetPackedG32(b));
int32_t bB = static_cast<int32_t>(SkGetPackedB32(b));
return aA == bA &&
SkAbs32(aR - bR) <= 1 &&
SkAbs32(aG - bG) <= 1 &&
SkAbs32(aB - bB) <= 1;
}
bool write_should_succeed(const SkImageInfo& dstInfo, const SkImageInfo& srcInfo, bool isGPU) {
if (!SkImageInfoValidConversion(dstInfo, srcInfo)) {
return false;
}
if (!isGPU) {
return true;
}
// The GPU backend supports writing unpremul data to a premul dst but not vice versa.
if (srcInfo.alphaType() == kPremul_SkAlphaType &&
dstInfo.alphaType() == kUnpremul_SkAlphaType) {
return false;
}
if (!SkColorTypeIsAlwaysOpaque(srcInfo.colorType()) &&
SkColorTypeIsAlwaysOpaque(dstInfo.colorType())) {
return false;
}
// The source has no alpha value and the dst is only alpha
if (SkColorTypeIsAlwaysOpaque(srcInfo.colorType()) &&
SkColorTypeIsAlphaOnly(dstInfo.colorType())) {
return false;
}
return true;
}
static bool check_write(skiatest::Reporter* reporter, SkSurface* surf, SkAlphaType surfaceAlphaType,
const SkBitmap& bitmap, int writeX, int writeY) {
size_t canvasRowBytes;
const uint32_t* canvasPixels;
// Can't use canvas->peekPixels(), as we are trying to look at GPU pixels sometimes as well.
// At some point this will be unsupported, as we won't allow accessBitmap() to magically call
// readPixels for the client.
SkBitmap secretDevBitmap;
secretDevBitmap.allocN32Pixels(surf->width(), surf->height());
if (!surf->readPixels(secretDevBitmap, 0, 0)) {
return false;
}
canvasRowBytes = secretDevBitmap.rowBytes();
canvasPixels = static_cast<const uint32_t*>(secretDevBitmap.getPixels());
if (nullptr == canvasPixels) {
return false;
}
if (surf->width() != DEV_W || surf->height() != DEV_H) {
return false;
}
const SkImageInfo& bmInfo = bitmap.info();
SkIRect writeRect = SkIRect::MakeXYWH(writeX, writeY, bitmap.width(), bitmap.height());
for (int cy = 0; cy < DEV_H; ++cy) {
for (int cx = 0; cx < DEV_W; ++cx) {
SkPMColor canvasPixel = canvasPixels[cx];
if (writeRect.contains(cx, cy)) {
int bx = cx - writeX;
int by = cy - writeY;
uint32_t bmpColor8888 = get_bitmap_color(bx, by, bitmap.width(),
bmInfo.colorType(), bmInfo.alphaType());
bool mul = (kUnpremul_SkAlphaType == bmInfo.alphaType());
SkPMColor bmpPMColor = convert_to_PMColor(bmInfo.colorType(), bmInfo.alphaType(),
bmpColor8888);
if (bmInfo.alphaType() == kOpaque_SkAlphaType ||
surfaceAlphaType == kOpaque_SkAlphaType) {
bmpPMColor |= 0xFF000000;
}
if (!check_pixel(bmpPMColor, canvasPixel, mul)) {
ERRORF(reporter, "Expected canvas pixel at %d, %d to be 0x%08x, got 0x%08x. "
"Write performed premul: %d", cx, cy, bmpPMColor, canvasPixel, mul);
return false;
}
} else {
SkPMColor testColor = get_canvas_color(cx, cy);
if (canvasPixel != testColor) {
ERRORF(reporter, "Canvas pixel outside write rect at %d, %d changed."
