src/gpu/gl/GrGLGpu.cpp

/*
 * 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 "GrGLGpu.h"
#include "GrGLStencilBuffer.h"
#include "GrGLTextureRenderTarget.h"
#include "GrGpuResourcePriv.h"
#include "GrPipeline.h"
#include "GrRenderTargetPriv.h"
#include "GrSurfacePriv.h"
#include "GrTemplates.h"
#include "GrTexturePriv.h"
#include "GrTypes.h"
#include "SkStrokeRec.h"
#include "SkTemplates.h"

#define GL_CALL(X) GR_GL_CALL(this->glInterface(), X)
#define GL_CALL_RET(RET, X) GR_GL_CALL_RET(this->glInterface(), RET, X)

#define SKIP_CACHE_CHECK    true

#if GR_GL_CHECK_ALLOC_WITH_GET_ERROR
    #define CLEAR_ERROR_BEFORE_ALLOC(iface)   GrGLClearErr(iface)
    #define GL_ALLOC_CALL(iface, call)        GR_GL_CALL_NOERRCHECK(iface, call)
    #define CHECK_ALLOC_ERROR(iface)          GR_GL_GET_ERROR(iface)
#else
    #define CLEAR_ERROR_BEFORE_ALLOC(iface)
    #define GL_ALLOC_CALL(iface, call)        GR_GL_CALL(iface, call)
    #define CHECK_ALLOC_ERROR(iface)          GR_GL_NO_ERROR
#endif


///////////////////////////////////////////////////////////////////////////////


static const GrGLenum gXfermodeCoeff2Blend[] = {
    GR_GL_ZERO,
    GR_GL_ONE,
    GR_GL_SRC_COLOR,
    GR_GL_ONE_MINUS_SRC_COLOR,
    GR_GL_DST_COLOR,
    GR_GL_ONE_MINUS_DST_COLOR,
    GR_GL_SRC_ALPHA,
    GR_GL_ONE_MINUS_SRC_ALPHA,
    GR_GL_DST_ALPHA,
    GR_GL_ONE_MINUS_DST_ALPHA,
    GR_GL_CONSTANT_COLOR,
    GR_GL_ONE_MINUS_CONSTANT_COLOR,
    GR_GL_CONSTANT_ALPHA,
    GR_GL_ONE_MINUS_CONSTANT_ALPHA,

    // extended blend coeffs
    GR_GL_SRC1_COLOR,
    GR_GL_ONE_MINUS_SRC1_COLOR,
    GR_GL_SRC1_ALPHA,
    GR_GL_ONE_MINUS_SRC1_ALPHA,
};

bool GrGLGpu::BlendCoeffReferencesConstant(GrBlendCoeff coeff) {
    static const bool gCoeffReferencesBlendConst[] = {
        false,
        false,
        false,
        false,
        false,
        false,
        false,
        false,
        false,
        false,
        true,
        true,
        true,
        true,

        // extended blend coeffs
        false,
        false,
        false,
        false,
    };
    return gCoeffReferencesBlendConst[coeff];
    GR_STATIC_ASSERT(kTotalGrBlendCoeffCount ==
                     SK_ARRAY_COUNT(gCoeffReferencesBlendConst));

    GR_STATIC_ASSERT(0 == kZero_GrBlendCoeff);
    GR_STATIC_ASSERT(1 == kOne_GrBlendCoeff);
    GR_STATIC_ASSERT(2 == kSC_GrBlendCoeff);
    GR_STATIC_ASSERT(3 == kISC_GrBlendCoeff);
    GR_STATIC_ASSERT(4 == kDC_GrBlendCoeff);
    GR_STATIC_ASSERT(5 == kIDC_GrBlendCoeff);
    GR_STATIC_ASSERT(6 == kSA_GrBlendCoeff);
    GR_STATIC_ASSERT(7 == kISA_GrBlendCoeff);
    GR_STATIC_ASSERT(8 == kDA_GrBlendCoeff);
    GR_STATIC_ASSERT(9 == kIDA_GrBlendCoeff);
    GR_STATIC_ASSERT(10 == kConstC_GrBlendCoeff);
    GR_STATIC_ASSERT(11 == kIConstC_GrBlendCoeff);
    GR_STATIC_ASSERT(12 == kConstA_GrBlendCoeff);
    GR_STATIC_ASSERT(13 == kIConstA_GrBlendCoeff);

    GR_STATIC_ASSERT(14 == kS2C_GrBlendCoeff);
    GR_STATIC_ASSERT(15 == kIS2C_GrBlendCoeff);
    GR_STATIC_ASSERT(16 == kS2A_GrBlendCoeff);
    GR_STATIC_ASSERT(17 == kIS2A_GrBlendCoeff);

    // assertion for gXfermodeCoeff2Blend have to be in GrGpu scope
    GR_STATIC_ASSERT(kTotalGrBlendCoeffCount ==
                     SK_ARRAY_COUNT(gXfermodeCoeff2Blend));
}

///////////////////////////////////////////////////////////////////////////////

static bool gPrintStartupSpew;

GrGLGpu::GrGLGpu(const GrGLContext& ctx, GrContext* context)
    : GrGpu(context)
    , fGLContext(ctx) {

    SkASSERT(ctx.isInitialized());
    fCaps.reset(SkRef(ctx.caps()));

    fHWBoundTextureUniqueIDs.reset(this->glCaps().maxFragmentTextureUnits());

    GrGLClearErr(fGLContext.interface());
    if (gPrintStartupSpew) {
        const GrGLubyte* vendor;
        const GrGLubyte* renderer;
        const GrGLubyte* version;
        GL_CALL_RET(vendor, GetString(GR_GL_VENDOR));
        GL_CALL_RET(renderer, GetString(GR_GL_RENDERER));
        GL_CALL_RET(version, GetString(GR_GL_VERSION));
        SkDebugf("------------------------- create GrGLGpu %p --------------\n",
                 this);
        SkDebugf("------ VENDOR %s\n", vendor);
        SkDebugf("------ RENDERER %s\n", renderer);
        SkDebugf("------ VERSION %s\n",  version);
        SkDebugf("------ EXTENSIONS\n");
        ctx.extensions().print();
        SkDebugf("\n");
        SkDebugf("%s", this->glCaps().dump().c_str());
    }

    fProgramCache = SkNEW_ARGS(ProgramCache, (this));

    SkASSERT(this->glCaps().maxVertexAttributes() >= GrGeometryProcessor::kMaxVertexAttribs);

    fLastSuccessfulStencilFmtIdx = 0;
    fHWProgramID = 0;

    if (this->glCaps().pathRenderingSupport()) {
        fPathRendering.reset(new GrGLPathRendering(this));
    }
}

GrGLGpu::~GrGLGpu() {
    if (0 != fHWProgramID) {
        // detach the current program so there is no confusion on OpenGL's part
        // that we want it to be deleted
        SkASSERT(fHWProgramID == fCurrentProgram->programID());
        GL_CALL(UseProgram(0));
    }

    if (fTempSrcFBO) {
        fTempSrcFBO->release(this->glInterface());
        fTempSrcFBO.reset(NULL);
    }
    if (fTempDstFBO) {
        fTempDstFBO->release(this->glInterface());
        fTempDstFBO.reset(NULL);
    }
    if (fStencilClearFBO) {
        fStencilClearFBO->release(this->glInterface());
        fStencilClearFBO.reset(NULL);
    }

    delete fProgramCache;
}

void GrGLGpu::contextAbandoned() {
    INHERITED::contextAbandoned();
    fProgramCache->abandon();
    fHWProgramID = 0;
    if (fTempSrcFBO) {
        fTempSrcFBO->abandon();
        fTempSrcFBO.reset(NULL);
    }
    if (fTempDstFBO) {
        fTempDstFBO->abandon();
        fTempDstFBO.reset(NULL);
    }
    if (fStencilClearFBO) {
        fStencilClearFBO->abandon();
        fStencilClearFBO.reset(NULL);
    }

    if (this->glCaps().pathRenderingSupport()) {
        this->glPathRendering()->abandonGpuResources();
    }
}

///////////////////////////////////////////////////////////////////////////////
GrPixelConfig GrGLGpu::preferredReadPixelsConfig(GrPixelConfig readConfig,
                                                 GrPixelConfig surfaceConfig) const {
    if (GR_GL_RGBA_8888_PIXEL_OPS_SLOW && kRGBA_8888_GrPixelConfig == readConfig) {
        return kBGRA_8888_GrPixelConfig;
    } else if (this->glContext().isMesa() &&
               GrBytesPerPixel(readConfig) == 4 &&
               GrPixelConfigSwapRAndB(readConfig) == surfaceConfig) {
        // Mesa 3D takes a slow path on when reading back  BGRA from an RGBA surface and vice-versa.
        // Perhaps this should be guarded by some compiletime or runtime check.
        return surfaceConfig;
    } else if (readConfig == kBGRA_8888_GrPixelConfig
            && !this->glCaps().readPixelsSupported(
                this->glInterface(),
                GR_GL_BGRA,
                GR_GL_UNSIGNED_BYTE,
                surfaceConfig
            )) {
        return kRGBA_8888_GrPixelConfig;
    } else {
        return readConfig;
    }
}

GrPixelConfig GrGLGpu::preferredWritePixelsConfig(GrPixelConfig writeConfig,
                                                  GrPixelConfig surfaceConfig) const {
    if (GR_GL_RGBA_8888_PIXEL_OPS_SLOW && kRGBA_8888_GrPixelConfig == writeConfig) {
        return kBGRA_8888_GrPixelConfig;
    } else {
        return writeConfig;
    }
}

bool GrGLGpu::canWriteTexturePixels(const GrTexture* texture, GrPixelConfig srcConfig) const {
    if (kIndex_8_GrPixelConfig == srcConfig || kIndex_8_GrPixelConfig == texture->config()) {
        return false;
    }
    if (srcConfig != texture->config() && kGLES_GrGLStandard == this->glStandard()) {
        // In general ES2 requires the internal format of the texture and the format of the src
        // pixels to match. However, It may or may not be possible to upload BGRA data to a RGBA
        // texture. It depends upon which extension added BGRA. The Apple extension allows it
        // (BGRA's internal format is RGBA) while the EXT extension does not (BGRA is its own
        // internal format).
        if (this->glCaps().isConfigTexturable(kBGRA_8888_GrPixelConfig) &&
            !this->glCaps().bgraIsInternalFormat() &&
            kBGRA_8888_GrPixelConfig == srcConfig &&
            kRGBA_8888_GrPixelConfig == texture->config()) {
            return true;
        } else {
            return false;
        }
    } else {
        return true;
    }
}

bool GrGLGpu::fullReadPixelsIsFasterThanPartial() const {
    return SkToBool(GR_GL_FULL_READPIXELS_FASTER_THAN_PARTIAL);
}

void GrGLGpu::onResetContext(uint32_t resetBits) {
    // we don't use the zb at all
    if (resetBits & kMisc_GrGLBackendState) {
        GL_CALL(Disable(GR_GL_DEPTH_TEST));
        GL_CALL(DepthMask(GR_GL_FALSE));

        fHWDrawFace = GrPipelineBuilder::kInvalid_DrawFace;
        fHWDitherEnabled = kUnknown_TriState;

        if (kGL_GrGLStandard == this->glStandard()) {
            // Desktop-only state that we never change
            if (!this->glCaps().isCoreProfile()) {
                GL_CALL(Disable(GR_GL_POINT_SMOOTH));
                GL_CALL(Disable(GR_GL_LINE_SMOOTH));
                GL_CALL(Disable(GR_GL_POLYGON_SMOOTH));
                GL_CALL(Disable(GR_GL_POLYGON_STIPPLE));
                GL_CALL(Disable(GR_GL_COLOR_LOGIC_OP));
                GL_CALL(Disable(GR_GL_INDEX_LOGIC_OP));
            }
            // The windows NVIDIA driver has GL_ARB_imaging in the extension string when using a
            // core profile. This seems like a bug since the core spec removes any mention of
            // GL_ARB_imaging.
            if (this->glCaps().imagingSupport() && !this->glCaps().isCoreProfile()) {
                GL_CALL(Disable(GR_GL_COLOR_TABLE));
            }
            GL_CALL(Disable(GR_GL_POLYGON_OFFSET_FILL));
            // Since ES doesn't support glPointSize at all we always use the VS to
            // set the point size
            GL_CALL(Enable(GR_GL_VERTEX_PROGRAM_POINT_SIZE));

            // We should set glPolygonMode(FRONT_AND_BACK,FILL) here, too. It isn't
            // currently part of our gl interface. There are probably others as
            // well.
        }

        if (kGLES_GrGLStandard == this->glStandard() &&
                fGLContext.hasExtension("GL_ARM_shader_framebuffer_fetch")) {
            // The arm extension requires specifically enabling MSAA fetching per sample.
            // On some devices this may have a perf hit.  Also multiple render targets are disabled
            GL_CALL(Enable(GR_GL_FETCH_PER_SAMPLE_ARM));
        }
        fHWWriteToColor = kUnknown_TriState;
        // we only ever use lines in hairline mode
        GL_CALL(LineWidth(1));
    }

    if (resetBits & kMSAAEnable_GrGLBackendState) {
        fMSAAEnabled = kUnknown_TriState;
    }

    fHWActiveTextureUnitIdx = -1; // invalid

    if (resetBits & kTextureBinding_GrGLBackendState) {
        for (int s = 0; s < fHWBoundTextureUniqueIDs.count(); ++s) {
            fHWBoundTextureUniqueIDs[s] = SK_InvalidUniqueID;
        }
    }

    if (resetBits & kBlend_GrGLBackendState) {
        fHWBlendState.invalidate();
    }

    if (resetBits & kView_GrGLBackendState) {
        fHWScissorSettings.invalidate();
        fHWViewport.invalidate();
    }

    if (resetBits & kStencil_GrGLBackendState) {
        fHWStencilSettings.invalidate();
        fHWStencilTestEnabled = kUnknown_TriState;
    }

    // Vertex
    if (resetBits & kVertex_GrGLBackendState) {
        fHWGeometryState.invalidate();
    }

    if (resetBits & kRenderTarget_GrGLBackendState) {
        for (size_t i = 0; i < SK_ARRAY_COUNT(fHWFBOBinding); ++i) {
            fHWFBOBinding[i].invalidate();
        }
    }

    if (resetBits & kPathRendering_GrGLBackendState) {
        if (this->caps()->pathRenderingSupport()) {
            this->glPathRendering()->resetContext();
        }
    }

