src/gpu/GrGpu.h

/*
 * Copyright 2011 Google Inc.
 *
 * Use of this source code is governed by a BSD-style license that can be
 * found in the LICENSE file.
 */

#ifndef GrGpu_DEFINED
#define GrGpu_DEFINED

#include "GrDrawTarget.h"
#include "GrPathRendering.h"
#include "GrProgramDesc.h"
#include "SkPath.h"

class GrContext;
class GrNonInstancedVertices;
class GrPath;
class GrPathRange;
class GrPathRenderer;
class GrPathRendererChain;
class GrPipeline;
class GrPrimitiveProcessor;
class GrStencilAttachment;
class GrVertices;

class GrGpu : public SkRefCnt {
public:
    /**
     * Create an instance of GrGpu that matches the specified backend. If the requested backend is
     * not supported (at compile-time or run-time) this returns NULL. The context will not be
     * fully constructed and should not be used by GrGpu until after this function returns.
     */
    static GrGpu* Create(GrBackend, GrBackendContext, GrContext* context);

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

    GrGpu(GrContext* context);
    ~GrGpu() override;

    GrContext* getContext() { return fContext; }
    const GrContext* getContext() const { return fContext; }

    /**
     * Gets the capabilities of the draw target.
     */
    const GrCaps* caps() const { return fCaps.get(); }

    GrPathRendering* pathRendering() { return fPathRendering.get(); }

    // Called by GrContext when the underlying backend context has been destroyed.
    // GrGpu should use this to ensure that no backend API calls will be made from
    // here onward, including in its destructor. Subclasses should call
    // INHERITED::contextAbandoned() if they override this.
    virtual void contextAbandoned();

    /**
     * The GrGpu object normally assumes that no outsider is setting state
     * within the underlying 3D API's context/device/whatever. This call informs
     * the GrGpu that the state was modified and it shouldn't make assumptions
     * about the state.
     */
    void markContextDirty(uint32_t state = kAll_GrBackendState) { fResetBits |= state; }

    /**
     * Creates a texture object. If kRenderTarget_GrSurfaceFlag the texture can
     * be used as a render target by calling GrTexture::asRenderTarget(). Not all
     * pixel configs can be used as render targets. Support for configs as textures
     * or render targets can be checked using GrCaps.
     *
     * @param desc        describes the texture to be created.
     * @param budgeted    does this texture count against the resource cache budget?
     * @param srcData     texel data to load texture. Begins with full-size
     *                    palette data for paletted textures. For compressed
     *                    formats it contains the compressed pixel data. Otherwise,
     *                    it contains width*height texels. If NULL texture data
     *                    is uninitialized.
     * @param rowBytes    the number of bytes between consecutive rows. Zero
     *                    means rows are tightly packed. This field is ignored
     *                    for compressed formats.
     *
     * @return    The texture object if successful, otherwise NULL.
     */
    GrTexture* createTexture(const GrSurfaceDesc& desc, bool budgeted,
                             const void* srcData, size_t rowBytes);

    /**
     * Implements GrContext::wrapBackendTexture
     */
    GrTexture* wrapBackendTexture(const GrBackendTextureDesc&);

    /**
     * Implements GrContext::wrapBackendTexture
     */
    GrRenderTarget* wrapBackendRenderTarget(const GrBackendRenderTargetDesc&);

    /**
     * Creates a vertex buffer.
     *
     * @param size    size in bytes of the vertex buffer
     * @param dynamic hints whether the data will be frequently changed
     *                by either GrVertexBuffer::map() or
     *                GrVertexBuffer::updateData().
     *
     * @return    The vertex buffer if successful, otherwise NULL.
     */
    GrVertexBuffer* createVertexBuffer(size_t size, bool dynamic);

    /**
     * Creates an index buffer.
     *
     * @param size    size in bytes of the index buffer
     * @param dynamic hints whether the data will be frequently changed
     *                by either GrIndexBuffer::map() or
     *                GrIndexBuffer::updateData().
     *
     * @return The index buffer if successful, otherwise NULL.
     */
    GrIndexBuffer* createIndexBuffer(size_t size, bool dynamic);

    /**
     * Resolves MSAA.
     */
    void resolveRenderTarget(GrRenderTarget* target);

