ronnie
2022-10-14 1504bb53e29d3d46222c0b3ea994fc494b48e153
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
/*
 * Copyright 2016 Google Inc.
 *
 * Use of this source code is governed by a BSD-style license that can be
 * found in the LICENSE file.
 */
 
#include "SkTypes.h"
#include "Test.h"
 
#include "GrClip.h"
#include "GrContext.h"
#include "GrContextPriv.h"
#include "GrGpuResource.h"
#include "GrMemoryPool.h"
#include "GrProxyProvider.h"
#include "GrRenderTargetContext.h"
#include "GrRenderTargetContextPriv.h"
#include "GrResourceProvider.h"
#include "glsl/GrGLSLFragmentProcessor.h"
#include "glsl/GrGLSLFragmentShaderBuilder.h"
#include "ops/GrFillRectOp.h"
#include "ops/GrMeshDrawOp.h"
#include "TestUtils.h"
 
#include <atomic>
#include <random>
 
namespace {
class TestOp : public GrMeshDrawOp {
public:
    DEFINE_OP_CLASS_ID
    static std::unique_ptr<GrDrawOp> Make(GrContext* context,
                                          std::unique_ptr<GrFragmentProcessor> fp) {
        GrOpMemoryPool* pool = context->contextPriv().opMemoryPool();
 
        return pool->allocate<TestOp>(std::move(fp));
    }
 
    const char* name() const override { return "TestOp"; }
 
    void visitProxies(const VisitProxyFunc& func, VisitorType) const override {
        fProcessors.visitProxies(func);
    }
 
    FixedFunctionFlags fixedFunctionFlags() const override { return FixedFunctionFlags::kNone; }
 
    GrProcessorSet::Analysis finalize(const GrCaps& caps, const GrAppliedClip* clip) override {
        static constexpr GrProcessorAnalysisColor kUnknownColor;
        SkPMColor4f overrideColor;
        return fProcessors.finalize(kUnknownColor, GrProcessorAnalysisCoverage::kNone, clip, false,
                                    caps, &overrideColor);
    }
 
private:
    friend class ::GrOpMemoryPool; // for ctor
 
    TestOp(std::unique_ptr<GrFragmentProcessor> fp)
            : INHERITED(ClassID()), fProcessors(std::move(fp)) {
        this->setBounds(SkRect::MakeWH(100, 100), HasAABloat::kNo, IsZeroArea::kNo);
    }
 
    void onPrepareDraws(Target* target) override { return; }
 
    GrProcessorSet fProcessors;
 
    typedef GrMeshDrawOp INHERITED;
};
 
/**
 * FP used to test ref/IO counts on owned GrGpuResources. Can also be a parent FP to test counts
 * of resources owned by child FPs.
 */
class TestFP : public GrFragmentProcessor {
public:
    static std::unique_ptr<GrFragmentProcessor> Make(std::unique_ptr<GrFragmentProcessor> child) {
        return std::unique_ptr<GrFragmentProcessor>(new TestFP(std::move(child)));
    }
    static std::unique_ptr<GrFragmentProcessor> Make(const SkTArray<sk_sp<GrTextureProxy>>& proxies,
                                                     const SkTArray<sk_sp<GrBuffer>>& buffers) {
        return std::unique_ptr<GrFragmentProcessor>(new TestFP(proxies, buffers));
    }
 
    const char* name() const override { return "test"; }
 
    void onGetGLSLProcessorKey(const GrShaderCaps&, GrProcessorKeyBuilder* b) const override {
        static std::atomic<int32_t> nextKey{0};
        b->add32(nextKey++);
    }
 
    std::unique_ptr<GrFragmentProcessor> clone() const override {
        return std::unique_ptr<GrFragmentProcessor>(new TestFP(*this));
    }
 
private:
    TestFP(const SkTArray<sk_sp<GrTextureProxy>>& proxies, const SkTArray<sk_sp<GrBuffer>>& buffers)
            : INHERITED(kTestFP_ClassID, kNone_OptimizationFlags), fSamplers(4) {
        for (const auto& proxy : proxies) {
            fSamplers.emplace_back(proxy);
        }
        this->setTextureSamplerCnt(fSamplers.count());
    }
 