" Should be 0x%08x, got 0x%08x. ", cx, cy, testColor, canvasPixel);
return false;
}
}
}
if (cy != DEV_H -1) {
const char* pad = reinterpret_cast<const char*>(canvasPixels + DEV_W);
for (size_t px = 0; px < canvasRowBytes - 4 * DEV_W; ++px) {
bool check;
REPORTER_ASSERT(reporter, check = (pad[px] == static_cast<char>(DEV_PAD)));
if (!check) {
return false;
}
}
}
canvasPixels += canvasRowBytes/4;
}
return true;
}
#include "include/core/SkMallocPixelRef.h"
// This is a tricky pattern, because we have to setConfig+rowBytes AND specify
// a custom pixelRef (which also has to specify its rowBytes), so we have to be
// sure that the two rowBytes match (and the infos match).
//
static bool alloc_row_bytes(SkBitmap* bm, const SkImageInfo& info, size_t rowBytes) {
if (!bm->setInfo(info, rowBytes)) {
return false;
}
sk_sp<SkPixelRef> pr = SkMallocPixelRef::MakeAllocate(info, rowBytes);
bm->setPixelRef(std::move(pr), 0, 0);
return true;
}
static void free_pixels(void* pixels, void* ctx) {
sk_free(pixels);
}
static bool setup_bitmap(SkBitmap* bm, SkColorType ct, SkAlphaType at, int w, int h, int tightRB) {
size_t rowBytes = tightRB ? 0 : 4 * w + 60;
SkImageInfo info = SkImageInfo::Make(w, h, ct, at);
if (!alloc_row_bytes(bm, info, rowBytes)) {
return false;
}
for (int y = 0; y < h; ++y) {
for (int x = 0; x < w; ++x) {
*bm->getAddr32(x, y) = get_bitmap_color(x, y, w, ct, at);
}
}
return true;
}
static void call_writepixels(SkSurface* surface) {
const SkImageInfo info = SkImageInfo::MakeN32Premul(1, 1);
SkPMColor pixel = 0;
surface->writePixels({info, &pixel, sizeof(SkPMColor)}, 0, 0);
}
DEF_TEST(WritePixelsSurfaceGenID, reporter) {
const SkImageInfo info = SkImageInfo::MakeN32Premul(100, 100);
auto surface(SkSurface::MakeRaster(info));
uint32_t genID1 = surface->generationID();
call_writepixels(surface.get());
uint32_t genID2 = surface->generationID();
REPORTER_ASSERT(reporter, genID1 != genID2);
}
static void test_write_pixels(skiatest::Reporter* reporter, SkSurface* surface,
const SkImageInfo& surfaceInfo) {
const SkIRect testRects[] = {
// entire thing
DEV_RECT,
// larger on all sides
SkIRect::MakeLTRB(-10, -10, DEV_W + 10, DEV_H + 10),
// fully contained
SkIRect::MakeLTRB(DEV_W / 4, DEV_H / 4, 3 * DEV_W / 4, 3 * DEV_H / 4),
// outside top left
SkIRect::MakeLTRB(-10, -10, -1, -1),
// touching top left corner
SkIRect::MakeLTRB(-10, -10, 0, 0),
// overlapping top left corner
SkIRect::MakeLTRB(-10, -10, DEV_W / 4, DEV_H / 4),
// overlapping top left and top right corners
SkIRect::MakeLTRB(-10, -10, DEV_W + 10, DEV_H / 4),