    // we assume these values
    if (resetBits & kPixelStore_GrGLBackendState) {
        if (this->glCaps().unpackRowLengthSupport()) {
            GL_CALL(PixelStorei(GR_GL_UNPACK_ROW_LENGTH, 0));
        }
        if (this->glCaps().packRowLengthSupport()) {
            GL_CALL(PixelStorei(GR_GL_PACK_ROW_LENGTH, 0));
        }
        if (this->glCaps().unpackFlipYSupport()) {
            GL_CALL(PixelStorei(GR_GL_UNPACK_FLIP_Y, GR_GL_FALSE));
        }
        if (this->glCaps().packFlipYSupport()) {
            GL_CALL(PixelStorei(GR_GL_PACK_REVERSE_ROW_ORDER, GR_GL_FALSE));
        }
    }

    if (resetBits & kProgram_GrGLBackendState) {
        fHWProgramID = 0;
    }
}

namespace {

GrSurfaceOrigin resolve_origin(GrSurfaceOrigin origin, bool renderTarget) {
    // By default, GrRenderTargets are GL's normal orientation so that they
    // can be drawn to by the outside world without the client having
    // to render upside down.
    if (kDefault_GrSurfaceOrigin == origin) {
        return renderTarget ? kBottomLeft_GrSurfaceOrigin : kTopLeft_GrSurfaceOrigin;
    } else {
        return origin;
    }
}

}

GrTexture* GrGLGpu::onWrapBackendTexture(const GrBackendTextureDesc& desc) {
    if (!this->configToGLFormats(desc.fConfig, false, NULL, NULL, NULL)) {
        return NULL;
    }

    if (0 == desc.fTextureHandle) {
        return NULL;
    }

    int maxSize = this->caps()->maxTextureSize();
    if (desc.fWidth > maxSize || desc.fHeight > maxSize) {
        return NULL;
    }

    GrGLTexture::IDDesc idDesc;
    GrSurfaceDesc surfDesc;

    idDesc.fTextureID = static_cast<GrGLuint>(desc.fTextureHandle);
    idDesc.fLifeCycle = GrGpuResource::kWrapped_LifeCycle;

    // next line relies on GrBackendTextureDesc's flags matching GrTexture's
    surfDesc.fFlags = (GrSurfaceFlags) desc.fFlags;
    surfDesc.fWidth = desc.fWidth;
    surfDesc.fHeight = desc.fHeight;
    surfDesc.fConfig = desc.fConfig;
    surfDesc.fSampleCnt = desc.fSampleCnt;
    bool renderTarget = SkToBool(desc.fFlags & kRenderTarget_GrBackendTextureFlag);
    // FIXME:  this should be calling resolve_origin(), but Chrome code is currently
    // assuming the old behaviour, which is that backend textures are always
    // BottomLeft, even for non-RT's.  Once Chrome is fixed, change this to:
    // glTexDesc.fOrigin = resolve_origin(desc.fOrigin, renderTarget);
    if (kDefault_GrSurfaceOrigin == desc.fOrigin) {
        surfDesc.fOrigin = kBottomLeft_GrSurfaceOrigin;
    } else {
        surfDesc.fOrigin = desc.fOrigin;
    }

    GrGLTexture* texture = NULL;
    if (renderTarget) {
        GrGLRenderTarget::IDDesc rtIDDesc;
        if (!this->createRenderTargetObjects(surfDesc, false, idDesc.fTextureID, &rtIDDesc)) {
            return NULL;
        }
        texture = SkNEW_ARGS(GrGLTextureRenderTarget, (this, surfDesc, idDesc, rtIDDesc));
    } else {
        texture = SkNEW_ARGS(GrGLTexture, (this, surfDesc, idDesc));
    }
    if (NULL == texture) {
        return NULL;
    }

    return texture;
}

GrRenderTarget* GrGLGpu::onWrapBackendRenderTarget(const GrBackendRenderTargetDesc& wrapDesc) {
    GrGLRenderTarget::IDDesc idDesc;
    GrGLuint fboID = static_cast<GrGLuint>(wrapDesc.fRenderTargetHandle);
    idDesc.fRenderFBO.reset(SkNEW_ARGS(GrGLFBO, (fboID)));
    idDesc.fMSColorRenderbufferID = 0;
    idDesc.fLifeCycle = GrGpuResource::kWrapped_LifeCycle;

    GrSurfaceDesc desc;
    desc.fConfig = wrapDesc.fConfig;
    desc.fFlags = kCheckAllocation_GrSurfaceFlag;
    desc.fWidth = wrapDesc.fWidth;
    desc.fHeight = wrapDesc.fHeight;
    desc.fSampleCnt = wrapDesc.fSampleCnt;
    desc.fOrigin = resolve_origin(wrapDesc.fOrigin, true);

    GrRenderTarget* tgt = SkNEW_ARGS(GrGLRenderTarget, (this, desc, idDesc));
    if (wrapDesc.fStencilBits) {
        GrGLStencilBuffer::IDDesc sbDesc;
        GrGLStencilBuffer::Format format;
        format.fInternalFormat = GrGLStencilBuffer::kUnknownInternalFormat;
        format.fPacked = false;
        format.fStencilBits = wrapDesc.fStencilBits;
        format.fTotalBits = wrapDesc.fStencilBits;
        GrGLStencilBuffer* sb = SkNEW_ARGS(GrGLStencilBuffer,
                                           (this,
                                            sbDesc,
                                            desc.fWidth,
                                            desc.fHeight,
                                            desc.fSampleCnt,
                                            format));
        tgt->renderTargetPriv().didAttachStencilBuffer(sb);
        sb->unref();
    }
    return tgt;
}

////////////////////////////////////////////////////////////////////////////////

bool GrGLGpu::onWriteTexturePixels(GrTexture* texture,
                                   int left, int top, int width, int height,
                                   GrPixelConfig config, const void* buffer,
                                   size_t rowBytes) {
    if (NULL == buffer) {
        return false;
    }
    GrGLTexture* glTex = static_cast<GrGLTexture*>(texture);

    this->setScratchTextureUnit();
    GL_CALL(BindTexture(GR_GL_TEXTURE_2D, glTex->textureID()));

    bool success = false;
    if (GrPixelConfigIsCompressed(glTex->desc().fConfig)) {
        // We check that config == desc.fConfig in GrGLGpu::canWriteTexturePixels()
        SkASSERT(config == glTex->desc().fConfig);
        success = this->uploadCompressedTexData(glTex->desc(), buffer, false, left, top, width,
                                                height);
    } else {
        success = this->uploadTexData(glTex->desc(), false, left, top, width, height, config,
                                      buffer, rowBytes);
    }

    if (success) {
        texture->texturePriv().dirtyMipMaps(true);
        return true;
    }

    return false;
}

static bool adjust_pixel_ops_params(int surfaceWidth,
                                    int surfaceHeight,
                                    size_t bpp,
                                    int* left, int* top, int* width, int* height,
                                    const void** data,
                                    size_t* rowBytes) {
    if (!*rowBytes) {
        *rowBytes = *width * bpp;
    }

    SkIRect subRect = SkIRect::MakeXYWH(*left, *top, *width, *height);
    SkIRect bounds = SkIRect::MakeWH(surfaceWidth, surfaceHeight);

    if (!subRect.intersect(bounds)) {
        return false;
    }
    *data = reinterpret_cast<const void*>(reinterpret_cast<intptr_t>(*data) +
          (subRect.fTop - *top) * *rowBytes + (subRect.fLeft - *left) * bpp);

    *left = subRect.fLeft;
    *top = subRect.fTop;
    *width = subRect.width();
    *height = subRect.height();
    return true;
}

static inline GrGLenum check_alloc_error(const GrSurfaceDesc& desc,
                                         const GrGLInterface* interface) {
    if (SkToBool(desc.fFlags & kCheckAllocation_GrSurfaceFlag)) {
        return GR_GL_GET_ERROR(interface);
    } else {
        return CHECK_ALLOC_ERROR(interface);
    }
}

bool GrGLGpu::uploadTexData(const GrSurfaceDesc& desc,
                            bool isNewTexture,
                            int left, int top, int width, int height,
                            GrPixelConfig dataConfig,
                            const void* data,
                            size_t rowBytes) {
    SkASSERT(data || isNewTexture);

    // If we're uploading compressed data then we should be using uploadCompressedTexData
    SkASSERT(!GrPixelConfigIsCompressed(dataConfig));

    size_t bpp = GrBytesPerPixel(dataConfig);
    if (!adjust_pixel_ops_params(desc.fWidth, desc.fHeight, bpp, &left, &top,
                                 &width, &height, &data, &rowBytes)) {
        return false;
    }
    size_t trimRowBytes = width * bpp;

    // in case we need a temporary, trimmed copy of the src pixels
    GrAutoMalloc<128 * 128> tempStorage;

    // We currently lazily create MIPMAPs when the we see a draw with
    // GrTextureParams::kMipMap_FilterMode. Using texture storage requires that the
    // MIP levels are all created when the texture is created. So for now we don't use
    // texture storage.
    bool useTexStorage = false &&
                         isNewTexture &&
                         this->glCaps().texStorageSupport();

    if (useTexStorage && kGL_GrGLStandard == this->glStandard()) {
        // 565 is not a sized internal format on desktop GL. So on desktop with
        // 565 we always use an unsized internal format to let the system pick
        // the best sized format to convert the 565 data to. Since TexStorage
        // only allows sized internal formats we will instead use TexImage2D.
        useTexStorage = desc.fConfig != kRGB_565_GrPixelConfig;
    }

    GrGLenum internalFormat = 0x0; // suppress warning
    GrGLenum externalFormat = 0x0; // suppress warning
    GrGLenum externalType = 0x0;   // suppress warning

    // glTexStorage requires sized internal formats on both desktop and ES. ES2 requires an unsized
    // format for glTexImage, unlike ES3 and desktop.
    bool useSizedFormat = useTexStorage;
    if (kGL_GrGLStandard == this->glStandard() ||
        (this->glVersion() >= GR_GL_VER(3, 0) &&
         // ES3 only works with sized BGRA8 format if "GL_APPLE_texture_format_BGRA8888" enabled
         (kBGRA_8888_GrPixelConfig != dataConfig || !this->glCaps().bgraIsInternalFormat())))  {
        useSizedFormat = true;
    }

    if (!this->configToGLFormats(dataConfig, useSizedFormat, &internalFormat,
                                 &externalFormat, &externalType)) {
        return false;
    }

    /*
     *  check whether to allocate a temporary buffer for flipping y or
     *  because our srcData has extra bytes past each row. If so, we need
     *  to trim those off here, since GL ES may not let us specify
     *  GL_UNPACK_ROW_LENGTH.
     */
    bool restoreGLRowLength = false;
    bool swFlipY = false;
    bool glFlipY = false;
    if (data) {
        if (kBottomLeft_GrSurfaceOrigin == desc.fOrigin) {
            if (this->glCaps().unpackFlipYSupport()) {
                glFlipY = true;
            } else {
                swFlipY = true;
            }
        }
        if (this->glCaps().unpackRowLengthSupport() && !swFlipY) {
            // can't use this for flipping, only non-neg values allowed. :(
            if (rowBytes != trimRowBytes) {
                GrGLint rowLength = static_cast<GrGLint>(rowBytes / bpp);
                GL_CALL(PixelStorei(GR_GL_UNPACK_ROW_LENGTH, rowLength));
                restoreGLRowLength = true;
            }
        } else {
            if (trimRowBytes != rowBytes || swFlipY) {
                // copy data into our new storage, skipping the trailing bytes
                size_t trimSize = height * trimRowBytes;
                const char* src = (const char*)data;
                if (swFlipY) {
                    src += (height - 1) * rowBytes;
                }
                char* dst = (char*)tempStorage.reset(trimSize);
                for (int y = 0; y < height; y++) {
                    memcpy(dst, src, trimRowBytes);
                    if (swFlipY) {
                        src -= rowBytes;
                    } else {
                        src += rowBytes;
                    }
                    dst += trimRowBytes;
                }
                // now point data to our copied version
                data = tempStorage.get();
            }
        }
        if (glFlipY) {
            GL_CALL(PixelStorei(GR_GL_UNPACK_FLIP_Y, GR_GL_TRUE));
        }
        GL_CALL(PixelStorei(GR_GL_UNPACK_ALIGNMENT,
              static_cast<GrGLint>(GrUnpackAlignment(dataConfig))));
    }
    bool succeeded = true;
    if (isNewTexture &&
        0 == left && 0 == top &&
        desc.fWidth == width && desc.fHeight == height) {
        CLEAR_ERROR_BEFORE_ALLOC(this->glInterface());
        if (useTexStorage) {
            // We never resize  or change formats of textures.
            GL_ALLOC_CALL(this->glInterface(),
                          TexStorage2D(GR_GL_TEXTURE_2D,
                                       1, // levels
                                       internalFormat,
                                       desc.fWidth, desc.fHeight));
        } else {
            GL_ALLOC_CALL(this->glInterface(),
                          TexImage2D(GR_GL_TEXTURE_2D,
                                     0, // level
                                     internalFormat,
                                     desc.fWidth, desc.fHeight,
                                     0, // border
                                     externalFormat, externalType,
                                     data));
        }
        GrGLenum error = check_alloc_error(desc, this->glInterface());
        if (error != GR_GL_NO_ERROR) {
            succeeded = false;
        } else {
            // if we have data and we used TexStorage to create the texture, we
            // now upload with TexSubImage.
            if (data && useTexStorage) {
                GL_CALL(TexSubImage2D(GR_GL_TEXTURE_2D,
                                      0, // level
                                      left, top,
                                      width, height,
                                      externalFormat, externalType,
                                      data));
            }
        }
    } else {
        if (swFlipY || glFlipY) {
            top = desc.fHeight - (top + height);
        }
        GL_CALL(TexSubImage2D(GR_GL_TEXTURE_2D,
                              0, // level
                              left, top,
                              width, height,
                              externalFormat, externalType, data));
    }

    if (restoreGLRowLength) {
        SkASSERT(this->glCaps().unpackRowLengthSupport());
        GL_CALL(PixelStorei(GR_GL_UNPACK_ROW_LENGTH, 0));
    }
    if (glFlipY) {
        GL_CALL(PixelStorei(GR_GL_UNPACK_FLIP_Y, GR_GL_FALSE));
    }
    return succeeded;
}

// TODO: This function is using a lot of wonky semantics like, if width == -1
// then set width = desc.fWdith ... blah. A better way to do it might be to
// create a CompressedTexData struct that takes a desc/ptr and figures out
// the proper upload semantics. Then users can construct this function how they
// see fit if they want to go against the "standard" way to do it.
bool GrGLGpu::uploadCompressedTexData(const GrSurfaceDesc& desc,
                                      const void* data,
                                      bool isNewTexture,
                                      int left, int top, int width, int height) {
    SkASSERT(data || isNewTexture);

    // No support for software flip y, yet...
    SkASSERT(kBottomLeft_GrSurfaceOrigin != desc.fOrigin);

    if (-1 == width) {
        width = desc.fWidth;
    }
#ifdef SK_DEBUG
    else {
        SkASSERT(width <= desc.fWidth);
    }
#endif

    if (-1 == height) {
        height = desc.fHeight;
    }
#ifdef SK_DEBUG
    else {
        SkASSERT(height <= desc.fHeight);
    }
#endif

    // Make sure that the width and height that we pass to OpenGL
    // is a multiple of the block size.
    size_t dataSize = GrCompressedFormatDataSize(desc.fConfig, width, height);