    TestFP(std::unique_ptr<GrFragmentProcessor> child)
            : INHERITED(kTestFP_ClassID, kNone_OptimizationFlags), fSamplers(4) {
        this->registerChildProcessor(std::move(child));
    }
 
    explicit TestFP(const TestFP& that)
            : INHERITED(kTestFP_ClassID, that.optimizationFlags()), fSamplers(4) {
        for (int i = 0; i < that.fSamplers.count(); ++i) {
            fSamplers.emplace_back(that.fSamplers[i]);
        }
        for (int i = 0; i < that.numChildProcessors(); ++i) {
            this->registerChildProcessor(that.childProcessor(i).clone());
        }
        this->setTextureSamplerCnt(fSamplers.count());
    }
 
    virtual GrGLSLFragmentProcessor* onCreateGLSLInstance() const override {
        class TestGLSLFP : public GrGLSLFragmentProcessor {
        public:
            TestGLSLFP() {}
            void emitCode(EmitArgs& args) override {
                GrGLSLFPFragmentBuilder* fragBuilder = args.fFragBuilder;
                fragBuilder->codeAppendf("%s = %s;", args.fOutputColor, args.fInputColor);
            }
 
        private:
        };
        return new TestGLSLFP();
    }
 
    bool onIsEqual(const GrFragmentProcessor&) const override { return false; }
    const TextureSampler& onTextureSampler(int i) const override { return fSamplers[i]; }
 
    GrTAllocator<TextureSampler> fSamplers;
    typedef GrFragmentProcessor INHERITED;
};
}
 
template <typename T>
inline void testingOnly_getIORefCnts(const T* resource, int* refCnt, int* readCnt, int* writeCnt) {
    *refCnt = resource->fRefCnt;
    *readCnt = resource->fPendingReads;
    *writeCnt = resource->fPendingWrites;
}
 
void testingOnly_getIORefCnts(GrTextureProxy* proxy, int* refCnt, int* readCnt, int* writeCnt) {
    *refCnt = proxy->getBackingRefCnt_TestOnly();
    *readCnt = proxy->getPendingReadCnt_TestOnly();
    *writeCnt = proxy->getPendingWriteCnt_TestOnly();
}
 
DEF_GPUTEST_FOR_ALL_CONTEXTS(ProcessorRefTest, reporter, ctxInfo) {
    GrContext* context = ctxInfo.grContext();
    GrProxyProvider* proxyProvider = context->contextPriv().proxyProvider();
 
    GrSurfaceDesc desc;
    desc.fWidth = 10;
    desc.fHeight = 10;
    desc.fConfig = kRGBA_8888_GrPixelConfig;
 
    const GrBackendFormat format =
            context->contextPriv().caps()->getBackendFormatFromColorType(kRGBA_8888_SkColorType);
 
    for (bool makeClone : {false, true}) {
        for (int parentCnt = 0; parentCnt < 2; parentCnt++) {
            sk_sp<GrRenderTargetContext> renderTargetContext(
                    context->contextPriv().makeDeferredRenderTargetContext(
                                                             format, SkBackingFit::kApprox, 1, 1,
                                                             kRGBA_8888_GrPixelConfig, nullptr));
            {
                sk_sp<GrTextureProxy> proxy1 = proxyProvider->createProxy(
                        format, desc, kTopLeft_GrSurfaceOrigin, SkBackingFit::kExact,
                        SkBudgeted::kYes);
                sk_sp<GrTextureProxy> proxy2 = proxyProvider->createProxy(
                        format, desc, kTopLeft_GrSurfaceOrigin, SkBackingFit::kExact,
                        SkBudgeted::kYes);
                sk_sp<GrTextureProxy> proxy3 = proxyProvider->createProxy(
                        format, desc, kTopLeft_GrSurfaceOrigin, SkBackingFit::kExact,
                        SkBudgeted::kYes);
                sk_sp<GrTextureProxy> proxy4 = proxyProvider->createProxy(
                        format, desc, kTopLeft_GrSurfaceOrigin, SkBackingFit::kExact,
                        SkBudgeted::kYes);
                {
                    SkTArray<sk_sp<GrTextureProxy>> proxies;
                    SkTArray<sk_sp<GrBuffer>> buffers;
                    proxies.push_back(proxy1);
                    auto fp = TestFP::Make(std::move(proxies), std::move(buffers));
                    for (int i = 0; i < parentCnt; ++i) {
                        fp = TestFP::Make(std::move(fp));
                    }
                    std::unique_ptr<GrFragmentProcessor> clone;
                    if (makeClone) {
                        clone = fp->clone();
                    }
                    std::unique_ptr<GrDrawOp> op(TestOp::Make(context, std::move(fp)));
                    renderTargetContext->priv().testingOnly_addDrawOp(std::move(op));
                    if (clone) {
                        op = TestOp::Make(context, std::move(clone));
                        renderTargetContext->priv().testingOnly_addDrawOp(std::move(op));
                    }
                }
                int refCnt, readCnt, writeCnt;
 