// touching entire top edge
SkIRect::MakeLTRB(-10, -10, DEV_W + 10, 0),
// overlapping top right corner
SkIRect::MakeLTRB(3 * DEV_W / 4, -10, DEV_W + 10, DEV_H / 4),
// contained in x, overlapping top edge
SkIRect::MakeLTRB(DEV_W / 4, -10, 3 * DEV_W / 4, DEV_H / 4),
// outside top right corner
SkIRect::MakeLTRB(DEV_W + 1, -10, DEV_W + 10, -1),
// touching top right corner
SkIRect::MakeLTRB(DEV_W, -10, DEV_W + 10, 0),
// overlapping top left and bottom left corners
SkIRect::MakeLTRB(-10, -10, DEV_W / 4, DEV_H + 10),
// touching entire left edge
SkIRect::MakeLTRB(-10, -10, 0, DEV_H + 10),
// overlapping bottom left corner
SkIRect::MakeLTRB(-10, 3 * DEV_H / 4, DEV_W / 4, DEV_H + 10),
// contained in y, overlapping left edge
SkIRect::MakeLTRB(-10, DEV_H / 4, DEV_W / 4, 3 * DEV_H / 4),
// outside bottom left corner
SkIRect::MakeLTRB(-10, DEV_H + 1, -1, DEV_H + 10),
// touching bottom left corner
SkIRect::MakeLTRB(-10, DEV_H, 0, DEV_H + 10),
// overlapping bottom left and bottom right corners
SkIRect::MakeLTRB(-10, 3 * DEV_H / 4, DEV_W + 10, DEV_H + 10),
// touching entire left edge
SkIRect::MakeLTRB(0, DEV_H, DEV_W, DEV_H + 10),
// overlapping bottom right corner
SkIRect::MakeLTRB(3 * DEV_W / 4, 3 * DEV_H / 4, DEV_W + 10, DEV_H + 10),
// overlapping top right and bottom right corners
SkIRect::MakeLTRB(3 * DEV_W / 4, -10, DEV_W + 10, DEV_H + 10),
};
SkCanvas* canvas = surface->getCanvas();
static const struct {
SkColorType fColorType;
SkAlphaType fAlphaType;
} gSrcConfigs[] = {
{kRGBA_8888_SkColorType, kPremul_SkAlphaType},
{kRGBA_8888_SkColorType, kUnpremul_SkAlphaType},
{kRGB_888x_SkColorType, kOpaque_SkAlphaType},
{kBGRA_8888_SkColorType, kPremul_SkAlphaType},
{kBGRA_8888_SkColorType, kUnpremul_SkAlphaType},
};
for (size_t r = 0; r < SK_ARRAY_COUNT(testRects); ++r) {
const SkIRect& rect = testRects[r];
for (int tightBmp = 0; tightBmp < 2; ++tightBmp) {
for (size_t c = 0; c < SK_ARRAY_COUNT(gSrcConfigs); ++c) {
const SkColorType ct = gSrcConfigs[c].fColorType;
const SkAlphaType at = gSrcConfigs[c].fAlphaType;
bool isGPU = SkToBool(surface->getCanvas()->recordingContext());
fill_surface(surface);
SkBitmap bmp;
REPORTER_ASSERT(reporter, setup_bitmap(&bmp, ct, at, rect.width(),
rect.height(), SkToBool(tightBmp)));
uint32_t idBefore = surface->generationID();
surface->writePixels(bmp, rect.fLeft, rect.fTop);
uint32_t idAfter = surface->generationID();
REPORTER_ASSERT(reporter, check_write(reporter, surface, surfaceInfo.alphaType(),
bmp, rect.fLeft, rect.fTop));
// we should change the genID iff pixels were actually written.