    // We only need the internal format for compressed 2D textures.
    GrGLenum internalFormat = 0;
    if (!this->configToGLFormats(desc.fConfig, false, &internalFormat, NULL, NULL)) {
        return false;
    }

    if (isNewTexture) {
        CLEAR_ERROR_BEFORE_ALLOC(this->glInterface());
        GL_ALLOC_CALL(this->glInterface(),
                      CompressedTexImage2D(GR_GL_TEXTURE_2D,
                                           0, // level
                                           internalFormat,
                                           width, height,
                                           0, // border
                                           SkToInt(dataSize),
                                           data));
        GrGLenum error = check_alloc_error(desc, this->glInterface());
        if (error != GR_GL_NO_ERROR) {
            return false;
        }
    } else {
        // Paletted textures can't be updated.
        if (GR_GL_PALETTE8_RGBA8 == internalFormat) {
            return false;
        }
        GL_CALL(CompressedTexSubImage2D(GR_GL_TEXTURE_2D,
                                        0, // level
                                        left, top,
                                        width, height,
                                        internalFormat,
                                        SkToInt(dataSize),
                                        data));
    }

    return true;
}

static bool renderbuffer_storage_msaa(GrGLContext& ctx,
                                      int sampleCount,
                                      GrGLenum format,
                                      int width, int height) {
    CLEAR_ERROR_BEFORE_ALLOC(ctx.interface());
    SkASSERT(GrGLCaps::kNone_MSFBOType != ctx.caps()->msFBOType());
    switch (ctx.caps()->msFBOType()) {
        case GrGLCaps::kDesktop_ARB_MSFBOType:
        case GrGLCaps::kDesktop_EXT_MSFBOType:
        case GrGLCaps::kES_3_0_MSFBOType:
            GL_ALLOC_CALL(ctx.interface(),
                            RenderbufferStorageMultisample(GR_GL_RENDERBUFFER,
                                                            sampleCount,
                                                            format,
                                                            width, height));
            break;
        case GrGLCaps::kES_Apple_MSFBOType:
            GL_ALLOC_CALL(ctx.interface(),
                            RenderbufferStorageMultisampleES2APPLE(GR_GL_RENDERBUFFER,
                                                                    sampleCount,
                                                                    format,
                                                                    width, height));
            break;
        case GrGLCaps::kES_EXT_MsToTexture_MSFBOType:
        case GrGLCaps::kES_IMG_MsToTexture_MSFBOType:
            GL_ALLOC_CALL(ctx.interface(),
                            RenderbufferStorageMultisampleES2EXT(GR_GL_RENDERBUFFER,
                                                                sampleCount,
                                                                format,
                                                                width, height));
            break;
        case GrGLCaps::kNone_MSFBOType:
            SkFAIL("Shouldn't be here if we don't support multisampled renderbuffers.");
            break;
    }
    return (GR_GL_NO_ERROR == CHECK_ALLOC_ERROR(ctx.interface()));
}

bool GrGLGpu::createRenderTargetObjects(const GrSurfaceDesc& desc, bool budgeted, GrGLuint texID,
                                        GrGLRenderTarget::IDDesc* idDesc) {
    idDesc->fMSColorRenderbufferID = 0;
    idDesc->fLifeCycle = budgeted ? GrGpuResource::kCached_LifeCycle :
                                    GrGpuResource::kUncached_LifeCycle;

    GrGLenum status;

    GrGLenum msColorFormat = 0; // suppress warning
    GrGLenum fboTarget = 0; // suppress warning

    if (desc.fSampleCnt > 0 && GrGLCaps::kNone_MSFBOType == this->glCaps().msFBOType()) {
        goto FAILED;
    }

    idDesc->fTextureFBO.reset(SkNEW_ARGS(GrGLFBO, (this->glInterface())));
    if (!idDesc->fTextureFBO->isValid()) {
        goto FAILED;
    }

    // If we are using multisampling we will create two FBOS. We render to one and then resolve to
    // the texture bound to the other. The exception is the IMG multisample extension. With this
    // extension the texture is multisampled when rendered to and then auto-resolves it when it is
    // rendered from.
    if (desc.fSampleCnt > 0 && this->glCaps().usesMSAARenderBuffers()) {
        idDesc->fRenderFBO.reset(SkNEW_ARGS(GrGLFBO, (this->glInterface())));
        if (!idDesc->fRenderFBO->isValid()) {
            goto FAILED;
        }
        GL_CALL(GenRenderbuffers(1, &idDesc->fMSColorRenderbufferID));
        if (!idDesc->fMSColorRenderbufferID ||
            !this->configToGLFormats(desc.fConfig,
                                     // ES2 and ES3 require sized internal formats for rb storage.
                                     kGLES_GrGLStandard == this->glStandard(),
                                     &msColorFormat,
                                     NULL,
                                     NULL)) {
            goto FAILED;
        }
    } else {
        idDesc->fRenderFBO.reset(SkRef(idDesc->fTextureFBO.get()));
    }

    if (idDesc->fRenderFBO != idDesc->fTextureFBO) {
        SkASSERT(desc.fSampleCnt > 0);
        GL_CALL(BindRenderbuffer(GR_GL_RENDERBUFFER, idDesc->fMSColorRenderbufferID));
        if (!renderbuffer_storage_msaa(fGLContext,
                                       desc.fSampleCnt,
                                       msColorFormat,
                                       desc.fWidth, desc.fHeight)) {
            goto FAILED;
        }
        fboTarget = this->bindFBO(kChangeAttachments_FBOBinding, idDesc->fRenderFBO);
        GL_CALL(FramebufferRenderbuffer(fboTarget,
                                        GR_GL_COLOR_ATTACHMENT0,
                                        GR_GL_RENDERBUFFER,
                                        idDesc->fMSColorRenderbufferID));
        if ((desc.fFlags & kCheckAllocation_GrSurfaceFlag) ||
            !this->glCaps().isConfigVerifiedColorAttachment(desc.fConfig)) {
            GL_CALL_RET(status, CheckFramebufferStatus(GR_GL_FRAMEBUFFER));
            if (status != GR_GL_FRAMEBUFFER_COMPLETE) {
                goto FAILED;
            }
            fGLContext.caps()->markConfigAsValidColorAttachment(desc.fConfig);
        }
    }
    fboTarget = this->bindFBO(kChangeAttachments_FBOBinding, idDesc->fTextureFBO);

    if (this->glCaps().usesImplicitMSAAResolve() && desc.fSampleCnt > 0) {
        GL_CALL(FramebufferTexture2DMultisample(fboTarget,
                                                GR_GL_COLOR_ATTACHMENT0,
                                                GR_GL_TEXTURE_2D,
                                                texID, 0, desc.fSampleCnt));
    } else {
        GL_CALL(FramebufferTexture2D(fboTarget,
                                     GR_GL_COLOR_ATTACHMENT0,
                                     GR_GL_TEXTURE_2D,
                                     texID, 0));
    }
    if ((desc.fFlags & kCheckAllocation_GrSurfaceFlag) ||
        !this->glCaps().isConfigVerifiedColorAttachment(desc.fConfig)) {
        GL_CALL_RET(status, CheckFramebufferStatus(fboTarget));
        if (status != GR_GL_FRAMEBUFFER_COMPLETE) {
            goto FAILED;
        }
        fGLContext.caps()->markConfigAsValidColorAttachment(desc.fConfig);
    }

    return true;

FAILED:
    if (idDesc->fMSColorRenderbufferID) {
        GL_CALL(DeleteRenderbuffers(1, &idDesc->fMSColorRenderbufferID));
    }
    if (idDesc->fRenderFBO) {
        idDesc->fRenderFBO->release(this->glInterface());
    }
    if (idDesc->fTextureFBO) {
        idDesc->fTextureFBO->release(this->glInterface());
    }
    return false;
}

// good to set a break-point here to know when createTexture fails
static GrTexture* return_null_texture() {
//    SkDEBUGFAIL("null texture");
    return NULL;
}

#if 0 && defined(SK_DEBUG)
static size_t as_size_t(int x) {
    return x;
}
#endif

GrTexture* GrGLGpu::onCreateTexture(const GrSurfaceDesc& origDesc, bool budgeted,
                                    const void* srcData, size_t rowBytes) {

    GrSurfaceDesc desc = origDesc;

    // Attempt to catch un- or wrongly initialized sample counts;
    SkASSERT(desc.fSampleCnt >= 0 && desc.fSampleCnt <= 64);
    // We fail if the MSAA was requested and is not available.
    if (GrGLCaps::kNone_MSFBOType == this->glCaps().msFBOType() && desc.fSampleCnt) {
        //SkDebugf("MSAA RT requested but not supported on this platform.");
        return return_null_texture();
    }

    bool renderTarget = SkToBool(desc.fFlags & kRenderTarget_GrSurfaceFlag);

    // If the sample count exceeds the max then we clamp it.
    desc.fSampleCnt = SkTMin(desc.fSampleCnt, this->caps()->maxSampleCount());
    desc.fOrigin = resolve_origin(desc.fOrigin, renderTarget);

    if (GrGLCaps::kNone_MSFBOType == this->glCaps().msFBOType() && desc.fSampleCnt) {
        //SkDebugf("MSAA RT requested but not supported on this platform.");
        return return_null_texture();
    }

    if (renderTarget) {
        int maxRTSize = this->caps()->maxRenderTargetSize();
        if (desc.fWidth > maxRTSize || desc.fHeight > maxRTSize) {
            return return_null_texture();
        }
    } else {
        int maxSize = this->caps()->maxTextureSize();
        if (desc.fWidth > maxSize || desc.fHeight > maxSize) {
            return return_null_texture();
        }
    }

    GrGLTexture::IDDesc idDesc;
    GL_CALL(GenTextures(1, &idDesc.fTextureID));
    idDesc.fLifeCycle = budgeted ? GrGpuResource::kCached_LifeCycle :
                                   GrGpuResource::kUncached_LifeCycle;

    if (!idDesc.fTextureID) {
        return return_null_texture();
    }

    this->setScratchTextureUnit();
    GL_CALL(BindTexture(GR_GL_TEXTURE_2D, idDesc.fTextureID));

    if (renderTarget && this->glCaps().textureUsageSupport()) {
        // provides a hint about how this texture will be used
        GL_CALL(TexParameteri(GR_GL_TEXTURE_2D,
                              GR_GL_TEXTURE_USAGE,
                              GR_GL_FRAMEBUFFER_ATTACHMENT));
    }

    // Some drivers like to know filter/wrap before seeing glTexImage2D. Some
    // drivers have a bug where an FBO won't be complete if it includes a
    // texture that is not mipmap complete (considering the filter in use).
    GrGLTexture::TexParams initialTexParams;
    // we only set a subset here so invalidate first
    initialTexParams.invalidate();
    initialTexParams.fMinFilter = GR_GL_NEAREST;
    initialTexParams.fMagFilter = GR_GL_NEAREST;
    initialTexParams.fWrapS = GR_GL_CLAMP_TO_EDGE;
    initialTexParams.fWrapT = GR_GL_CLAMP_TO_EDGE;
    GL_CALL(TexParameteri(GR_GL_TEXTURE_2D,
                          GR_GL_TEXTURE_MAG_FILTER,
                          initialTexParams.fMagFilter));
    GL_CALL(TexParameteri(GR_GL_TEXTURE_2D,
                          GR_GL_TEXTURE_MIN_FILTER,
                          initialTexParams.fMinFilter));
    GL_CALL(TexParameteri(GR_GL_TEXTURE_2D,
                          GR_GL_TEXTURE_WRAP_S,
                          initialTexParams.fWrapS));
    GL_CALL(TexParameteri(GR_GL_TEXTURE_2D,
                          GR_GL_TEXTURE_WRAP_T,
                          initialTexParams.fWrapT));
    if (!this->uploadTexData(desc, true, 0, 0,
                             desc.fWidth, desc.fHeight,
                             desc.fConfig, srcData, rowBytes)) {
        GL_CALL(DeleteTextures(1, &idDesc.fTextureID));
        return return_null_texture();
    }

    GrGLTexture* tex;
    if (renderTarget) {
        // unbind the texture from the texture unit before binding it to the frame buffer
        GL_CALL(BindTexture(GR_GL_TEXTURE_2D, 0));
        GrGLRenderTarget::IDDesc rtIDDesc;

        if (!this->createRenderTargetObjects(desc, budgeted, idDesc.fTextureID, &rtIDDesc)) {
            GL_CALL(DeleteTextures(1, &idDesc.fTextureID));
            return return_null_texture();
        }
        tex = SkNEW_ARGS(GrGLTextureRenderTarget, (this, desc, idDesc, rtIDDesc));
    } else {
        tex = SkNEW_ARGS(GrGLTexture, (this, desc, idDesc));
    }
    tex->setCachedTexParams(initialTexParams, this->getResetTimestamp());
#ifdef TRACE_TEXTURE_CREATION
    SkDebugf("--- new texture [%d] size=(%d %d) config=%d\n",
             glTexDesc.fTextureID, desc.fWidth, desc.fHeight, desc.fConfig);
#endif
    return tex;
}

GrTexture* GrGLGpu::onCreateCompressedTexture(const GrSurfaceDesc& origDesc, bool budgeted,
                                              const void* srcData) {

    if(SkToBool(origDesc.fFlags & kRenderTarget_GrSurfaceFlag) || origDesc.fSampleCnt > 0) {
        return return_null_texture();
    }

    // Make sure that we're not flipping Y.
    GrSurfaceOrigin texOrigin = resolve_origin(origDesc.fOrigin, false);
    if (kBottomLeft_GrSurfaceOrigin == texOrigin) {
        return return_null_texture();
    }
    GrSurfaceDesc desc = origDesc;
    desc.fOrigin = texOrigin;

    int maxSize = this->caps()->maxTextureSize();
    if (desc.fWidth > maxSize || desc.fHeight > maxSize) {
        return return_null_texture();
    }

    GrGLTexture::IDDesc idDesc;
    GL_CALL(GenTextures(1, &idDesc.fTextureID));
    idDesc.fLifeCycle = budgeted ? GrGpuResource::kCached_LifeCycle :
                                   GrGpuResource::kUncached_LifeCycle;

    if (!idDesc.fTextureID) {
        return return_null_texture();
    }

    this->setScratchTextureUnit();
    GL_CALL(BindTexture(GR_GL_TEXTURE_2D, idDesc.fTextureID));

    // Some drivers like to know filter/wrap before seeing glTexImage2D. Some
    // drivers have a bug where an FBO won't be complete if it includes a
    // texture that is not mipmap complete (considering the filter in use).
    GrGLTexture::TexParams initialTexParams;
    // we only set a subset here so invalidate first
    initialTexParams.invalidate();
    initialTexParams.fMinFilter = GR_GL_NEAREST;
    initialTexParams.fMagFilter = GR_GL_NEAREST;
    initialTexParams.fWrapS = GR_GL_CLAMP_TO_EDGE;
    initialTexParams.fWrapT = GR_GL_CLAMP_TO_EDGE;
    GL_CALL(TexParameteri(GR_GL_TEXTURE_2D,
                          GR_GL_TEXTURE_MAG_FILTER,
                          initialTexParams.fMagFilter));
    GL_CALL(TexParameteri(GR_GL_TEXTURE_2D,
                          GR_GL_TEXTURE_MIN_FILTER,
                          initialTexParams.fMinFilter));
    GL_CALL(TexParameteri(GR_GL_TEXTURE_2D,
                          GR_GL_TEXTURE_WRAP_S,
                          initialTexParams.fWrapS));
    GL_CALL(TexParameteri(GR_GL_TEXTURE_2D,
                          GR_GL_TEXTURE_WRAP_T,
                          initialTexParams.fWrapT));

    if (!this->uploadCompressedTexData(desc, srcData)) {
        GL_CALL(DeleteTextures(1, &idDesc.fTextureID));
        return return_null_texture();
    }