                testingOnly_getIORefCnts(proxy1.get(), &refCnt, &readCnt, &writeCnt);
                // IO counts should be double if there is a clone of the FP.
                int ioRefMul = makeClone ? 2 : 1;
                REPORTER_ASSERT(reporter, -1 == refCnt);
                REPORTER_ASSERT(reporter, ioRefMul * 1 == readCnt);
                REPORTER_ASSERT(reporter, ioRefMul * 0 == writeCnt);
 
                context->flush();
 
                testingOnly_getIORefCnts(proxy1.get(), &refCnt, &readCnt, &writeCnt);
                REPORTER_ASSERT(reporter, 1 == refCnt);
                REPORTER_ASSERT(reporter, ioRefMul * 0 == readCnt);
                REPORTER_ASSERT(reporter, ioRefMul * 0 == writeCnt);
 
            }
        }
    }
}
 
// This test uses the random GrFragmentProcessor test factory, which relies on static initializers.
#if SK_ALLOW_STATIC_GLOBAL_INITIALIZERS
 
#include "SkCommandLineFlags.h"
DEFINE_bool(randomProcessorTest, false, "Use non-deterministic seed for random processor tests?");
DEFINE_uint32(processorSeed, 0, "Use specific seed for processor tests. Overridden by " \
                                "--randomProcessorTest.");
 
#if GR_TEST_UTILS
 
static GrColor input_texel_color(int i, int j, SkScalar delta) {
    // Delta must be less than 0.5 to prevent over/underflow issues with the input color
    SkASSERT(delta <= 0.5);
 
    SkColor color = SkColorSetARGB((uint8_t)i, (uint8_t)j, (uint8_t)(i + j), (uint8_t)(2 * j - i));
    SkColor4f color4f = SkColor4f::FromColor(color);
    for (int i = 0; i < 4; i++) {
        if (color4f[i] > 0.5) {
            color4f[i] -= delta;
        } else {
            color4f[i] += delta;
        }
    }
    return color4f.premul().toBytes_RGBA();
}
 
void test_draw_op(GrContext* context,
                  GrRenderTargetContext* rtc,
                  std::unique_ptr<GrFragmentProcessor> fp,
                  sk_sp<GrTextureProxy> inputDataProxy) {
    GrPaint paint;
    paint.addColorTextureProcessor(std::move(inputDataProxy), SkMatrix::I());
    paint.addColorFragmentProcessor(std::move(fp));
    paint.setPorterDuffXPFactory(SkBlendMode::kSrc);
 
    auto op = GrFillRectOp::Make(context, std::move(paint), GrAAType::kNone, SkMatrix::I(),
                                 SkRect::MakeWH(rtc->width(), rtc->height()));
    rtc->addDrawOp(GrNoClip(), std::move(op));
}
 
// This assumes that the output buffer will be the same size as inputDataProxy
void render_fp(GrContext* context, GrRenderTargetContext* rtc, GrFragmentProcessor* fp,
               sk_sp<GrTextureProxy> inputDataProxy, GrColor* buffer) {
    int width = inputDataProxy->width();
    int height = inputDataProxy->height();
 
    // test_draw_op needs to take ownership of an FP, so give it a clone that it can own
    test_draw_op(context, rtc, fp->clone(), inputDataProxy);
    memset(buffer, 0x0, sizeof(GrColor) * width * height);
           rtc->readPixels(SkImageInfo::Make(width, height, kRGBA_8888_SkColorType,
                                             kPremul_SkAlphaType),
                           buffer, 0, 0, 0);
}
 
/** Initializes the two test texture proxies that are available to the FP test factories. */
bool init_test_textures(GrProxyProvider* proxyProvider, SkRandom* random,
                        sk_sp<GrTextureProxy> proxies[2]) {
    static const int kTestTextureSize = 256;
 
    {
        // Put premul data into the RGBA texture that the test FPs can optionally use.
        std::unique_ptr<GrColor[]> rgbaData(new GrColor[kTestTextureSize * kTestTextureSize]);
        for (int y = 0; y < kTestTextureSize; ++y) {
            for (int x = 0; x < kTestTextureSize; ++x) {
                rgbaData[kTestTextureSize * y + x] = input_texel_color(
                        random->nextULessThan(256), random->nextULessThan(256), 0.0f);
            }
        }
 