    /**
     * Gets a preferred 8888 config to use for writing/reading pixel data to/from a surface with
     * config surfaceConfig. The returned config must have at least as many bits per channel as the
     * readConfig or writeConfig param.
     */
    virtual GrPixelConfig preferredReadPixelsConfig(GrPixelConfig readConfig,
                                                    GrPixelConfig surfaceConfig) const {
        return readConfig;
    }
    virtual GrPixelConfig preferredWritePixelsConfig(GrPixelConfig writeConfig,
                                                     GrPixelConfig surfaceConfig) const {
        return writeConfig;
    }

    /**
     * Called before uploading writing pixels to a GrTexture when the src pixel config doesn't
     * match the texture's config.
     */
    virtual bool canWriteTexturePixels(const GrTexture*, GrPixelConfig srcConfig) const = 0;

    /**
     * OpenGL's readPixels returns the result bottom-to-top while the skia
     * API is top-to-bottom. Thus we have to do a y-axis flip. The obvious
     * solution is to have the subclass do the flip using either the CPU or GPU.
     * However, the caller (GrContext) may have transformations to apply and can
     * simply fold in the y-flip for free. On the other hand, the subclass may
     * be able to do it for free itself. For example, the subclass may have to
     * do memcpys to handle rowBytes that aren't tight. It could do the y-flip
     * concurrently.
     *
     * This function returns true if a y-flip is required to put the pixels in
     * top-to-bottom order and the subclass cannot do it for free.
     *
     * See read pixels for the params
     * @return true if calling readPixels with the same set of params will
     *              produce bottom-to-top data
     */
     virtual bool readPixelsWillPayForYFlip(GrRenderTarget* renderTarget,
                                            int left, int top,
                                            int width, int height,
                                            GrPixelConfig config,
                                            size_t rowBytes) const = 0;
     /**
      * This should return true if reading a NxM rectangle of pixels from a
      * render target is faster if the target has dimensons N and M and the read
      * rectangle has its top-left at 0,0.
      */
     virtual bool fullReadPixelsIsFasterThanPartial() const { return false; };

    /**
     * Reads a rectangle of pixels from a render target.
     *
     * @param renderTarget  the render target to read from. NULL means the
     *                      current render target.
     * @param left          left edge of the rectangle to read (inclusive)
     * @param top           top edge of the rectangle to read (inclusive)
     * @param width         width of rectangle to read in pixels.
     * @param height        height of rectangle to read in pixels.
     * @param config        the pixel config of the destination buffer
     * @param buffer        memory to read the rectangle into.
     * @param rowBytes      the number of bytes between consecutive rows. Zero
     *                      means rows are tightly packed.
     * @param invertY       buffer should be populated bottom-to-top as opposed
     *                      to top-to-bottom (skia's usual order)
     *
     * @return true if the read succeeded, false if not. The read can fail
     *              because of a unsupported pixel config or because no render
     *              target is currently set.
     */
    bool readPixels(GrRenderTarget* renderTarget,
                    int left, int top, int width, int height,
                    GrPixelConfig config, void* buffer, size_t rowBytes);

    /**
     * Updates the pixels in a rectangle of a texture.
     *
     * @param left          left edge of the rectangle to write (inclusive)
     * @param top           top edge of the rectangle to write (inclusive)
     * @param width         width of rectangle to write in pixels.
     * @param height        height of rectangle to write in pixels.
     * @param config        the pixel config of the source buffer
     * @param buffer        memory to read pixels from
     * @param rowBytes      number of bytes between consecutive rows. Zero
     *                      means rows are tightly packed.
     */
    bool writeTexturePixels(GrTexture* texture,
                            int left, int top, int width, int height,
                            GrPixelConfig config, const void* buffer,
                            size_t rowBytes);

    /**
     * Clear the passed in render target. Ignores the draw state and clip. Clears the whole thing if
     * rect is NULL, otherwise just the rect. If canIgnoreRect is set then the entire render target
     * can be optionally cleared.
     */
    void clear(const SkIRect* rect, GrColor color, bool canIgnoreRect,GrRenderTarget* renderTarget);


    void clearStencilClip(const SkIRect& rect, bool insideClip, GrRenderTarget* renderTarget);

    /**
     * Discards the contents render target. NULL indicates that the current render target should
     * be discarded.
     **/
    virtual void discard(GrRenderTarget* = NULL) = 0;

    /**
     * This is can be called before allocating a texture to be a dst for copySurface. It will
     * populate the origin, config, and flags fields of the desc such that copySurface can
     * efficiently succeed. It should only succeed if it can allow copySurface to perform a copy
     * that would be more effecient than drawing the src to a dst render target.
     */
    virtual bool initCopySurfaceDstDesc(const GrSurface* src, GrSurfaceDesc* desc) = 0;