SkIRect canvasRect = SkIRect::MakeSize(canvas->getBaseLayerSize());
SkIRect writeRect = SkIRect::MakeXYWH(rect.fLeft, rect.fTop,
bmp.width(), bmp.height());
bool expectSuccess = SkIRect::Intersects(canvasRect, writeRect) &&
write_should_succeed(surfaceInfo, bmp.info(), isGPU);
REPORTER_ASSERT(reporter, expectSuccess == (idBefore != idAfter));
}
}
}
}
DEF_TEST(WritePixels, reporter) {
const SkImageInfo info = SkImageInfo::MakeN32Premul(DEV_W, DEV_H);
for (auto& tightRowBytes : { true, false }) {
const size_t rowBytes = tightRowBytes ? info.minRowBytes() : 4 * DEV_W + 100;
const size_t size = info.computeByteSize(rowBytes);
void* pixels = sk_malloc_throw(size);
// if rowBytes isn't tight then set the padding to a known value
if (!tightRowBytes) {
memset(pixels, DEV_PAD, size);
}
auto surface(SkSurface::MakeRasterDirectReleaseProc(info, pixels, rowBytes,
free_pixels, nullptr));
test_write_pixels(reporter, surface.get(), info);
}
}
static void test_write_pixels(skiatest::Reporter* reporter,
GrRecordingContext* rContext,
int sampleCnt) {
const SkImageInfo ii = SkImageInfo::MakeN32Premul(DEV_W, DEV_H);
for (auto& origin : { kTopLeft_GrSurfaceOrigin, kBottomLeft_GrSurfaceOrigin }) {
sk_sp<SkSurface> surface(SkSurface::MakeRenderTarget(rContext,
SkBudgeted::kNo, ii, sampleCnt,
origin, nullptr));
if (surface) {
test_write_pixels(reporter, surface.get(), ii);
}
}
}
DEF_GPUTEST_FOR_RENDERING_CONTEXTS(WritePixels_Gpu, reporter, ctxInfo) {
test_write_pixels(reporter, ctxInfo.directContext(), 1);
}
DEF_GPUTEST_FOR_RENDERING_CONTEXTS(WritePixelsMSAA_Gpu, reporter, ctxInfo) {
test_write_pixels(reporter, ctxInfo.directContext(), 1);
}
static void test_write_pixels_non_texture(skiatest::Reporter* reporter,
GrDirectContext* dContext,
int sampleCnt) {
// Dawn currently doesn't support writePixels to a texture-as-render-target.
// See http://skbug.com/10336.
if (GrBackendApi::kDawn == dContext->backend()) {
return;
}
for (auto& origin : { kTopLeft_GrSurfaceOrigin, kBottomLeft_GrSurfaceOrigin }) {
SkColorType colorType = kN32_SkColorType;
auto surface = sk_gpu_test::MakeBackendRenderTargetSurface(dContext,
{DEV_W, DEV_H},
origin,
sampleCnt,
colorType);
if (surface) {
auto ii = SkImageInfo::MakeN32Premul(DEV_W, DEV_H);
test_write_pixels(reporter, surface.get(), ii);
}
}
}
DEF_GPUTEST_FOR_RENDERING_CONTEXTS(WritePixelsNonTexture_Gpu, reporter, ctxInfo) {
test_write_pixels_non_texture(reporter, ctxInfo.directContext(), 1);
}
DEF_GPUTEST_FOR_RENDERING_CONTEXTS(WritePixelsNonTextureMSAA_Gpu, reporter, ctxInfo) {
test_write_pixels_non_texture(reporter, ctxInfo.directContext(), 4);
}
static sk_sp<SkSurface> create_surf(GrRecordingContext* rContext, int width, int height) {
const SkImageInfo ii = SkImageInfo::Make(width, height,
kRGBA_8888_SkColorType, kPremul_SkAlphaType);
sk_sp<SkSurface> surf = SkSurface::MakeRenderTarget(rContext, SkBudgeted::kYes, ii);
surf->flushAndSubmit();
return surf;
}
static sk_sp<SkImage> upload(const sk_sp<SkSurface>& surf, SkColor color) {
const SkImageInfo smII = SkImageInfo::Make(16, 16, kRGBA_8888_SkColorType, kPremul_SkAlphaType);
SkBitmap bm;
bm.allocPixels(smII);
bm.eraseColor(color);
surf->writePixels(bm, 0, 0);
return surf->makeImageSnapshot();
}
// This is tests whether the first writePixels is completed before the
// second writePixels takes effect (i.e., that writePixels correctly flushes
// in between uses of the shared backing resource).
// The unit test fails on Nexus 6P/Android M with driver 129.0 without the
// "DisallowTexSubImageForUnormConfigTexturesEverBoundToFBO" workaround enabled.