        SkImageInfo ii = SkImageInfo::Make(kTestTextureSize, kTestTextureSize,
                                           kRGBA_8888_SkColorType, kPremul_SkAlphaType);
        SkPixmap pixmap(ii, rgbaData.get(), ii.minRowBytes());
        sk_sp<SkImage> img = SkImage::MakeRasterCopy(pixmap);
        proxies[0] = proxyProvider->createTextureProxy(img, kNone_GrSurfaceFlags, 1,
                                                       SkBudgeted::kYes, SkBackingFit::kExact);
    }
 
    {
        // Put random values into the alpha texture that the test FPs can optionally use.
        std::unique_ptr<uint8_t[]> alphaData(new uint8_t[kTestTextureSize * kTestTextureSize]);
        for (int y = 0; y < kTestTextureSize; ++y) {
            for (int x = 0; x < kTestTextureSize; ++x) {
                alphaData[kTestTextureSize * y + x] = random->nextULessThan(256);
            }
        }
 
        SkImageInfo ii = SkImageInfo::Make(kTestTextureSize, kTestTextureSize,
                                           kAlpha_8_SkColorType, kPremul_SkAlphaType);
        SkPixmap pixmap(ii, alphaData.get(), ii.minRowBytes());
        sk_sp<SkImage> img = SkImage::MakeRasterCopy(pixmap);
        proxies[1] = proxyProvider->createTextureProxy(img, kNone_GrSurfaceFlags, 1,
                                                       SkBudgeted::kYes, SkBackingFit::kExact);
    }
 
    return proxies[0] && proxies[1];
}
 
// Creates a texture of premul colors used as the output of the fragment processor that precedes
// the fragment processor under test. Color values are those provided by input_texel_color().
sk_sp<GrTextureProxy> make_input_texture(GrProxyProvider* proxyProvider, int width, int height,
                                         SkScalar delta) {
    std::unique_ptr<GrColor[]> data(new GrColor[width * height]);
    for (int y = 0; y < width; ++y) {
        for (int x = 0; x < height; ++x) {
            data.get()[width * y + x] = input_texel_color(x, y, delta);
        }
    }
 
    SkImageInfo ii = SkImageInfo::Make(width, height, kRGBA_8888_SkColorType, kPremul_SkAlphaType);
    SkPixmap pixmap(ii, data.get(), ii.minRowBytes());
    sk_sp<SkImage> img = SkImage::MakeRasterCopy(pixmap);
    return proxyProvider->createTextureProxy(img, kNone_GrSurfaceFlags, 1,
                                             SkBudgeted::kYes, SkBackingFit::kExact);
}
 
bool log_surface_context(sk_sp<GrSurfaceContext> src, SkString* dst) {
    SkImageInfo ii = SkImageInfo::Make(src->width(), src->height(), kRGBA_8888_SkColorType,
                                       kPremul_SkAlphaType);
    SkBitmap bm;
    SkAssertResult(bm.tryAllocPixels(ii));
    SkAssertResult(src->readPixels(ii, bm.getPixels(), bm.rowBytes(), 0, 0));
 
    return bitmap_to_base64_data_uri(bm, dst);
}
 
bool log_surface_proxy(GrContext* context, sk_sp<GrSurfaceProxy> src, SkString* dst) {
    sk_sp<GrSurfaceContext> sContext(context->contextPriv().makeWrappedSurfaceContext(src));
    return log_surface_context(sContext, dst);
}
 
bool fuzzy_color_equals(const SkPMColor4f& c1, const SkPMColor4f& c2) {
    // With the loss of precision of rendering into 32-bit color, then estimating the FP's output
    // from that, it is not uncommon for a valid output to differ from estimate by up to 0.01
    // (really 1/128 ~ .0078, but frequently floating point issues make that tolerance a little
    // too unforgiving).
    static constexpr SkScalar kTolerance = 0.01f;
    for (int i = 0; i < 4; i++) {
        if (!SkScalarNearlyEqual(c1[i], c2[i], kTolerance)) {
            return false;
        }
    }
    return true;
}
 
int modulation_index(int channelIndex, bool alphaModulation) {
    return alphaModulation ? 3 : channelIndex;
}
 