    GrGLTexture* tex;
    tex = SkNEW_ARGS(GrGLTexture, (this, desc, idDesc));
    tex->setCachedTexParams(initialTexParams, this->getResetTimestamp());
#ifdef TRACE_TEXTURE_CREATION
    SkDebugf("--- new compressed texture [%d] size=(%d %d) config=%d\n",
             glTexDesc.fTextureID, desc.fWidth, desc.fHeight, desc.fConfig);
#endif
    return tex;
}

namespace {

const GrGLuint kUnknownBitCount = GrGLStencilBuffer::kUnknownBitCount;

void inline get_stencil_rb_sizes(const GrGLInterface* gl,
                                 GrGLStencilBuffer::Format* format) {

    // we shouldn't ever know one size and not the other
    SkASSERT((kUnknownBitCount == format->fStencilBits) ==
             (kUnknownBitCount == format->fTotalBits));
    if (kUnknownBitCount == format->fStencilBits) {
        GR_GL_GetRenderbufferParameteriv(gl, GR_GL_RENDERBUFFER,
                                         GR_GL_RENDERBUFFER_STENCIL_SIZE,
                                         (GrGLint*)&format->fStencilBits);
        if (format->fPacked) {
            GR_GL_GetRenderbufferParameteriv(gl, GR_GL_RENDERBUFFER,
                                             GR_GL_RENDERBUFFER_DEPTH_SIZE,
                                             (GrGLint*)&format->fTotalBits);
            format->fTotalBits += format->fStencilBits;
        } else {
            format->fTotalBits = format->fStencilBits;
        }
    }
}
}

bool GrGLGpu::createStencilBufferForRenderTarget(GrRenderTarget* rt, int width, int height) {
    // All internally created RTs are also textures. We don't create
    // SBs for a client's standalone RT (that is a RT that isn't also a texture).
    SkASSERT(rt->asTexture());
    SkASSERT(width >= rt->width());
    SkASSERT(height >= rt->height());

    int samples = rt->numSamples();
    GrGLStencilBuffer::IDDesc sbDesc;

    int stencilFmtCnt = this->glCaps().stencilFormats().count();
    for (int i = 0; i < stencilFmtCnt; ++i) {
        if (!sbDesc.fRenderbufferID) {
            GL_CALL(GenRenderbuffers(1, &sbDesc.fRenderbufferID));
        }
        if (!sbDesc.fRenderbufferID) {
            return false;
        }
        GL_CALL(BindRenderbuffer(GR_GL_RENDERBUFFER, sbDesc.fRenderbufferID));
        // we start with the last stencil format that succeeded in hopes
        // that we won't go through this loop more than once after the
        // first (painful) stencil creation.
        int sIdx = (i + fLastSuccessfulStencilFmtIdx) % stencilFmtCnt;
        const GrGLCaps::StencilFormat& sFmt = this->glCaps().stencilFormats()[sIdx];
        CLEAR_ERROR_BEFORE_ALLOC(this->glInterface());
        // we do this "if" so that we don't call the multisample
        // version on a GL that doesn't have an MSAA extension.
        bool created;
        if (samples > 0) {
            created = renderbuffer_storage_msaa(fGLContext,
                                                samples,
                                                sFmt.fInternalFormat,
                                                width, height);
        } else {
            GL_ALLOC_CALL(this->glInterface(), RenderbufferStorage(GR_GL_RENDERBUFFER,
                                                                   sFmt.fInternalFormat,
                                                                   width, height));
            created = (GR_GL_NO_ERROR == check_alloc_error(rt->desc(), this->glInterface()));
        }
        if (created) {
            fStats.incStencilBufferCreates();
            // After sized formats we attempt an unsized format and take
            // whatever sizes GL gives us. In that case we query for the size.
            GrGLStencilBuffer::Format format = sFmt;
            get_stencil_rb_sizes(this->glInterface(), &format);
            SkAutoTUnref<GrGLStencilBuffer> sb(SkNEW_ARGS(GrGLStencilBuffer,
                                                  (this, sbDesc, width, height, samples, format)));
            if (this->attachStencilBufferToRenderTarget(sb, rt)) {
                fLastSuccessfulStencilFmtIdx = sIdx;
                rt->renderTargetPriv().didAttachStencilBuffer(sb);

                // Clear the stencil buffer. We use a special purpose FBO for this so that the
                // entire stencil buffer is cleared, even if it is attached to an FBO with a
                // smaller color target.
                if (!fStencilClearFBO) {
                    fStencilClearFBO.reset(SkNEW_ARGS(GrGLFBO, (this->glInterface())));
                }
                SkASSERT(fStencilClearFBO->isValid());
                GrGLenum fboTarget = this->bindFBO(kClear_FBOBinding, fStencilClearFBO);

                GL_CALL(FramebufferRenderbuffer(fboTarget,
                                                GR_GL_STENCIL_ATTACHMENT,
                                                GR_GL_RENDERBUFFER, sbDesc.fRenderbufferID));
                if (sFmt.fPacked) {
                    GL_CALL(FramebufferRenderbuffer(fboTarget,
                                                    GR_GL_DEPTH_ATTACHMENT,
                                                    GR_GL_RENDERBUFFER, sbDesc.fRenderbufferID));
                }

                GL_CALL(ClearStencil(0));
                // Many GL implementations seem to have trouble with clearing an FBO with only
                // a stencil buffer.
                GrGLuint tempRB;
                GL_CALL(GenRenderbuffers(1, &tempRB));
                GL_CALL(BindRenderbuffer(GR_GL_RENDERBUFFER, tempRB));

                if (samples > 0) {
                    renderbuffer_storage_msaa(fGLContext, samples, GR_GL_RGBA8, width, height);
                } else {
                    GL_CALL(RenderbufferStorage(GR_GL_RENDERBUFFER, GR_GL_RGBA8, width, height));
                }
                GL_CALL(FramebufferRenderbuffer(fboTarget,
                                                GR_GL_COLOR_ATTACHMENT0,
                                                GR_GL_RENDERBUFFER, tempRB));

                GL_CALL(Clear(GR_GL_STENCIL_BUFFER_BIT));

                GL_CALL(FramebufferRenderbuffer(fboTarget,
                                                GR_GL_COLOR_ATTACHMENT0,
                                                GR_GL_RENDERBUFFER, 0));
                GL_CALL(DeleteRenderbuffers(1, &tempRB));

                // Unbind the SB from the FBO so that we don't keep it alive.
                GL_CALL(FramebufferRenderbuffer(fboTarget,
                                                GR_GL_STENCIL_ATTACHMENT,
                                                GR_GL_RENDERBUFFER, 0));
                if (sFmt.fPacked) {
                    GL_CALL(FramebufferRenderbuffer(fboTarget,
                                                    GR_GL_DEPTH_ATTACHMENT,
                                                    GR_GL_RENDERBUFFER, 0));
                }

                return true;
            }
            // Remove the scratch key from this resource so we don't grab it from the cache ever
            // again.
            sb->resourcePriv().removeScratchKey();
            // Set this to 0 since we handed the valid ID off to the failed stencil buffer resource.
            sbDesc.fRenderbufferID = 0;
        }
    }
    GL_CALL(DeleteRenderbuffers(1, &sbDesc.fRenderbufferID));
    return false;
}

bool GrGLGpu::attachStencilBufferToRenderTarget(GrStencilBuffer* sb, GrRenderTarget* rt) {
    GrGLRenderTarget* glrt = static_cast<GrGLRenderTarget*>(rt);
    if (NULL == sb) {
        if (rt->renderTargetPriv().getStencilBuffer()) {
            GL_CALL(FramebufferRenderbuffer(GR_GL_FRAMEBUFFER,
                                            GR_GL_STENCIL_ATTACHMENT,
                                            GR_GL_RENDERBUFFER, 0));
            GL_CALL(FramebufferRenderbuffer(GR_GL_FRAMEBUFFER,
                                            GR_GL_DEPTH_ATTACHMENT,
                                            GR_GL_RENDERBUFFER, 0));
#ifdef SK_DEBUG
            GrGLenum status;
            GL_CALL_RET(status, CheckFramebufferStatus(GR_GL_FRAMEBUFFER));
            SkASSERT(GR_GL_FRAMEBUFFER_COMPLETE == status);
#endif
        }
        return true;
    } else {
        GrGLStencilBuffer* glsb = static_cast<GrGLStencilBuffer*>(sb);
        GrGLuint rb = glsb->renderbufferID();
        GrGLenum fboTarget = this->bindFBO(kChangeAttachments_FBOBinding, glrt->renderFBO());

        GL_CALL(FramebufferRenderbuffer(fboTarget,
                                        GR_GL_STENCIL_ATTACHMENT,
                                        GR_GL_RENDERBUFFER, rb));
        if (glsb->format().fPacked) {
            GL_CALL(FramebufferRenderbuffer(fboTarget,
                                            GR_GL_DEPTH_ATTACHMENT,
                                            GR_GL_RENDERBUFFER, rb));
        } else {
            GL_CALL(FramebufferRenderbuffer(fboTarget,
                                            GR_GL_DEPTH_ATTACHMENT,
                                            GR_GL_RENDERBUFFER, 0));
        }

        GrGLenum status;
        if (!this->glCaps().isColorConfigAndStencilFormatVerified(rt->config(), glsb->format())) {
            GL_CALL_RET(status, CheckFramebufferStatus(GR_GL_FRAMEBUFFER));
            if (status != GR_GL_FRAMEBUFFER_COMPLETE) {
                GL_CALL(FramebufferRenderbuffer(fboTarget,
                                                GR_GL_STENCIL_ATTACHMENT,
                                                GR_GL_RENDERBUFFER, 0));
                if (glsb->format().fPacked) {
                    GL_CALL(FramebufferRenderbuffer(fboTarget,
                                                    GR_GL_DEPTH_ATTACHMENT,
                                                    GR_GL_RENDERBUFFER, 0));
                }
                return false;
            } else {
                fGLContext.caps()->markColorConfigAndStencilFormatAsVerified(
                    rt->config(),
                    glsb->format());
            }
        }
        return true;
    }
}

////////////////////////////////////////////////////////////////////////////////

GrVertexBuffer* GrGLGpu::onCreateVertexBuffer(size_t size, bool dynamic) {
    GrGLVertexBuffer::Desc desc;
    desc.fDynamic = dynamic;
    desc.fSizeInBytes = size;

    if (this->glCaps().useNonVBOVertexAndIndexDynamicData() && desc.fDynamic) {
        desc.fID = 0;
        GrGLVertexBuffer* vertexBuffer = SkNEW_ARGS(GrGLVertexBuffer, (this, desc));
        return vertexBuffer;
    } else {
        GL_CALL(GenBuffers(1, &desc.fID));
        if (desc.fID) {
            fHWGeometryState.setVertexBufferID(this, desc.fID);
            CLEAR_ERROR_BEFORE_ALLOC(this->glInterface());
            // make sure driver can allocate memory for this buffer
            GL_ALLOC_CALL(this->glInterface(),
                          BufferData(GR_GL_ARRAY_BUFFER,
                                     (GrGLsizeiptr) desc.fSizeInBytes,
                                     NULL,   // data ptr
                                     desc.fDynamic ? GR_GL_DYNAMIC_DRAW : GR_GL_STATIC_DRAW));
            if (CHECK_ALLOC_ERROR(this->glInterface()) != GR_GL_NO_ERROR) {
                GL_CALL(DeleteBuffers(1, &desc.fID));
                this->notifyVertexBufferDelete(desc.fID);
                return NULL;
            }
            GrGLVertexBuffer* vertexBuffer = SkNEW_ARGS(GrGLVertexBuffer, (this, desc));
            return vertexBuffer;
        }
        return NULL;
    }
}

GrIndexBuffer* GrGLGpu::onCreateIndexBuffer(size_t size, bool dynamic) {
    GrGLIndexBuffer::Desc desc;
    desc.fDynamic = dynamic;
    desc.fSizeInBytes = size;

    if (this->glCaps().useNonVBOVertexAndIndexDynamicData() && desc.fDynamic) {
        desc.fID = 0;
        GrIndexBuffer* indexBuffer = SkNEW_ARGS(GrGLIndexBuffer, (this, desc));
        return indexBuffer;
    } else {
        GL_CALL(GenBuffers(1, &desc.fID));
        if (desc.fID) {
            fHWGeometryState.setIndexBufferIDOnDefaultVertexArray(this, desc.fID);
            CLEAR_ERROR_BEFORE_ALLOC(this->glInterface());
            // make sure driver can allocate memory for this buffer
            GL_ALLOC_CALL(this->glInterface(),
                          BufferData(GR_GL_ELEMENT_ARRAY_BUFFER,
                                     (GrGLsizeiptr) desc.fSizeInBytes,
                                     NULL,  // data ptr
                                     desc.fDynamic ? GR_GL_DYNAMIC_DRAW : GR_GL_STATIC_DRAW));
            if (CHECK_ALLOC_ERROR(this->glInterface()) != GR_GL_NO_ERROR) {
                GL_CALL(DeleteBuffers(1, &desc.fID));
                this->notifyIndexBufferDelete(desc.fID);
                return NULL;
            }
            GrIndexBuffer* indexBuffer = SkNEW_ARGS(GrGLIndexBuffer, (this, desc));
            return indexBuffer;
        }
        return NULL;
    }
}

void GrGLGpu::flushScissor(const GrScissorState& scissorState,
                           const GrGLIRect& rtViewport,
                           GrSurfaceOrigin rtOrigin) {
    if (scissorState.enabled()) {
        GrGLIRect scissor;
        scissor.setRelativeTo(rtViewport,
                              scissorState.rect().fLeft,
                              scissorState.rect().fTop,
                              scissorState.rect().width(),
                              scissorState.rect().height(),
                              rtOrigin);
        // if the scissor fully contains the viewport then we fall through and
        // disable the scissor test.
        if (!scissor.contains(rtViewport)) {
            if (fHWScissorSettings.fRect != scissor) {
                scissor.pushToGLScissor(this->glInterface());
                fHWScissorSettings.fRect = scissor;
            }
            if (kYes_TriState != fHWScissorSettings.fEnabled) {
                GL_CALL(Enable(GR_GL_SCISSOR_TEST));
                fHWScissorSettings.fEnabled = kYes_TriState;
            }
            return;
        }
    }