    // After the client interacts directly with the 3D context state the GrGpu
    // must resync its internal state and assumptions about 3D context state.
    // Each time this occurs the GrGpu bumps a timestamp.
    // state of the 3D context
    // At 10 resets / frame and 60fps a 64bit timestamp will overflow in about
    // a billion years.
    typedef uint64_t ResetTimestamp;

    // This timestamp is always older than the current timestamp
    static const ResetTimestamp kExpiredTimestamp = 0;
    // Returns a timestamp based on the number of times the context was reset.
    // This timestamp can be used to lazily detect when cached 3D context state
    // is dirty.
    ResetTimestamp getResetTimestamp() const { return fResetTimestamp; }

    virtual void buildProgramDesc(GrProgramDesc*,
                                  const GrPrimitiveProcessor&,
                                  const GrPipeline&,
                                  const GrBatchTracker&) const = 0;

    // Called to determine whether a copySurface call would succeed or not. Derived
    // classes must keep this consistent with their implementation of onCopySurface(). Fallbacks
    // to issuing a draw from the src to dst take place at the GrDrawTarget level and this function
    // should only return true if a faster copy path exists. The rect and point are pre-clipped. The
    // src rect and implied dst rect are guaranteed to be within the src/dst bounds and non-empty.
    virtual bool canCopySurface(const GrSurface* dst,
                                const GrSurface* src,
                                const SkIRect& srcRect,
                                const SkIPoint& dstPoint) = 0;

    // Called to perform a surface to surface copy. Fallbacks to issuing a draw from the src to dst
    // take place at the GrDrawTarget level and this function implement faster copy paths. The rect
    // and point are pre-clipped. The src rect and implied dst rect are guaranteed to be within the
    // src/dst bounds and non-empty.
    virtual bool copySurface(GrSurface* dst,
                             GrSurface* src,
                             const SkIRect& srcRect,
                             const SkIPoint& dstPoint) = 0;

    // Called before certain draws in order to guarantee coherent results from dst reads.
    virtual void xferBarrier(GrRenderTarget*, GrXferBarrierType) = 0;

    struct DrawArgs {
        DrawArgs(const GrPrimitiveProcessor* primProc,
                 const GrPipeline* pipeline,
                 const GrProgramDesc* desc,
                 const GrBatchTracker* batchTracker)
            : fPrimitiveProcessor(primProc)
            , fPipeline(pipeline)
            , fDesc(desc)
            , fBatchTracker(batchTracker) {
            SkASSERT(primProc && pipeline && desc && batchTracker);
        }
        const GrPrimitiveProcessor* fPrimitiveProcessor;
        const GrPipeline* fPipeline;
        const GrProgramDesc* fDesc;
        const GrBatchTracker* fBatchTracker;
    };

    void draw(const DrawArgs&, const GrVertices&);

    /** None of these params are optional, pointers used just to avoid making copies. */
    struct StencilPathState {
        bool fUseHWAA;
        GrRenderTarget* fRenderTarget;
        const SkMatrix* fViewMatrix;
        const GrStencilSettings* fStencil;
        const GrScissorState* fScissor;
    };

    void stencilPath(const GrPath*, const StencilPathState&);

    void drawPath(const DrawArgs&, const GrPath*, const GrStencilSettings&);
    void drawPaths(const DrawArgs&,
                   const GrPathRange*,
                   const void* indices,
                   GrDrawTarget::PathIndexType,
                   const float transformValues[],
                   GrDrawTarget::PathTransformType,
                   int count,
                   const GrStencilSettings&);

    ///////////////////////////////////////////////////////////////////////////
    // Debugging and Stats

    class Stats {
    public:
#if GR_GPU_STATS
        Stats() { this->reset(); }

        void reset() {
            fRenderTargetBinds = 0;
            fShaderCompilations = 0;
            fTextureCreates = 0;
            fTextureUploads = 0;
            fStencilAttachmentCreates = 0;
        }

        int renderTargetBinds() const { return fRenderTargetBinds; }
        void incRenderTargetBinds() { fRenderTargetBinds++; }
        int shaderCompilations() const { return fShaderCompilations; }
        void incShaderCompilations() { fShaderCompilations++; }
        int textureCreates() const { return fTextureCreates; }
        void incTextureCreates() { fTextureCreates++; }
        int textureUploads() const { return fTextureUploads; }
        void incTextureUploads() { fTextureUploads++; }
        void incStencilAttachmentCreates() { fStencilAttachmentCreates++; }
        void dump(SkString*);

    private:
        int fRenderTargetBinds;
        int fShaderCompilations;
        int fTextureCreates;
        int fTextureUploads;
        int fStencilAttachmentCreates;
#else
        void dump(SkString*) {};
        void incRenderTargetBinds() {}
        void incShaderCompilations() {}
        void incTextureCreates() {}
        void incTextureUploads() {}
        void incStencilAttachmentCreates() {}
#endif
    };