// Given three input colors (color preceding the FP being tested), and the output of the FP, this
// ensures that the out1 = fp * in1.a, out2 = fp * in2.a, and out3 = fp * in3.a, where fp is the
// pre-modulated color that should not be changing across frames (FP's state doesn't change).
//
// When alphaModulation is false, this tests the very similar conditions that out1 = fp * in1,
// etc. using per-channel modulation instead of modulation by just the input alpha channel.
// - This estimates the pre-modulated fp color from one of the input/output pairs and confirms the
//   conditions hold for the other two pairs.
bool legal_modulation(const GrColor& in1, const GrColor& in2, const GrColor& in3,
                      const GrColor& out1, const GrColor& out2, const GrColor& out3,
                      bool alphaModulation) {
    // Convert to floating point, which is the number space the FP operates in (more or less)
    SkPMColor4f in1f = SkPMColor4f::FromBytes_RGBA(in1);
    SkPMColor4f in2f = SkPMColor4f::FromBytes_RGBA(in2);
    SkPMColor4f in3f = SkPMColor4f::FromBytes_RGBA(in3);
    SkPMColor4f out1f = SkPMColor4f::FromBytes_RGBA(out1);
    SkPMColor4f out2f = SkPMColor4f::FromBytes_RGBA(out2);
    SkPMColor4f out3f = SkPMColor4f::FromBytes_RGBA(out3);
 
    // Reconstruct the output of the FP before the shader modulated its color with the input value.
    // When the original input is very small, it may cause the final output color to round
    // to 0, in which case we estimate the pre-modulated color using one of the stepped frames that
    // will then have a guaranteed larger channel value (since the offset will be added to it).
    SkPMColor4f fpPreModulation;
    for (int i = 0; i < 4; i++) {
        int modulationIndex = modulation_index(i, alphaModulation);
        if (in1f[modulationIndex] < 0.2f) {
            // Use the stepped frame
            fpPreModulation[i] = out2f[i] / in2f[modulationIndex];
        } else {
            fpPreModulation[i] = out1f[i] / in1f[modulationIndex];
        }
    }
 
    // With reconstructed pre-modulated FP output, derive the expected value of fp * input for each
    // of the transformed input colors.
    SkPMColor4f expected1 = alphaModulation ? (fpPreModulation * in1f.fA)
                                            : (fpPreModulation * in1f);
    SkPMColor4f expected2 = alphaModulation ? (fpPreModulation * in2f.fA)
                                            : (fpPreModulation * in2f);
    SkPMColor4f expected3 = alphaModulation ? (fpPreModulation * in3f.fA)
                                            : (fpPreModulation * in3f);
 
    return fuzzy_color_equals(out1f, expected1) &&
           fuzzy_color_equals(out2f, expected2) &&
           fuzzy_color_equals(out3f, expected3);
}
 
DEF_GPUTEST_FOR_GL_RENDERING_CONTEXTS(ProcessorOptimizationValidationTest, reporter, ctxInfo) {
    GrContext* context = ctxInfo.grContext();
    GrProxyProvider* proxyProvider = context->contextPriv().proxyProvider();
    auto resourceProvider = context->contextPriv().resourceProvider();
    using FPFactory = GrFragmentProcessorTestFactory;
 
    uint32_t seed = FLAGS_processorSeed;
    if (FLAGS_randomProcessorTest) {
        std::random_device rd;
        seed = rd();
    }
    // If a non-deterministic bot fails this test, check the output to see what seed it used, then
    // use --processorSeed <seed> (without --randomProcessorTest) to reproduce.
    SkRandom random(seed);
 
    const GrBackendFormat format =
            context->contextPriv().caps()->getBackendFormatFromColorType(kRGBA_8888_SkColorType);
 
    // Make the destination context for the test.
    static constexpr int kRenderSize = 256;
    sk_sp<GrRenderTargetContext> rtc = context->contextPriv().makeDeferredRenderTargetContext(
            format, SkBackingFit::kExact, kRenderSize, kRenderSize, kRGBA_8888_GrPixelConfig,
            nullptr);
 
    sk_sp<GrTextureProxy> proxies[2];
    if (!init_test_textures(proxyProvider, &random, proxies)) {
        ERRORF(reporter, "Could not create test textures");
        return;
    }
    GrProcessorTestData testData(&random, context, rtc.get(), proxies);
 
    // Coverage optimization uses three frames with a linearly transformed input texture.  The first
    // frame has no offset, second frames add .2 and .4, which should then be present as a fixed
    // difference between the frame outputs if the FP is properly following the modulation
    // requirements of the coverage optimization.
    static constexpr SkScalar kInputDelta = 0.2f;
    auto inputTexture1 = make_input_texture(proxyProvider, kRenderSize, kRenderSize, 0.0f);
    auto inputTexture2 = make_input_texture(proxyProvider, kRenderSize, kRenderSize, kInputDelta);
    auto inputTexture3 = make_input_texture(proxyProvider, kRenderSize, kRenderSize, 2*kInputDelta);
 