    // See fall through note above
    this->disableScissor();
}

bool GrGLGpu::flushGLState(const DrawArgs& args) {
    GrXferProcessor::BlendInfo blendInfo;
    const GrPipeline& pipeline = *args.fPipeline;
    args.fPipeline->getXferProcessor()->getBlendInfo(&blendInfo);

    this->flushDither(pipeline.isDitherState());
    this->flushColorWrite(blendInfo.fWriteColor);
    this->flushDrawFace(pipeline.getDrawFace());

    fCurrentProgram.reset(fProgramCache->getProgram(args));
    if (NULL == fCurrentProgram.get()) {
        SkDEBUGFAIL("Failed to create program!");
        return false;
    }

    fCurrentProgram.get()->ref();

    GrGLuint programID = fCurrentProgram->programID();
    if (fHWProgramID != programID) {
        GL_CALL(UseProgram(programID));
        fHWProgramID = programID;
    }


    fCurrentProgram->setData(*args.fPrimitiveProcessor, pipeline, *args.fBatchTracker);

    GrGLRenderTarget* glRT = static_cast<GrGLRenderTarget*>(pipeline.getRenderTarget());

    this->flushStencil(pipeline.getStencil());
    this->flushScissor(pipeline.getScissorState(), glRT->getViewport(), glRT->origin());
    this->flushHWAAState(glRT, pipeline.isHWAntialiasState());

    // This must come after textures are flushed because a texture may need
    // to be msaa-resolved (which will modify bound FBO and scissor state).
    this->bindFBO(kDraw_FBOBinding, glRT->renderFBO());
    this->setViewport(glRT->getViewport());
    if (blendInfo.fWriteColor) {
        this->flushBlend(blendInfo);
        this->markSurfaceContentsDirty(glRT, NULL);
    }

    return true;
}

void GrGLGpu::setupGeometry(const GrPrimitiveProcessor& primProc,
                            const GrDrawTarget::DrawInfo& info,
                            size_t* indexOffsetInBytes) {
    GrGLVertexBuffer* vbuf;
    vbuf = (GrGLVertexBuffer*) info.vertexBuffer();

    SkASSERT(vbuf);
    SkASSERT(!vbuf->isMapped());

    GrGLIndexBuffer* ibuf = NULL;
    if (info.isIndexed()) {
        SkASSERT(indexOffsetInBytes);

        *indexOffsetInBytes = 0;
        ibuf = (GrGLIndexBuffer*)info.indexBuffer();

        SkASSERT(ibuf);
        SkASSERT(!ibuf->isMapped());
        *indexOffsetInBytes += ibuf->baseOffset();
    }
    GrGLAttribArrayState* attribState =
        fHWGeometryState.bindArrayAndBuffersToDraw(this, vbuf, ibuf);

    int vaCount = primProc.numAttribs();
    if (vaCount > 0) {

        GrGLsizei stride = static_cast<GrGLsizei>(primProc.getVertexStride());

        size_t vertexOffsetInBytes = stride * info.startVertex();

        vertexOffsetInBytes += vbuf->baseOffset();

        uint32_t usedAttribArraysMask = 0;
        size_t offset = 0;

        for (int attribIndex = 0; attribIndex < vaCount; attribIndex++) {
            const GrGeometryProcessor::Attribute& attrib = primProc.getAttrib(attribIndex);
            usedAttribArraysMask |= (1 << attribIndex);
            GrVertexAttribType attribType = attrib.fType;
            attribState->set(this,
                             attribIndex,
                             vbuf,
                             GrGLAttribTypeToLayout(attribType).fCount,
                             GrGLAttribTypeToLayout(attribType).fType,
                             GrGLAttribTypeToLayout(attribType).fNormalized,
                             stride,
                             reinterpret_cast<GrGLvoid*>(vertexOffsetInBytes + offset));
            offset += attrib.fOffset;
        }
        attribState->disableUnusedArrays(this, usedAttribArraysMask);
    }
}

void GrGLGpu::buildProgramDesc(GrProgramDesc* desc,
                               const GrPrimitiveProcessor& primProc,
                               const GrPipeline& pipeline,
                               const GrBatchTracker& batchTracker) const {
    if (!GrGLProgramDescBuilder::Build(desc, primProc, pipeline, this, batchTracker)) {
        SkDEBUGFAIL("Failed to generate GL program descriptor");
    }
}

void GrGLGpu::disableScissor() {
    if (kNo_TriState != fHWScissorSettings.fEnabled) {
        GL_CALL(Disable(GR_GL_SCISSOR_TEST));
        fHWScissorSettings.fEnabled = kNo_TriState;
        return;
    }
}

void GrGLGpu::onClear(GrRenderTarget* target, const SkIRect* rect, GrColor color,
                      bool canIgnoreRect) {
    // parent class should never let us get here with no RT
    SkASSERT(target);
    GrGLRenderTarget* glRT = static_cast<GrGLRenderTarget*>(target);

    if (canIgnoreRect && this->glCaps().fullClearIsFree()) {
        rect = NULL;
    }

    SkIRect clippedRect;
    if (rect) {
        // flushScissor expects rect to be clipped to the target.
        clippedRect = *rect;
        SkIRect rtRect = SkIRect::MakeWH(target->width(), target->height());
        if (clippedRect.intersect(rtRect)) {
            rect = &clippedRect;
        } else {
            return;
        }
    }

    this->bindFBO(kClear_FBOBinding, glRT->renderFBO());
    this->markSurfaceContentsDirty(glRT, rect);
    GrScissorState scissorState;
    if (rect) {
        scissorState.set(*rect);
    }
    this->flushScissor(scissorState, glRT->getViewport(), glRT->origin());

    GrGLfloat r, g, b, a;
    static const GrGLfloat scale255 = 1.f / 255.f;
    a = GrColorUnpackA(color) * scale255;
    GrGLfloat scaleRGB = scale255;
    r = GrColorUnpackR(color) * scaleRGB;
    g = GrColorUnpackG(color) * scaleRGB;
    b = GrColorUnpackB(color) * scaleRGB;

    GL_CALL(ColorMask(GR_GL_TRUE, GR_GL_TRUE, GR_GL_TRUE, GR_GL_TRUE));
    fHWWriteToColor = kYes_TriState;
    GL_CALL(ClearColor(r, g, b, a));
    GL_CALL(Clear(GR_GL_COLOR_BUFFER_BIT));
}

void GrGLGpu::discard(GrRenderTarget* renderTarget) {
    SkASSERT(renderTarget);
    if (!this->caps()->discardRenderTargetSupport()) {
        return;
    }

    GrGLRenderTarget* glRT = static_cast<GrGLRenderTarget*>(renderTarget);
    GrGLenum fboTarget = this->bindFBO(kDiscard_FBOBinding, glRT->renderFBO());
    switch (this->glCaps().invalidateFBType()) {
        case GrGLCaps::kNone_InvalidateFBType:
            SkFAIL("Should never get here.");
            break;
        case GrGLCaps::kInvalidate_InvalidateFBType:
            if (glRT->renderFBO()->isDefaultFramebuffer()) {
                //  When rendering to the default framebuffer the legal values for attachments
                //  are GL_COLOR, GL_DEPTH, GL_STENCIL, ... rather than the various FBO attachment
                //  types.
                static const GrGLenum attachments[] = { GR_GL_COLOR };
                GL_CALL(InvalidateFramebuffer(fboTarget, SK_ARRAY_COUNT(attachments),
                        attachments));
            } else {
                static const GrGLenum attachments[] = { GR_GL_COLOR_ATTACHMENT0 };
                GL_CALL(InvalidateFramebuffer(fboTarget, SK_ARRAY_COUNT(attachments),
                        attachments));
            }
            break;
        case GrGLCaps::kDiscard_InvalidateFBType: {
            if (glRT->renderFBO()->isDefaultFramebuffer()) {
                //  When rendering to the default framebuffer the legal values for attachments
                //  are GL_COLOR, GL_DEPTH, GL_STENCIL, ... rather than the various FBO attachment
                //  types. See glDiscardFramebuffer() spec.
                static const GrGLenum attachments[] = { GR_GL_COLOR };
                GL_CALL(DiscardFramebuffer(fboTarget, SK_ARRAY_COUNT(attachments),
                        attachments));
            } else {
                static const GrGLenum attachments[] = { GR_GL_COLOR_ATTACHMENT0 };
                GL_CALL(DiscardFramebuffer(fboTarget, SK_ARRAY_COUNT(attachments),
                        attachments));
            }
            break;
        }
    }
    renderTarget->flagAsResolved();
}


void GrGLGpu::clearStencil(GrRenderTarget* target) {
    if (NULL == target) {
        return;
    }
    GrGLRenderTarget* glRT = static_cast<GrGLRenderTarget*>(target);
    this->bindFBO(kClear_FBOBinding, glRT->renderFBO());

    this->disableScissor();

    GL_CALL(StencilMask(0xffffffff));
    GL_CALL(ClearStencil(0));
    GL_CALL(Clear(GR_GL_STENCIL_BUFFER_BIT));
    fHWStencilSettings.invalidate();
}

void GrGLGpu::onClearStencilClip(GrRenderTarget* target, const SkIRect& rect, bool insideClip) {
    SkASSERT(target);

    GrStencilBuffer* sb = target->renderTargetPriv().getStencilBuffer();
    // this should only be called internally when we know we have a
    // stencil buffer.
    SkASSERT(sb);
    GrGLint stencilBitCount =  sb->bits();
#if 0
    SkASSERT(stencilBitCount > 0);
    GrGLint clipStencilMask  = (1 << (stencilBitCount - 1));
#else
    // we could just clear the clip bit but when we go through
    // ANGLE a partial stencil mask will cause clears to be
    // turned into draws. Our contract on GrDrawTarget says that
    // changing the clip between stencil passes may or may not
    // zero the client's clip bits. So we just clear the whole thing.
    static const GrGLint clipStencilMask  = ~0;
#endif
    GrGLint value;
    if (insideClip) {
        value = (1 << (stencilBitCount - 1));
    } else {
        value = 0;
    }
    GrGLRenderTarget* glRT = static_cast<GrGLRenderTarget*>(target);
    this->bindFBO(kClear_FBOBinding, glRT->renderFBO());

    GrScissorState scissorState;
    scissorState.set(rect);
    this->flushScissor(scissorState, glRT->getViewport(), glRT->origin());

    GL_CALL(StencilMask((uint32_t) clipStencilMask));
    GL_CALL(ClearStencil(value));
    GL_CALL(Clear(GR_GL_STENCIL_BUFFER_BIT));
    fHWStencilSettings.invalidate();
}

bool GrGLGpu::readPixelsWillPayForYFlip(GrRenderTarget* renderTarget,
                                        int left, int top,
                                        int width, int height,
                                        GrPixelConfig config,
                                        size_t rowBytes) const {
    // If this rendertarget is aready TopLeft, we don't need to flip.
    if (kTopLeft_GrSurfaceOrigin == renderTarget->origin()) {
        return false;
    }

    // if GL can do the flip then we'll never pay for it.
    if (this->glCaps().packFlipYSupport()) {
        return false;
    }

    // If we have to do memcpy to handle non-trim rowBytes then we
    // get the flip for free. Otherwise it costs.
    if (this->glCaps().packRowLengthSupport()) {
        return true;
    }
    // If we have to do memcpys to handle rowBytes then y-flip is free
    // Note the rowBytes might be tight to the passed in data, but if data
    // gets clipped in x to the target the rowBytes will no longer be tight.
    if (left >= 0 && (left + width) < renderTarget->width()) {
           return 0 == rowBytes ||
                  GrBytesPerPixel(config) * width == rowBytes;
    } else {
        return false;
    }
}

bool GrGLGpu::onReadPixels(GrRenderTarget* target,
                           int left, int top,
                           int width, int height,
                           GrPixelConfig config,
                           void* buffer,
                           size_t rowBytes) {
    // We cannot read pixels into a compressed buffer
    if (GrPixelConfigIsCompressed(config)) {
        return false;
    }

    GrGLenum format = 0;
    GrGLenum type = 0;
    bool flipY = kBottomLeft_GrSurfaceOrigin == target->origin();
    if (!this->configToGLFormats(config, false, NULL, &format, &type)) {
        return false;
    }
    size_t bpp = GrBytesPerPixel(config);
    if (!adjust_pixel_ops_params(target->width(), target->height(), bpp,
                                 &left, &top, &width, &height,
                                 const_cast<const void**>(&buffer),
                                 &rowBytes)) {
        return false;
    }

    // resolve the render target if necessary
    GrGLRenderTarget* tgt = static_cast<GrGLRenderTarget*>(target);
    if (tgt->getResolveType() == GrGLRenderTarget::kCantResolve_ResolveType) {
        return false;
    }
    if (tgt->getResolveType() == GrGLRenderTarget::kCanResolve_ResolveType) {
        this->onResolveRenderTarget(tgt);
    }
    this->bindFBO(kReadPixels_FBOBinding, tgt->textureFBO());

    const GrGLIRect& glvp = tgt->getViewport();

    // the read rect is viewport-relative
    GrGLIRect readRect;
    readRect.setRelativeTo(glvp, left, top, width, height, target->origin());

    size_t tightRowBytes = bpp * width;
    if (0 == rowBytes) {
        rowBytes = tightRowBytes;
    }
    size_t readDstRowBytes = tightRowBytes;
    void* readDst = buffer;

    // determine if GL can read using the passed rowBytes or if we need
    // a scratch buffer.
    GrAutoMalloc<32 * sizeof(GrColor)> scratch;
    if (rowBytes != tightRowBytes) {
        if (this->glCaps().packRowLengthSupport()) {
            SkASSERT(!(rowBytes % sizeof(GrColor)));
            GL_CALL(PixelStorei(GR_GL_PACK_ROW_LENGTH,
                                static_cast<GrGLint>(rowBytes / sizeof(GrColor))));
            readDstRowBytes = rowBytes;
        } else {
            scratch.reset(tightRowBytes * height);
            readDst = scratch.get();
        }
    }
    if (flipY && this->glCaps().packFlipYSupport()) {
        GL_CALL(PixelStorei(GR_GL_PACK_REVERSE_ROW_ORDER, 1));
    }
    GL_CALL(ReadPixels(readRect.fLeft, readRect.fBottom,
                       readRect.fWidth, readRect.fHeight,
                       format, type, readDst));
    if (readDstRowBytes != tightRowBytes) {
        SkASSERT(this->glCaps().packRowLengthSupport());
        GL_CALL(PixelStorei(GR_GL_PACK_ROW_LENGTH, 0));
    }
    if (flipY && this->glCaps().packFlipYSupport()) {
        GL_CALL(PixelStorei(GR_GL_PACK_REVERSE_ROW_ORDER, 0));
        flipY = false;
    }