// skbug.com/11834
DEF_GPUTEST_FOR_RENDERING_CONTEXTS(WritePixelsPendingIO, reporter, ctxInfo) {
auto context = ctxInfo.directContext();
GrProxyProvider* proxyProvider = context->priv().proxyProvider();
const GrCaps* caps = context->priv().caps();
static const int kFullSize = 62;
static const int kHalfSize = 31;
static const uint32_t kLeftColor = 0xFF222222;
static const uint32_t kRightColor = 0xFFAAAAAA;
const SkImageInfo fullII = SkImageInfo::Make(kFullSize, kFullSize,
kRGBA_8888_SkColorType, kPremul_SkAlphaType);
const SkImageInfo halfII = SkImageInfo::Make(kHalfSize, kFullSize,
kRGBA_8888_SkColorType, kPremul_SkAlphaType);
sk_sp<SkSurface> dest = SkSurface::MakeRenderTarget(context, SkBudgeted::kYes, fullII);
{
// Seed the resource cached with a scratch texture that will be reused by writePixels
static constexpr SkISize kDims = {32, 64};
const GrBackendFormat format = caps->getDefaultBackendFormat(GrColorType::kRGBA_8888,
GrRenderable::kNo);
sk_sp<GrTextureProxy> temp = proxyProvider->createProxy(
format, kDims, GrRenderable::kNo, 1, GrMipmapped::kNo, SkBackingFit::kApprox,
SkBudgeted::kYes, GrProtected::kNo);
temp->instantiate(context->priv().resourceProvider());
}
// Create the surfaces and flush them to ensure there is no lingering pendingIO
sk_sp<SkSurface> leftSurf = create_surf(context, kHalfSize, kFullSize);
sk_sp<SkSurface> rightSurf = create_surf(context, kHalfSize, kFullSize);
sk_sp<SkImage> leftImg = upload(std::move(leftSurf), kLeftColor);
dest->getCanvas()->drawImage(std::move(leftImg), 0, 0);
sk_sp<SkImage> rightImg = upload(std::move(rightSurf), kRightColor);
dest->getCanvas()->drawImage(std::move(rightImg), kHalfSize, 0);
SkBitmap bm;
bm.allocPixels(fullII);
SkAssertResult(dest->readPixels(bm, 0, 0));
bool isCorrect = true;
for (int y = 0; isCorrect && y < 16; ++y) {
const uint32_t* sl = bm.getAddr32(0, y);
for (int x = 0; x < 16; ++x) {
if (kLeftColor != sl[x]) {
isCorrect = false;
break;
}
}
for (int x = kHalfSize; x < kHalfSize+16; ++x) {
if (kRightColor != sl[x]) {
isCorrect = false;
break;
}
}
}
REPORTER_ASSERT(reporter, isCorrect);
}
DEF_TEST(WritePixels_InvalidRowBytes, reporter) {
auto dstII = SkImageInfo::Make({10, 10}, kRGBA_8888_SkColorType, kPremul_SkAlphaType);
auto surf = SkSurface::MakeRaster(dstII);
for (int ct = 0; ct < kLastEnum_SkColorType + 1; ++ct) {
auto colorType = static_cast<SkColorType>(ct);
size_t bpp = SkColorTypeBytesPerPixel(colorType);
if (bpp <= 1) {
continue;
}
auto srcII = dstII.makeColorType(colorType);
size_t badRowBytes = (surf->width() + 1)*bpp - 1;
auto storage = std::make_unique<char[]>(badRowBytes*surf->height());
memset(storage.get(), 0, badRowBytes * surf->height());
// SkSurface::writePixels doesn't report bool, SkCanvas's does.
REPORTER_ASSERT(reporter,
!surf->getCanvas()->writePixels(srcII, storage.get(), badRowBytes, 0, 0));
}
}