    // Encoded images are very verbose and this tests many potential images, so only export the
    // first failure (subsequent failures have a reasonable chance of being related).
    bool loggedFirstFailure = false;
    bool loggedFirstWarning = false;
 
    // Storage for the three frames required for coverage compatibility optimization. Each frame
    // uses the correspondingly numbered inputTextureX.
    std::unique_ptr<GrColor[]> readData1(new GrColor[kRenderSize * kRenderSize]);
    std::unique_ptr<GrColor[]> readData2(new GrColor[kRenderSize * kRenderSize]);
    std::unique_ptr<GrColor[]> readData3(new GrColor[kRenderSize * kRenderSize]);
 
    // Because processor factories configure themselves in random ways, this is not exhaustive.
    for (int i = 0; i < FPFactory::Count(); ++i) {
        int timesToInvokeFactory = 5;
        // Increase the number of attempts if the FP has child FPs since optimizations likely depend
        // on child optimizations being present.
        std::unique_ptr<GrFragmentProcessor> fp = FPFactory::MakeIdx(i, &testData);
        for (int j = 0; j < fp->numChildProcessors(); ++j) {
            // This value made a reasonable trade off between time and coverage when this test was
            // written.
            timesToInvokeFactory *= FPFactory::Count() / 2;
        }
        for (int j = 0; j < timesToInvokeFactory; ++j) {
            fp = FPFactory::MakeIdx(i, &testData);
            if (!fp->instantiate(resourceProvider)) {
                continue;
            }
 
            if (!fp->hasConstantOutputForConstantInput() && !fp->preservesOpaqueInput() &&
                !fp->compatibleWithCoverageAsAlpha()) {
                continue;
            }
 
            if (fp->compatibleWithCoverageAsAlpha()) {
                // 2nd and 3rd frames are only used when checking coverage optimization
                render_fp(context, rtc.get(), fp.get(), inputTexture2, readData2.get());
                render_fp(context, rtc.get(), fp.get(), inputTexture3, readData3.get());
            }
            // Draw base frame last so that rtc holds the original FP behavior if we need to
            // dump the image to the log.
            render_fp(context, rtc.get(), fp.get(), inputTexture1, readData1.get());
 
            if (0) {  // Useful to see what FPs are being tested.
                SkString children;
                for (int c = 0; c < fp->numChildProcessors(); ++c) {
                    if (!c) {
                        children.append("(");
                    }
                    children.append(fp->childProcessor(c).name());
                    children.append(c == fp->numChildProcessors() - 1 ? ")" : ", ");
                }
                SkDebugf("%s %s\n", fp->name(), children.c_str());
            }
 
            // This test has a history of being flaky on a number of devices. If an FP is logically
            // violating the optimizations, it's reasonable to expect it to violate requirements on
            // a large number of pixels in the image. Sporadic pixel violations are more indicative
            // of device errors and represents a separate problem.
#if defined(SK_BUILD_FOR_SKQP)
            static constexpr int kMaxAcceptableFailedPixels = 0; // Strict when running as SKQP
#else
            static constexpr int kMaxAcceptableFailedPixels = 2 * kRenderSize; // ~0.7% of the image
#endif
 
            int failedPixelCount = 0;
            // Collect first optimization failure message, to be output later as a warning or an
            // error depending on whether the rendering "passed" or failed.
            SkString coverageMessage;
            SkString opaqueMessage;
            SkString constMessage;
            for (int y = 0; y < kRenderSize; ++y) {
                for (int x = 0; x < kRenderSize; ++x) {
                    bool passing = true;
                    GrColor input = input_texel_color(x, y, 0.0f);
                    GrColor output = readData1.get()[y * kRenderSize + x];
 
                    if (fp->compatibleWithCoverageAsAlpha()) {
                        GrColor i2 = input_texel_color(x, y, kInputDelta);
                        GrColor i3 = input_texel_color(x, y, 2 * kInputDelta);
 
                        GrColor o2 = readData2.get()[y * kRenderSize + x];
                        GrColor o3 = readData3.get()[y * kRenderSize + x];
 
                        // A compatible processor is allowed to modulate either the input color or
                        // just the input alpha.
                        bool legalAlphaModulation = legal_modulation(input, i2, i3, output, o2, o3,
                                                                     /* alpha */ true);
                        bool legalColorModulation = legal_modulation(input, i2, i3, output, o2, o3,
                                                                     /* alpha */ false);
 
                        if (!legalColorModulation && !legalAlphaModulation) {
                            passing = false;
 