    // now reverse the order of the rows, since GL's are bottom-to-top, but our
    // API presents top-to-bottom. We must preserve the padding contents. Note
    // that the above readPixels did not overwrite the padding.
    if (readDst == buffer) {
        SkASSERT(rowBytes == readDstRowBytes);
        if (flipY) {
            scratch.reset(tightRowBytes);
            void* tmpRow = scratch.get();
            // flip y in-place by rows
            const int halfY = height >> 1;
            char* top = reinterpret_cast<char*>(buffer);
            char* bottom = top + (height - 1) * rowBytes;
            for (int y = 0; y < halfY; y++) {
                memcpy(tmpRow, top, tightRowBytes);
                memcpy(top, bottom, tightRowBytes);
                memcpy(bottom, tmpRow, tightRowBytes);
                top += rowBytes;
                bottom -= rowBytes;
            }
        }
    } else {
        SkASSERT(readDst != buffer);
        SkASSERT(rowBytes != tightRowBytes);
        // copy from readDst to buffer while flipping y
        // const int halfY = height >> 1;
        const char* src = reinterpret_cast<const char*>(readDst);
        char* dst = reinterpret_cast<char*>(buffer);
        if (flipY) {
            dst += (height-1) * rowBytes;
        }
        for (int y = 0; y < height; y++) {
            memcpy(dst, src, tightRowBytes);
            src += readDstRowBytes;
            if (!flipY) {
                dst += rowBytes;
            } else {
                dst -= rowBytes;
            }
        }
    }
    return true;
}

GrGLenum GrGLGpu::bindFBO(FBOBinding binding, const GrGLFBO* fbo) {
    SkASSERT(fbo);
    SkASSERT(fbo->isValid());
    
    enum {
        kDraw = 0,
        kRead = 1
    };

    bool useGLFramebuffer = !this->glCaps().usesMSAARenderBuffers() ||
        (this->glCaps().preferBindingToReadAndDrawFramebuffer() &&
         kBlitSrc_FBOBinding != binding && kBlitDst_FBOBinding != binding);

    if (useGLFramebuffer) {
        SkASSERT(kBlitSrc_FBOBinding != binding);
        fStats.incRenderTargetBinds();
        GL_CALL(BindFramebuffer(GR_GL_FRAMEBUFFER, fbo->fboID()));
        fHWFBOBinding[kDraw].fFBO.reset(SkRef(fbo));
        fHWFBOBinding[kRead].fFBO.reset(SkRef(fbo));
        return GR_GL_FRAMEBUFFER;
    }
    GrGLenum target = 0;
    HWFBOBinding* hwFBOBinding = NULL;
    switch (binding) {
        case kDraw_FBOBinding:
        case kClear_FBOBinding:
        case kDiscard_FBOBinding:
        case kChangeAttachments_FBOBinding:
        case kBlitDst_FBOBinding:
            target = GR_GL_DRAW_FRAMEBUFFER;
            hwFBOBinding = &fHWFBOBinding[kDraw];
            break;

        case kReadPixels_FBOBinding:
        case kBlitSrc_FBOBinding:
        case kCopyTexSrc_FBOBinding:
            target = GR_GL_READ_FRAMEBUFFER;
            hwFBOBinding = &fHWFBOBinding[kRead];
            break;
    }
    fStats.incRenderTargetBinds();
    GL_CALL(BindFramebuffer(target, fbo->fboID()));
    hwFBOBinding->fFBO.reset(SkRef(fbo));
    return target;
}

void GrGLGpu::setViewport(const GrGLIRect& viewport) {
    if (viewport != fHWViewport) {
        viewport.pushToGLViewport(this->glInterface());
        fHWViewport = viewport;
    }
}

void GrGLGpu::markSurfaceContentsDirty(GrSurface* surface, const SkIRect* bounds) {
    if (NULL == bounds || !bounds->isEmpty()) {
        GrGLRenderTarget* rt = static_cast<GrGLRenderTarget*>(surface->asRenderTarget());
        if (rt) {
            rt->flagAsNeedingResolve(bounds);
        }
        GrGLTexture* texture = static_cast<GrGLTexture*>(surface->asTexture());
        if (texture) {
            texture->texturePriv().dirtyMipMaps(true);
        }
    }
}

GrGLenum gPrimitiveType2GLMode[] = {
    GR_GL_TRIANGLES,
    GR_GL_TRIANGLE_STRIP,
    GR_GL_TRIANGLE_FAN,
    GR_GL_POINTS,
    GR_GL_LINES,
    GR_GL_LINE_STRIP
};

#define SWAP_PER_DRAW 0

#if SWAP_PER_DRAW
    #if defined(SK_BUILD_FOR_MAC)
        #include <AGL/agl.h>
    #elif defined(SK_BUILD_FOR_WIN32)
        #include <gl/GL.h>
        void SwapBuf() {
            DWORD procID = GetCurrentProcessId();
            HWND hwnd = GetTopWindow(GetDesktopWindow());
            while(hwnd) {
                DWORD wndProcID = 0;
                GetWindowThreadProcessId(hwnd, &wndProcID);
                if(wndProcID == procID) {
                    SwapBuffers(GetDC(hwnd));
                }
                hwnd = GetNextWindow(hwnd, GW_HWNDNEXT);
            }
         }
    #endif
#endif

void GrGLGpu::onDraw(const DrawArgs& args, const GrDrawTarget::DrawInfo& info) {
    if (!this->flushGLState(args)) {
        return;
    }

    size_t indexOffsetInBytes = 0;
    this->setupGeometry(*args.fPrimitiveProcessor, info, &indexOffsetInBytes);

    SkASSERT((size_t)info.primitiveType() < SK_ARRAY_COUNT(gPrimitiveType2GLMode));

    if (info.isIndexed()) {
        GrGLvoid* indices =
            reinterpret_cast<GrGLvoid*>(indexOffsetInBytes + sizeof(uint16_t) * info.startIndex());
        // info.startVertex() was accounted for by setupGeometry.
        GL_CALL(DrawElements(gPrimitiveType2GLMode[info.primitiveType()],
                             info.indexCount(),
                             GR_GL_UNSIGNED_SHORT,
                             indices));
    } else {
        // Pass 0 for parameter first. We have to adjust glVertexAttribPointer() to account for
        // startVertex in the DrawElements case. So we always rely on setupGeometry to have
        // accounted for startVertex.
        GL_CALL(DrawArrays(gPrimitiveType2GLMode[info.primitiveType()], 0, info.vertexCount()));
    }
#if SWAP_PER_DRAW
    glFlush();
    #if defined(SK_BUILD_FOR_MAC)
        aglSwapBuffers(aglGetCurrentContext());
        int set_a_break_pt_here = 9;
        aglSwapBuffers(aglGetCurrentContext());
    #elif defined(SK_BUILD_FOR_WIN32)
        SwapBuf();
        int set_a_break_pt_here = 9;
        SwapBuf();
    #endif
#endif
}

void GrGLGpu::onStencilPath(const GrPath* path, const StencilPathState& state) {
    this->flushColorWrite(false);
    this->flushDrawFace(GrPipelineBuilder::kBoth_DrawFace);

    GrGLRenderTarget* rt = static_cast<GrGLRenderTarget*>(state.fRenderTarget);
    SkISize size = SkISize::Make(rt->width(), rt->height());
    this->glPathRendering()->setProjectionMatrix(*state.fViewMatrix, size, rt->origin());
    this->flushScissor(*state.fScissor, rt->getViewport(), rt->origin());
    this->flushHWAAState(rt, state.fUseHWAA);
    this->bindFBO(kDraw_FBOBinding, rt->renderFBO());
    this->setViewport(rt->getViewport());
    fPathRendering->stencilPath(path, *state.fStencil);
}

void GrGLGpu::onDrawPath(const DrawArgs& args, const GrPath* path,
                         const GrStencilSettings& stencil) {
    if (!this->flushGLState(args)) {
        return;
    }
    fPathRendering->drawPath(path, stencil);
}

void GrGLGpu::onDrawPaths(const DrawArgs& args,
                          const GrPathRange* pathRange,
                          const void* indices,
                          GrDrawTarget::PathIndexType indexType,
                          const float transformValues[],
                          GrDrawTarget::PathTransformType transformType,
                           int count,
                           const GrStencilSettings& stencil) {
    if (!this->flushGLState(args)) {
        return;
    }
    fPathRendering->drawPaths(pathRange, indices, indexType, transformValues,
                              transformType, count, stencil);
}

void GrGLGpu::onResolveRenderTarget(GrRenderTarget* target) {
    GrGLRenderTarget* rt = static_cast<GrGLRenderTarget*>(target);
    if (rt->needsResolve()) {
        // Some extensions automatically resolves the texture when it is read.
        if (this->glCaps().usesMSAARenderBuffers()) {
            SkASSERT(rt->textureFBO() != rt->renderFBO());
            this->bindFBO(kBlitSrc_FBOBinding, rt->renderFBO());
            this->bindFBO(kBlitDst_FBOBinding, rt->textureFBO());
            const GrGLIRect& vp = rt->getViewport();
            const SkIRect dirtyRect = rt->getResolveRect();

            if (GrGLCaps::kES_Apple_MSFBOType == this->glCaps().msFBOType()) {
                // Apple's extension uses the scissor as the blit bounds.
                GrScissorState scissorState;
                scissorState.set(dirtyRect);
                this->flushScissor(scissorState, vp, rt->origin());
                GL_CALL(ResolveMultisampleFramebuffer());
            } else {
                GrGLIRect r;
                r.setRelativeTo(vp, dirtyRect.fLeft, dirtyRect.fTop,
                                dirtyRect.width(), dirtyRect.height(), target->origin());

                int right = r.fLeft + r.fWidth;
                int top = r.fBottom + r.fHeight;

                // BlitFrameBuffer respects the scissor, so disable it.
                this->disableScissor();
                GL_CALL(BlitFramebuffer(r.fLeft, r.fBottom, right, top,
                                        r.fLeft, r.fBottom, right, top,
                                        GR_GL_COLOR_BUFFER_BIT, GR_GL_NEAREST));
            }
        }
        rt->flagAsResolved();
    }
}

namespace {


GrGLenum gr_to_gl_stencil_op(GrStencilOp op) {
    static const GrGLenum gTable[] = {
        GR_GL_KEEP,        // kKeep_StencilOp
        GR_GL_REPLACE,     // kReplace_StencilOp
        GR_GL_INCR_WRAP,   // kIncWrap_StencilOp
        GR_GL_INCR,        // kIncClamp_StencilOp
        GR_GL_DECR_WRAP,   // kDecWrap_StencilOp
        GR_GL_DECR,        // kDecClamp_StencilOp
        GR_GL_ZERO,        // kZero_StencilOp
        GR_GL_INVERT,      // kInvert_StencilOp
    };
    GR_STATIC_ASSERT(SK_ARRAY_COUNT(gTable) == kStencilOpCount);
    GR_STATIC_ASSERT(0 == kKeep_StencilOp);
    GR_STATIC_ASSERT(1 == kReplace_StencilOp);
    GR_STATIC_ASSERT(2 == kIncWrap_StencilOp);
    GR_STATIC_ASSERT(3 == kIncClamp_StencilOp);
    GR_STATIC_ASSERT(4 == kDecWrap_StencilOp);
    GR_STATIC_ASSERT(5 == kDecClamp_StencilOp);
    GR_STATIC_ASSERT(6 == kZero_StencilOp);
    GR_STATIC_ASSERT(7 == kInvert_StencilOp);
    SkASSERT((unsigned) op < kStencilOpCount);
    return gTable[op];
}

void set_gl_stencil(const GrGLInterface* gl,
                    const GrStencilSettings& settings,
                    GrGLenum glFace,
                    GrStencilSettings::Face grFace) {
    GrGLenum glFunc = GrToGLStencilFunc(settings.func(grFace));
    GrGLenum glFailOp = gr_to_gl_stencil_op(settings.failOp(grFace));
    GrGLenum glPassOp = gr_to_gl_stencil_op(settings.passOp(grFace));

    GrGLint ref = settings.funcRef(grFace);
    GrGLint mask = settings.funcMask(grFace);
    GrGLint writeMask = settings.writeMask(grFace);

    if (GR_GL_FRONT_AND_BACK == glFace) {
        // we call the combined func just in case separate stencil is not
        // supported.
        GR_GL_CALL(gl, StencilFunc(glFunc, ref, mask));
        GR_GL_CALL(gl, StencilMask(writeMask));
        GR_GL_CALL(gl, StencilOp(glFailOp, glPassOp, glPassOp));
    } else {
        GR_GL_CALL(gl, StencilFuncSeparate(glFace, glFunc, ref, mask));
        GR_GL_CALL(gl, StencilMaskSeparate(glFace, writeMask));
        GR_GL_CALL(gl, StencilOpSeparate(glFace, glFailOp, glPassOp, glPassOp));
    }
}
}

void GrGLGpu::flushStencil(const GrStencilSettings& stencilSettings) {
    if (fHWStencilSettings != stencilSettings) {
        if (stencilSettings.isDisabled()) {
            if (kNo_TriState != fHWStencilTestEnabled) {
                GL_CALL(Disable(GR_GL_STENCIL_TEST));
                fHWStencilTestEnabled = kNo_TriState;
            }
        } else {
            if (kYes_TriState != fHWStencilTestEnabled) {
                GL_CALL(Enable(GR_GL_STENCIL_TEST));
                fHWStencilTestEnabled = kYes_TriState;
            }
        }
        if (!stencilSettings.isDisabled()) {
            if (this->caps()->twoSidedStencilSupport()) {
                set_gl_stencil(this->glInterface(),
                               stencilSettings,
                               GR_GL_FRONT,
                               GrStencilSettings::kFront_Face);
                set_gl_stencil(this->glInterface(),
                               stencilSettings,
                               GR_GL_BACK,
                               GrStencilSettings::kBack_Face);
            } else {
                set_gl_stencil(this->glInterface(),
                               stencilSettings,
                               GR_GL_FRONT_AND_BACK,
                               GrStencilSettings::kFront_Face);
            }
        }
        fHWStencilSettings = stencilSettings;
    }
}

void GrGLGpu::flushHWAAState(GrRenderTarget* rt, bool useHWAA) {
    SkASSERT(!useHWAA || rt->isMultisampled());

    if (kGL_GrGLStandard == this->glStandard()) {
        if (useHWAA) {
            if (kYes_TriState != fMSAAEnabled) {
                GL_CALL(Enable(GR_GL_MULTISAMPLE));
                fMSAAEnabled = kYes_TriState;
            }
        } else {
            if (kNo_TriState != fMSAAEnabled) {
                GL_CALL(Disable(GR_GL_MULTISAMPLE));
                fMSAAEnabled = kNo_TriState;
            }
        }
    }
}

void GrGLGpu::flushBlend(const GrXferProcessor::BlendInfo& blendInfo) {
    // Any optimization to disable blending should have already been applied and
    // tweaked the coeffs to (1, 0).
    