                            if (coverageMessage.isEmpty()) {
                                coverageMessage.printf("\"Modulating\" processor %s did not match "
                                        "alpha-modulation nor color-modulation rules. "
                                        "Input: 0x%08x, Output: 0x%08x, pixel (%d, %d).",
                                        fp->name(), input, output, x, y);
                            }
                        }
                    }
 
                    SkPMColor4f input4f = SkPMColor4f::FromBytes_RGBA(input);
                    SkPMColor4f output4f = SkPMColor4f::FromBytes_RGBA(output);
                    SkPMColor4f expected4f;
                    if (fp->hasConstantOutputForConstantInput(input4f, &expected4f)) {
                        float rDiff = fabsf(output4f.fR - expected4f.fR);
                        float gDiff = fabsf(output4f.fG - expected4f.fG);
                        float bDiff = fabsf(output4f.fB - expected4f.fB);
                        float aDiff = fabsf(output4f.fA - expected4f.fA);
                        static constexpr float kTol = 4 / 255.f;
                        if (rDiff > kTol || gDiff > kTol || bDiff > kTol || aDiff > kTol) {
                            if (constMessage.isEmpty()) {
                                passing = false;
 
                                constMessage.printf("Processor %s claimed output for const input "
                                        "doesn't match actual output. Error: %f, Tolerance: %f, "
                                        "input: (%f, %f, %f, %f), actual: (%f, %f, %f, %f), "
                                        "expected(%f, %f, %f, %f)", fp->name(),
                                        SkTMax(rDiff, SkTMax(gDiff, SkTMax(bDiff, aDiff))), kTol,
                                        input4f.fR, input4f.fG, input4f.fB, input4f.fA,
                                        output4f.fR, output4f.fG, output4f.fB, output4f.fA,
                                        expected4f.fR, expected4f.fG, expected4f.fB, expected4f.fA);
                            }
                        }
                    }
                    if (input4f.isOpaque() && fp->preservesOpaqueInput() && !output4f.isOpaque()) {
                        passing = false;
 
                        if (opaqueMessage.isEmpty()) {
                            opaqueMessage.printf("Processor %s claimed opaqueness is preserved but "
                                    "it is not. Input: 0x%08x, Output: 0x%08x.",
                                    fp->name(), input, output);
                        }
                    }
 
                    if (!passing) {
                        // Regardless of how many optimizations the pixel violates, count it as a
                        // single bad pixel.
                        failedPixelCount++;
                    }
                }
            }
 
            // Finished analyzing the entire image, see if the number of pixel failures meets the
            // threshold for an FP violating the optimization requirements.
            if (failedPixelCount > kMaxAcceptableFailedPixels) {
                ERRORF(reporter, "Processor violated %d of %d pixels, seed: 0x%08x, processor: %s"
                       ", first failing pixel details are below:",
                       failedPixelCount, kRenderSize * kRenderSize, seed,
                       fp->dumpInfo().c_str());
 
                // Print first failing pixel's details.
                if (!coverageMessage.isEmpty()) {
                    ERRORF(reporter, coverageMessage.c_str());
                }
                if (!constMessage.isEmpty()) {
                    ERRORF(reporter, constMessage.c_str());
                }
                if (!opaqueMessage.isEmpty()) {
                    ERRORF(reporter, opaqueMessage.c_str());
                }
 
                if (!loggedFirstFailure) {
                    // Print with ERRORF to make sure the encoded image is output
                    SkString input;
                    log_surface_proxy(context, inputTexture1, &input);
                    SkString output;
                    log_surface_context(rtc, &output);
                    ERRORF(reporter, "Input image: %s\n\n"
                           "===========================================================\n\n"
                           "Output image: %s\n", input.c_str(), output.c_str());
                    loggedFirstFailure = true;
                }
            } else if(failedPixelCount > 0) {
                // Don't trigger an error, but don't just hide the failures either.
                INFOF(reporter, "Processor violated %d of %d pixels (below error threshold), seed: "
                      "0x%08x, processor: %s", failedPixelCount, kRenderSize * kRenderSize,
                      seed, fp->dumpInfo().c_str());
                if (!coverageMessage.isEmpty()) {
                    INFOF(reporter, coverageMessage.c_str());
                }
                if (!constMessage.isEmpty()) {
                    INFOF(reporter, constMessage.c_str());
                }
                if (!opaqueMessage.isEmpty()) {
                    INFOF(reporter, opaqueMessage.c_str());
                }
                if (!loggedFirstWarning) {
                    SkString input;
                    log_surface_proxy(context, inputTexture1, &input);
                    SkString output;
                    log_surface_context(rtc, &output);
                    INFOF(reporter, "Input image: %s\n\n"
                          "===========================================================\n\n"
                          "Output image: %s\n", input.c_str(), output.c_str());
                    loggedFirstWarning = true;
                }
            }
        }
    }
}
 
// Tests that fragment processors returned by GrFragmentProcessor::clone() are equivalent to their
// progenitors.
DEF_GPUTEST_FOR_GL_RENDERING_CONTEXTS(ProcessorCloneTest, reporter, ctxInfo) {
    GrContext* context = ctxInfo.grContext();
    GrProxyProvider* proxyProvider = context->contextPriv().proxyProvider();
    auto resourceProvider = context->contextPriv().resourceProvider();
 
    SkRandom random;
 
    const GrBackendFormat format =
            context->contextPriv().caps()->getBackendFormatFromColorType(kRGBA_8888_SkColorType);
 
    // Make the destination context for the test.
    static constexpr int kRenderSize = 1024;
    sk_sp<GrRenderTargetContext> rtc = context->contextPriv().makeDeferredRenderTargetContext(
            format, SkBackingFit::kExact, kRenderSize, kRenderSize, kRGBA_8888_GrPixelConfig,
            nullptr);
 
    sk_sp<GrTextureProxy> proxies[2];
    if (!init_test_textures(proxyProvider, &random, proxies)) {
        ERRORF(reporter, "Could not create test textures");
        return;
    }
    GrProcessorTestData testData(&random, context, rtc.get(), proxies);
 
    auto inputTexture = make_input_texture(proxyProvider, kRenderSize, kRenderSize, 0.0f);
    std::unique_ptr<GrColor[]> readData1(new GrColor[kRenderSize * kRenderSize]);
    std::unique_ptr<GrColor[]> readData2(new GrColor[kRenderSize * kRenderSize]);
    auto readInfo = SkImageInfo::Make(kRenderSize, kRenderSize, kRGBA_8888_SkColorType,
                                      kPremul_SkAlphaType);
 
    // Because processor factories configure themselves in random ways, this is not exhaustive.
    for (int i = 0; i < GrFragmentProcessorTestFactory::Count(); ++i) {
        static constexpr int kTimesToInvokeFactory = 10;
        for (int j = 0; j < kTimesToInvokeFactory; ++j) {
            auto fp = GrFragmentProcessorTestFactory::MakeIdx(i, &testData);
            auto clone = fp->clone();
            if (!clone) {
                ERRORF(reporter, "Clone of processor %s failed.", fp->name());
                continue;
            }
            const char* name = fp->name();
            if (!fp->instantiate(resourceProvider) || !clone->instantiate(resourceProvider)) {
                continue;
            }
            REPORTER_ASSERT(reporter, !strcmp(fp->name(), clone->name()));
            REPORTER_ASSERT(reporter, fp->compatibleWithCoverageAsAlpha() ==
                                      clone->compatibleWithCoverageAsAlpha());
            REPORTER_ASSERT(reporter, fp->isEqual(*clone));
            REPORTER_ASSERT(reporter, fp->preservesOpaqueInput() == clone->preservesOpaqueInput());
            REPORTER_ASSERT(reporter, fp->hasConstantOutputForConstantInput() ==
                                      clone->hasConstantOutputForConstantInput());
            REPORTER_ASSERT(reporter, fp->numChildProcessors() == clone->numChildProcessors());
            REPORTER_ASSERT(reporter, fp->usesLocalCoords() == clone->usesLocalCoords());
            // Draw with original and read back the results.
            render_fp(context, rtc.get(), fp.get(), inputTexture, readData1.get());
 
            // Draw with clone and read back the results.
            render_fp(context, rtc.get(), clone.get(), inputTexture, readData2.get());
 
            // Check that the results are the same.
            bool passing = true;
            for (int y = 0; y < kRenderSize && passing; ++y) {
                for (int x = 0; x < kRenderSize && passing; ++x) {
                    int idx = y * kRenderSize + x;
                    if (readData1[idx] != readData2[idx]) {
                        ERRORF(reporter,
                               "Processor %s made clone produced different output. "
                               "Input color: 0x%08x, Original Output Color: 0x%08x, "
                               "Clone Output Color: 0x%08x..",
                               name, input_texel_color(x, y, 0.0f), readData1[idx], readData2[idx]);
                        passing = false;
                    }
                }
            }
        }
    }
}
 
#endif  // GR_TEST_UTILS
#endif  // SK_ALLOW_STATIC_GLOBAL_INITIALIZERS