    GrBlendCoeff srcCoeff = blendInfo.fSrcBlend;
    GrBlendCoeff dstCoeff = blendInfo.fDstBlend;
    bool blendOff = kOne_GrBlendCoeff == srcCoeff && kZero_GrBlendCoeff == dstCoeff;
    if (blendOff) {
        if (kNo_TriState != fHWBlendState.fEnabled) {
            GL_CALL(Disable(GR_GL_BLEND));
            fHWBlendState.fEnabled = kNo_TriState;
        }
    } else {
        if (kYes_TriState != fHWBlendState.fEnabled) {
            GL_CALL(Enable(GR_GL_BLEND));
            fHWBlendState.fEnabled = kYes_TriState;
        }
        if (fHWBlendState.fSrcCoeff != srcCoeff ||
            fHWBlendState.fDstCoeff != dstCoeff) {
            GL_CALL(BlendFunc(gXfermodeCoeff2Blend[srcCoeff],
                              gXfermodeCoeff2Blend[dstCoeff]));
            fHWBlendState.fSrcCoeff = srcCoeff;
            fHWBlendState.fDstCoeff = dstCoeff;
        }
        GrColor blendConst = blendInfo.fBlendConstant;
        if ((BlendCoeffReferencesConstant(srcCoeff) ||
             BlendCoeffReferencesConstant(dstCoeff)) &&
            (!fHWBlendState.fConstColorValid ||
             fHWBlendState.fConstColor != blendConst)) {
            GrGLfloat c[4];
            GrColorToRGBAFloat(blendConst, c);
            GL_CALL(BlendColor(c[0], c[1], c[2], c[3]));
            fHWBlendState.fConstColor = blendConst;
            fHWBlendState.fConstColorValid = true;
        }
    }
}

static inline GrGLenum tile_to_gl_wrap(SkShader::TileMode tm) {
    static const GrGLenum gWrapModes[] = {
        GR_GL_CLAMP_TO_EDGE,
        GR_GL_REPEAT,
        GR_GL_MIRRORED_REPEAT
    };
    GR_STATIC_ASSERT(SkShader::kTileModeCount == SK_ARRAY_COUNT(gWrapModes));
    GR_STATIC_ASSERT(0 == SkShader::kClamp_TileMode);
    GR_STATIC_ASSERT(1 == SkShader::kRepeat_TileMode);
    GR_STATIC_ASSERT(2 == SkShader::kMirror_TileMode);
    return gWrapModes[tm];
}

void GrGLGpu::bindTexture(int unitIdx, const GrTextureParams& params, GrGLTexture* texture) {
    SkASSERT(texture);

    // If we created a rt/tex and rendered to it without using a texture and now we're texturing
    // from the rt it will still be the last bound texture, but it needs resolving. So keep this
    // out of the "last != next" check.
    GrGLRenderTarget* texRT = static_cast<GrGLRenderTarget*>(texture->asRenderTarget());
    if (texRT) {
        this->onResolveRenderTarget(texRT);
    }

    uint32_t textureID = texture->getUniqueID();
    if (fHWBoundTextureUniqueIDs[unitIdx] != textureID) {
        this->setTextureUnit(unitIdx);
        GL_CALL(BindTexture(GR_GL_TEXTURE_2D, texture->textureID()));
        fHWBoundTextureUniqueIDs[unitIdx] = textureID;
    }

    ResetTimestamp timestamp;
    const GrGLTexture::TexParams& oldTexParams = texture->getCachedTexParams(&timestamp);
    bool setAll = timestamp < this->getResetTimestamp();
    GrGLTexture::TexParams newTexParams;

    static GrGLenum glMinFilterModes[] = {
        GR_GL_NEAREST,
        GR_GL_LINEAR,
        GR_GL_LINEAR_MIPMAP_LINEAR
    };
    static GrGLenum glMagFilterModes[] = {
        GR_GL_NEAREST,
        GR_GL_LINEAR,
        GR_GL_LINEAR
    };
    GrTextureParams::FilterMode filterMode = params.filterMode();

    if (GrTextureParams::kMipMap_FilterMode == filterMode) {
        if (!this->caps()->mipMapSupport() || GrPixelConfigIsCompressed(texture->config())) {
            filterMode = GrTextureParams::kBilerp_FilterMode;
        }
    }

    newTexParams.fMinFilter = glMinFilterModes[filterMode];
    newTexParams.fMagFilter = glMagFilterModes[filterMode];

    if (GrTextureParams::kMipMap_FilterMode == filterMode &&
        texture->texturePriv().mipMapsAreDirty()) {
        GL_CALL(GenerateMipmap(GR_GL_TEXTURE_2D));
        texture->texturePriv().dirtyMipMaps(false);
    }

    newTexParams.fWrapS = tile_to_gl_wrap(params.getTileModeX());
    newTexParams.fWrapT = tile_to_gl_wrap(params.getTileModeY());
    memcpy(newTexParams.fSwizzleRGBA,
           GrGLShaderBuilder::GetTexParamSwizzle(texture->config(), this->glCaps()),
           sizeof(newTexParams.fSwizzleRGBA));
    if (setAll || newTexParams.fMagFilter != oldTexParams.fMagFilter) {
        this->setTextureUnit(unitIdx);
        GL_CALL(TexParameteri(GR_GL_TEXTURE_2D,
                              GR_GL_TEXTURE_MAG_FILTER,
                              newTexParams.fMagFilter));
    }
    if (setAll || newTexParams.fMinFilter != oldTexParams.fMinFilter) {
        this->setTextureUnit(unitIdx);
        GL_CALL(TexParameteri(GR_GL_TEXTURE_2D,
                              GR_GL_TEXTURE_MIN_FILTER,
                              newTexParams.fMinFilter));
    }
    if (setAll || newTexParams.fWrapS != oldTexParams.fWrapS) {
        this->setTextureUnit(unitIdx);
        GL_CALL(TexParameteri(GR_GL_TEXTURE_2D,
                              GR_GL_TEXTURE_WRAP_S,
                              newTexParams.fWrapS));
    }
    if (setAll || newTexParams.fWrapT != oldTexParams.fWrapT) {
        this->setTextureUnit(unitIdx);
        GL_CALL(TexParameteri(GR_GL_TEXTURE_2D,
                              GR_GL_TEXTURE_WRAP_T,
                              newTexParams.fWrapT));
    }
    if (this->glCaps().textureSwizzleSupport() &&
        (setAll || memcmp(newTexParams.fSwizzleRGBA,
                          oldTexParams.fSwizzleRGBA,
                          sizeof(newTexParams.fSwizzleRGBA)))) {
        this->setTextureUnit(unitIdx);
        if (this->glStandard() == kGLES_GrGLStandard) {
            // ES3 added swizzle support but not GL_TEXTURE_SWIZZLE_RGBA.
            const GrGLenum* swizzle = newTexParams.fSwizzleRGBA;
            GL_CALL(TexParameteri(GR_GL_TEXTURE_2D, GR_GL_TEXTURE_SWIZZLE_R, swizzle[0]));
            GL_CALL(TexParameteri(GR_GL_TEXTURE_2D, GR_GL_TEXTURE_SWIZZLE_G, swizzle[1]));
            GL_CALL(TexParameteri(GR_GL_TEXTURE_2D, GR_GL_TEXTURE_SWIZZLE_B, swizzle[2]));
            GL_CALL(TexParameteri(GR_GL_TEXTURE_2D, GR_GL_TEXTURE_SWIZZLE_A, swizzle[3]));
        } else {
            GR_STATIC_ASSERT(sizeof(newTexParams.fSwizzleRGBA[0]) == sizeof(GrGLint));
            const GrGLint* swizzle = reinterpret_cast<const GrGLint*>(newTexParams.fSwizzleRGBA);
            GL_CALL(TexParameteriv(GR_GL_TEXTURE_2D, GR_GL_TEXTURE_SWIZZLE_RGBA, swizzle));
        }
    }
    texture->setCachedTexParams(newTexParams, this->getResetTimestamp());
}

void GrGLGpu::flushDither(bool dither) {
    if (dither) {
        if (kYes_TriState != fHWDitherEnabled) {
            GL_CALL(Enable(GR_GL_DITHER));
            fHWDitherEnabled = kYes_TriState;
        }
    } else {
        if (kNo_TriState != fHWDitherEnabled) {
            GL_CALL(Disable(GR_GL_DITHER));
            fHWDitherEnabled = kNo_TriState;
        }
    }
}

void GrGLGpu::flushColorWrite(bool writeColor) {
    if (!writeColor) {
        if (kNo_TriState != fHWWriteToColor) {
            GL_CALL(ColorMask(GR_GL_FALSE, GR_GL_FALSE,
                              GR_GL_FALSE, GR_GL_FALSE));
            fHWWriteToColor = kNo_TriState;
        }
    } else {
        if (kYes_TriState != fHWWriteToColor) {
            GL_CALL(ColorMask(GR_GL_TRUE, GR_GL_TRUE, GR_GL_TRUE, GR_GL_TRUE));
            fHWWriteToColor = kYes_TriState;
        }
    }
}

void GrGLGpu::flushDrawFace(GrPipelineBuilder::DrawFace face) {
    if (fHWDrawFace != face) {
        switch (face) {
            case GrPipelineBuilder::kCCW_DrawFace:
                GL_CALL(Enable(GR_GL_CULL_FACE));
                GL_CALL(CullFace(GR_GL_BACK));
                break;
            case GrPipelineBuilder::kCW_DrawFace:
                GL_CALL(Enable(GR_GL_CULL_FACE));
                GL_CALL(CullFace(GR_GL_FRONT));
                break;
            case GrPipelineBuilder::kBoth_DrawFace:
                GL_CALL(Disable(GR_GL_CULL_FACE));
                break;
            default:
                SkFAIL("Unknown draw face.");
        }
        fHWDrawFace = face;
    }
}

bool GrGLGpu::configToGLFormats(GrPixelConfig config,
                                bool getSizedInternalFormat,
                                GrGLenum* internalFormat,
                                GrGLenum* externalFormat,
                                GrGLenum* externalType) {
    GrGLenum dontCare;
    if (NULL == internalFormat) {
        internalFormat = &dontCare;
    }
    if (NULL == externalFormat) {
        externalFormat = &dontCare;
    }
    if (NULL == externalType) {
        externalType = &dontCare;
    }

    if(!this->glCaps().isConfigTexturable(config)) {
        return false;
    }

    switch (config) {
        case kRGBA_8888_GrPixelConfig:
            *internalFormat = GR_GL_RGBA;
            *externalFormat = GR_GL_RGBA;
            if (getSizedInternalFormat) {
                *internalFormat = GR_GL_RGBA8;
            } else {
                *internalFormat = GR_GL_RGBA;
            }
            *externalType = GR_GL_UNSIGNED_BYTE;
            break;
        case kBGRA_8888_GrPixelConfig:
            if (this->glCaps().bgraIsInternalFormat()) {
                if (getSizedInternalFormat) {
                    *internalFormat = GR_GL_BGRA8;
                } else {
                    *internalFormat = GR_GL_BGRA;
                }
            } else {
                if (getSizedInternalFormat) {
                    *internalFormat = GR_GL_RGBA8;
                } else {
                    *internalFormat = GR_GL_RGBA;
                }
            }
            *externalFormat = GR_GL_BGRA;
            *externalType = GR_GL_UNSIGNED_BYTE;
            break;
        case kSRGBA_8888_GrPixelConfig:
            *internalFormat = GR_GL_SRGB_ALPHA;
            *externalFormat = GR_GL_SRGB_ALPHA;
            if (getSizedInternalFormat) {
                *internalFormat = GR_GL_SRGB8_ALPHA8;
            } else {
                *internalFormat = GR_GL_SRGB_ALPHA;
            }
            *externalType = GR_GL_UNSIGNED_BYTE;
            break;
        case kRGB_565_GrPixelConfig:
            *internalFormat = GR_GL_RGB;
            *externalFormat = GR_GL_RGB;
            if (getSizedInternalFormat) {
                if (!this->glCaps().ES2CompatibilitySupport()) {
                    *internalFormat = GR_GL_RGB5;
                } else {
                    *internalFormat = GR_GL_RGB565;
                }
            } else {
                *internalFormat = GR_GL_RGB;
            }
            *externalType = GR_GL_UNSIGNED_SHORT_5_6_5;
            break;
        case kRGBA_4444_GrPixelConfig:
            *internalFormat = GR_GL_RGBA;
            *externalFormat = GR_GL_RGBA;
            if (getSizedInternalFormat) {
                *internalFormat = GR_GL_RGBA4;
            } else {
                *internalFormat = GR_GL_RGBA;
            }
            *externalType = GR_GL_UNSIGNED_SHORT_4_4_4_4;
            break;
        case kIndex_8_GrPixelConfig:
            // no sized/unsized internal format distinction here
            *internalFormat = GR_GL_PALETTE8_RGBA8;
            break;
        case kAlpha_8_GrPixelConfig:
            if (this->glCaps().textureRedSupport()) {
                *internalFormat = GR_GL_RED;
                *externalFormat = GR_GL_RED;
                if (getSizedInternalFormat) {
                    *internalFormat = GR_GL_R8;
                } else {
                    *internalFormat = GR_GL_RED;
                }
                *externalType = GR_GL_UNSIGNED_BYTE;
            } else {
                *internalFormat = GR_GL_ALPHA;
                *externalFormat = GR_GL_ALPHA;
                if (getSizedInternalFormat) {
                    *internalFormat = GR_GL_ALPHA8;
                } else {
                    *internalFormat = GR_GL_ALPHA;
                }
                *externalType = GR_GL_UNSIGNED_BYTE;
            }
            break;
        case kETC1_GrPixelConfig:
            *internalFormat = GR_GL_COMPRESSED_RGB8_ETC1;
            break;
        case kLATC_GrPixelConfig:
            switch(this->glCaps().latcAlias()) {
                case GrGLCaps::kLATC_LATCAlias:
                    *internalFormat = GR_GL_COMPRESSED_LUMINANCE_LATC1;
                    break;
                case GrGLCaps::kRGTC_LATCAlias:
                    *internalFormat = GR_GL_COMPRESSED_RED_RGTC1;
                    break;
                case GrGLCaps::k3DC_LATCAlias:
                    *internalFormat = GR_GL_COMPRESSED_3DC_X;
                    break;
            }
            break;
        case kR11_EAC_GrPixelConfig:
            *internalFormat = GR_GL_COMPRESSED_R11;
            break;

        case kASTC_12x12_GrPixelConfig:
            *internalFormat = GR_GL_COMPRESSED_RGBA_ASTC_12x12;
            break;

        case kRGBA_float_GrPixelConfig:
            *internalFormat = GR_GL_RGBA32F;
            *externalFormat = GR_GL_RGBA;
            *externalType = GR_GL_FLOAT;
            break;

        case kAlpha_half_GrPixelConfig:
            if (this->glCaps().textureRedSupport()) {
                if (getSizedInternalFormat) {
                    *internalFormat = GR_GL_R16F;
                } else {
                    *internalFormat = GR_GL_RED;
                }
                *externalFormat = GR_GL_RED;
            } else {
                if (getSizedInternalFormat) {
                    *internalFormat = GR_GL_ALPHA16F;
                } else {
                    *internalFormat = GR_GL_ALPHA;
                }
                *externalFormat = GR_GL_ALPHA;
            }
            if (kGL_GrGLStandard == this->glStandard() || this->glVersion() >= GR_GL_VER(3, 0)) {
                *externalType = GR_GL_HALF_FLOAT;
            } else {
                *externalType = GR_GL_HALF_FLOAT_OES;
            }
            break;
            
        default:
            return false;
    }
    return true;
}

void GrGLGpu::setTextureUnit(int unit) {
    SkASSERT(unit >= 0 && unit < fHWBoundTextureUniqueIDs.count());
    if (unit != fHWActiveTextureUnitIdx) {
        GL_CALL(ActiveTexture(GR_GL_TEXTURE0 + unit));
        fHWActiveTextureUnitIdx = unit;
    }
}

void GrGLGpu::setScratchTextureUnit() {
    // Bind the last texture unit since it is the least likely to be used by GrGLProgram.
    int lastUnitIdx = fHWBoundTextureUniqueIDs.count() - 1;
    if (lastUnitIdx != fHWActiveTextureUnitIdx) {
        GL_CALL(ActiveTexture(GR_GL_TEXTURE0 + lastUnitIdx));
        fHWActiveTextureUnitIdx = lastUnitIdx;
    }
    // clear out the this field so that if a program does use this unit it will rebind the correct
    // texture.
    fHWBoundTextureUniqueIDs[lastUnitIdx] = SK_InvalidUniqueID;
}

namespace {
// Determines whether glBlitFramebuffer could be used between src and dst.
inline bool can_blit_framebuffer(const GrSurface* dst,
                                 const GrSurface* src,
                                 const GrGLGpu* gpu) {
    if (gpu->glCaps().isConfigRenderable(dst->config(), dst->desc().fSampleCnt > 0) &&
        gpu->glCaps().isConfigRenderable(src->config(), src->desc().fSampleCnt > 0) &&
        gpu->glCaps().usesMSAARenderBuffers()) {
        // ES3 doesn't allow framebuffer blits when the src has MSAA and the configs don't match
        // or the rects are not the same (not just the same size but have the same edges).
        if (GrGLCaps::kES_3_0_MSFBOType == gpu->glCaps().msFBOType() &&
            (src->desc().fSampleCnt > 0 || src->config() != dst->config())) {
           return false;
        }
        return true;
    } else {
        return false;
    }
}

inline bool can_copy_texsubimage(const GrSurface* dst,
                                 const GrSurface* src,
                                 const GrGLGpu* gpu) {
    // Table 3.9 of the ES2 spec indicates the supported formats with CopyTexSubImage
    // and BGRA isn't in the spec. There doesn't appear to be any extension that adds it. Perhaps
    // many drivers would allow it to work, but ANGLE does not.
    if (kGLES_GrGLStandard == gpu->glStandard() && gpu->glCaps().bgraIsInternalFormat() &&
        (kBGRA_8888_GrPixelConfig == dst->config() || kBGRA_8888_GrPixelConfig == src->config())) {
        return false;
    }
    const GrGLRenderTarget* dstRT = static_cast<const GrGLRenderTarget*>(dst->asRenderTarget());
    // If dst is multisampled (and uses an extension where there is a separate MSAA renderbuffer)
    // then we don't want to copy to the texture but to the MSAA buffer.
    if (dstRT && dstRT->renderFBO() != dstRT->textureFBO()) {
        return false;
    }
    const GrGLRenderTarget* srcRT = static_cast<const GrGLRenderTarget*>(src->asRenderTarget());
    // If the src is multisampled (and uses an extension where there is a separate MSAA
    // renderbuffer) then it is an invalid operation to call CopyTexSubImage
    if (srcRT && srcRT->renderFBO() != srcRT->textureFBO()) {
        return false;
    }
    if (gpu->glCaps().isConfigRenderable(src->config(), src->desc().fSampleCnt > 0) &&
        dst->asTexture() &&
        dst->origin() == src->origin() &&
        !GrPixelConfigIsCompressed(src->config())) {
        return true;
    } else {
        return false;
    }
}

}

// If a temporary FBO was created, its non-zero ID is returned. The viewport that the copy rect is
// relative to is output.
GrGLGpu::FBOBinding GrGLGpu::bindSurfaceAsFBOForCopy(GrSurface* surface, FBOBinding binding,
                                                     GrGLIRect* viewport) {
    GrGLRenderTarget* rt = static_cast<GrGLRenderTarget*>(surface->asRenderTarget());
    if (NULL == rt) {
        SkASSERT(surface->asTexture());
        GrGLuint texID = static_cast<GrGLTexture*>(surface->asTexture())->textureID();
        GrGLFBO* tempFBO;

        if (kBlitSrc_FBOBinding == binding || kCopyTexSrc_FBOBinding == binding) {
            if (!fTempSrcFBO) {
                fTempSrcFBO.reset(SkNEW_ARGS(GrGLFBO, (this->glInterface())));
                SkASSERT(fTempSrcFBO->isValid());
            }
            tempFBO = fTempSrcFBO;
        } else {
            SkASSERT(kBlitDst_FBOBinding == binding);
            if (!fTempDstFBO) {
                fTempDstFBO.reset(SkNEW_ARGS(GrGLFBO, (this->glInterface())));
                SkASSERT(fTempDstFBO->isValid());
            }
            tempFBO = fTempDstFBO;
        }

        GrGLenum target = this->bindFBO(binding, tempFBO);
        GR_GL_CALL(this->glInterface(), FramebufferTexture2D(target,
                                                             GR_GL_COLOR_ATTACHMENT0,
                                                             GR_GL_TEXTURE_2D,
                                                             texID,
                                                             0));
        viewport->fLeft = 0;
        viewport->fBottom = 0;
        viewport->fWidth = surface->width();
        viewport->fHeight = surface->height();
        return binding;
    } else {
        this->bindFBO(binding, rt->renderFBO());
        *viewport = rt->getViewport();
        return kInvalidFBOBinding;
    }
}

void GrGLGpu::unbindSurfaceAsFBOForCopy(FBOBinding binding) {
    if (kInvalidFBOBinding == binding) {
        return;
    }
    GrGLFBO* tempFBO = kBlitDst_FBOBinding == binding ? fTempSrcFBO : fTempDstFBO;
    GrGLenum target = this->bindFBO(binding, tempFBO);
    GR_GL_CALL(this->glInterface(), FramebufferTexture2D(target,
                                                         GR_GL_COLOR_ATTACHMENT0,
                                                         GR_GL_TEXTURE_2D,
                                                         0,
                                                         0));
}

bool GrGLGpu::initCopySurfaceDstDesc(const GrSurface* src, GrSurfaceDesc* desc) {
    // In here we look for opportunities to use CopyTexSubImage, or fbo blit. If neither are
    // possible and we return false to fallback to creating a render target dst for render-to-
    // texture. This code prefers CopyTexSubImage to fbo blit and avoids triggering temporary fbo
    // creation. It isn't clear that avoiding temporary fbo creation is actually optimal.

    // Check for format issues with glCopyTexSubImage2D
    if (kGLES_GrGLStandard == this->glStandard() && this->glCaps().bgraIsInternalFormat() &&
        kBGRA_8888_GrPixelConfig == src->config()) {
        // glCopyTexSubImage2D doesn't work with this config. If the bgra can be used with fbo blit
        // then we set up for that, otherwise fail.
        if (this->caps()->isConfigRenderable(kBGRA_8888_GrPixelConfig, false)) {
            desc->fOrigin = kDefault_GrSurfaceOrigin;
            desc->fFlags = kRenderTarget_GrSurfaceFlag;
            desc->fConfig = kBGRA_8888_GrPixelConfig;
            return true;
        }
        return false;
    } else if (NULL == src->asRenderTarget()) {
        // CopyTexSubImage2D or fbo blit would require creating a temp fbo for the src.
        return false;
    }

    const GrGLRenderTarget* srcRT = static_cast<const GrGLRenderTarget*>(src->asRenderTarget());
    if (srcRT && srcRT->renderFBO() != srcRT->textureFBO()) {
        // It's illegal to call CopyTexSubImage2D on a MSAA renderbuffer. Set up for FBO blit or
        // fail.
        if (this->caps()->isConfigRenderable(src->config(), false)) {
            desc->fOrigin = kDefault_GrSurfaceOrigin;
            desc->fFlags = kRenderTarget_GrSurfaceFlag;
            desc->fConfig = src->config();
            return true;
        }
        return false;
    }

    // We'll do a CopyTexSubImage. Make the dst a plain old texture.
    desc->fConfig = src->config();
    desc->fOrigin = src->origin();
    desc->fFlags = kNone_GrSurfaceFlags;
    return true;
}

bool GrGLGpu::copySurface(GrSurface* dst,
                          GrSurface* src,
                          const SkIRect& srcRect,
                          const SkIPoint& dstPoint) {
    bool copied = false;
    SkIRect dstRect = SkIRect::MakeXYWH(dstPoint.fX, dstPoint.fY,
                                        srcRect.width(), srcRect.height());
    if (can_copy_texsubimage(dst, src, this)) {
        GrGLIRect srcVP;
        FBOBinding srcFBOBinding = this->bindSurfaceAsFBOForCopy(src, kCopyTexSrc_FBOBinding,
                                                                 &srcVP);
        GrGLTexture* dstTex = static_cast<GrGLTexture*>(dst->asTexture());
        SkASSERT(dstTex);
        GrGLIRect srcGLRect;
        srcGLRect.setRelativeTo(srcVP,
                                srcRect.fLeft,
                                srcRect.fTop,
                                srcRect.width(),
                                srcRect.height(),
                                src->origin());

        this->setScratchTextureUnit();
        GL_CALL(BindTexture(GR_GL_TEXTURE_2D, dstTex->textureID()));
        GrGLint dstY;
        if (kBottomLeft_GrSurfaceOrigin == dst->origin()) {
            dstY = dst->height() - (dstPoint.fY + srcGLRect.fHeight);
        } else {
            dstY = dstPoint.fY;
        }
        GL_CALL(CopyTexSubImage2D(GR_GL_TEXTURE_2D, 0,
                                  dstPoint.fX, dstY,
                                  srcGLRect.fLeft, srcGLRect.fBottom,
                                  srcGLRect.fWidth, srcGLRect.fHeight));
        copied = true;
        this->unbindSurfaceAsFBOForCopy(srcFBOBinding);
    } else if (can_blit_framebuffer(dst, src, this)) {
        bool selfOverlap = false;
        if (dst == src) {
            selfOverlap = SkIRect::IntersectsNoEmptyCheck(dstRect, srcRect);
        }

        if (!selfOverlap) {
            GrGLIRect dstVP;
            GrGLIRect srcVP;
            FBOBinding dstFBOBinding = this->bindSurfaceAsFBOForCopy(dst, kBlitDst_FBOBinding,
                                                                     &dstVP);
            FBOBinding srcFBOBinding = this->bindSurfaceAsFBOForCopy(src, kBlitSrc_FBOBinding,
                                                                     &srcVP);

            GrGLIRect srcGLRect;
            GrGLIRect dstGLRect;
            srcGLRect.setRelativeTo(srcVP,
                                    srcRect.fLeft,
                                    srcRect.fTop,
                                    srcRect.width(),
                                    srcRect.height(),
                                    src->origin());
            dstGLRect.setRelativeTo(dstVP,
                                    dstRect.fLeft,
                                    dstRect.fTop,
                                    dstRect.width(),
                                    dstRect.height(),
                                    dst->origin());

            // BlitFrameBuffer respects the scissor, so disable it.
            this->disableScissor();

            GrGLint srcY0;
            GrGLint srcY1;
            // Does the blit need to y-mirror or not?
            if (src->origin() == dst->origin()) {
                srcY0 = srcGLRect.fBottom;
                srcY1 = srcGLRect.fBottom + srcGLRect.fHeight;
            } else {
                srcY0 = srcGLRect.fBottom + srcGLRect.fHeight;
                srcY1 = srcGLRect.fBottom;
            }
            GL_CALL(BlitFramebuffer(srcGLRect.fLeft,
                                    srcY0,
                                    srcGLRect.fLeft + srcGLRect.fWidth,
                                    srcY1,
                                    dstGLRect.fLeft,
                                    dstGLRect.fBottom,
                                    dstGLRect.fLeft + dstGLRect.fWidth,
                                    dstGLRect.fBottom + dstGLRect.fHeight,
                                    GR_GL_COLOR_BUFFER_BIT, GR_GL_NEAREST));
            this->unbindSurfaceAsFBOForCopy(dstFBOBinding);
            this->unbindSurfaceAsFBOForCopy(srcFBOBinding);
            copied = true;
        }
    }
    if (copied) {
        this->markSurfaceContentsDirty(dst, &dstRect);
    }
    return copied;
}

bool GrGLGpu::canCopySurface(const GrSurface* dst,
                             const GrSurface* src,
                             const SkIRect& srcRect,
                             const SkIPoint& dstPoint) {
    // This mirrors the logic in onCopySurface.
    if (can_copy_texsubimage(dst, src, this)) {
        return true;
    }
    if (can_blit_framebuffer(dst, src, this)) {
        if (dst == src) {
            SkIRect dstRect = SkIRect::MakeXYWH(dstPoint.fX, dstPoint.fY,
                                                srcRect.width(), srcRect.height());
            if(!SkIRect::IntersectsNoEmptyCheck(dstRect, srcRect)) {
                return true;
            }
        } else {
            return true;
        }
    }
    return false;
}

void GrGLGpu::didAddGpuTraceMarker() {
    if (this->caps()->gpuTracingSupport()) {
        const GrTraceMarkerSet& markerArray = this->getActiveTraceMarkers();
        SkString markerString = markerArray.toStringLast();
#if GR_FORCE_GPU_TRACE_DEBUGGING
        SkDebugf("%s\n", markerString.c_str());
#else
        GL_CALL(PushGroupMarker(0, markerString.c_str()));
#endif
    }
}

void GrGLGpu::didRemoveGpuTraceMarker() {
    if (this->caps()->gpuTracingSupport()) {
#if GR_FORCE_GPU_TRACE_DEBUGGING
        SkDebugf("Pop trace marker.\n");
#else
        GL_CALL(PopGroupMarker());
#endif
    }
}

///////////////////////////////////////////////////////////////////////////////

GrGLAttribArrayState* GrGLGpu::HWGeometryState::bindArrayAndBuffersToDraw(
                                                GrGLGpu* gpu,
                                                const GrGLVertexBuffer* vbuffer,
                                                const GrGLIndexBuffer* ibuffer) {
    SkASSERT(vbuffer);
    GrGLAttribArrayState* attribState;

    // We use a vertex array if we're on a core profile and the verts are in a VBO.
    if (gpu->glCaps().isCoreProfile() && !vbuffer->isCPUBacked()) {
        if (NULL == fVBOVertexArray || fVBOVertexArray->wasDestroyed()) {
            SkSafeUnref(fVBOVertexArray);
            GrGLuint arrayID;
            GR_GL_CALL(gpu->glInterface(), GenVertexArrays(1, &arrayID));
            int attrCount = gpu->glCaps().maxVertexAttributes();
            fVBOVertexArray = SkNEW_ARGS(GrGLVertexArray, (gpu, arrayID, attrCount));
        }
        attribState = fVBOVertexArray->bindWithIndexBuffer(ibuffer);
    } else {
        if (ibuffer) {
            this->setIndexBufferIDOnDefaultVertexArray(gpu, ibuffer->bufferID());
        } else {
            this->setVertexArrayID(gpu, 0);
        }
        int attrCount = gpu->glCaps().maxVertexAttributes();
        if (fDefaultVertexArrayAttribState.count() != attrCount) {
            fDefaultVertexArrayAttribState.resize(attrCount);
        }
        attribState = &fDefaultVertexArrayAttribState;
    }
    return attribState;
}