    Stats* stats() { return &fStats; }

    /**
     * Called at start and end of gpu trace marking
     * GR_CREATE_GPU_TRACE_MARKER(marker_str, target) will automatically call these at the start
     * and end of a code block respectively
     */
    void addGpuTraceMarker(const GrGpuTraceMarker* marker);
    void removeGpuTraceMarker(const GrGpuTraceMarker* marker);

    /**
     * Takes the current active set of markers and stores them for later use. Any current marker
     * in the active set is removed from the active set and the targets remove function is called.
     * These functions do not work as a stack so you cannot call save a second time before calling
     * restore. Also, it is assumed that when restore is called the current active set of markers
     * is empty. When the stored markers are added back into the active set, the targets add marker
     * is called.
     */
    void saveActiveTraceMarkers();
    void restoreActiveTraceMarkers();

    // Given a rt, find or create a stencil buffer and attach it
    bool attachStencilAttachmentToRenderTarget(GrRenderTarget* target);

protected:
    // Functions used to map clip-respecting stencil tests into normal
    // stencil funcs supported by GPUs.
    static GrStencilFunc ConvertStencilFunc(bool stencilInClip,
                                            GrStencilFunc func);
    static void ConvertStencilFuncAndMask(GrStencilFunc func,
                                          bool clipInStencil,
                                          unsigned int clipBit,
                                          unsigned int userBits,
                                          unsigned int* ref,
                                          unsigned int* mask);

    const GrTraceMarkerSet& getActiveTraceMarkers() const { return fActiveTraceMarkers; }

    Stats                                   fStats;
    SkAutoTDelete<GrPathRendering>          fPathRendering;
    // Subclass must initialize this in its constructor.
    SkAutoTUnref<const GrCaps>    fCaps;

private:
    // called when the 3D context state is unknown. Subclass should emit any
    // assumed 3D context state and dirty any state cache.
    virtual void onResetContext(uint32_t resetBits) = 0;

    // overridden by backend-specific derived class to create objects.
    // Texture size and sample size will have already been validated in base class before
    // onCreateTexture/CompressedTexture are called.
    virtual GrTexture* onCreateTexture(const GrSurfaceDesc& desc,
                                       GrGpuResource::LifeCycle lifeCycle,
                                       const void* srcData, size_t rowBytes) = 0;
    virtual GrTexture* onCreateCompressedTexture(const GrSurfaceDesc& desc,
                                                 GrGpuResource::LifeCycle lifeCycle,
                                                 const void* srcData) = 0;
    virtual GrTexture* onWrapBackendTexture(const GrBackendTextureDesc&) = 0;
    virtual GrRenderTarget* onWrapBackendRenderTarget(const GrBackendRenderTargetDesc&) = 0;
    virtual GrVertexBuffer* onCreateVertexBuffer(size_t size, bool dynamic) = 0;
    virtual GrIndexBuffer* onCreateIndexBuffer(size_t size, bool dynamic) = 0;

    // overridden by backend-specific derived class to perform the clear.
    virtual void onClear(GrRenderTarget*, const SkIRect* rect, GrColor color,
                         bool canIgnoreRect) = 0;


    // Overridden by backend specific classes to perform a clear of the stencil clip bits.  This is
    // ONLY used by the the clip target
    virtual void onClearStencilClip(GrRenderTarget*, const SkIRect& rect, bool insideClip) = 0;

    // overridden by backend-specific derived class to perform the draw call.
    virtual void onDraw(const DrawArgs&, const GrNonInstancedVertices&) = 0;
    virtual void onStencilPath(const GrPath*, const StencilPathState&) = 0;

    virtual void onDrawPath(const DrawArgs&, const GrPath*, const GrStencilSettings&) = 0;
    virtual void onDrawPaths(const DrawArgs&,
                             const GrPathRange*,
                             const void* indices,
                             GrDrawTarget::PathIndexType,
                             const float transformValues[],
                             GrDrawTarget::PathTransformType,
                             int count,
                             const GrStencilSettings&) = 0;

    // overridden by backend-specific derived class to perform the read pixels.
    virtual bool onReadPixels(GrRenderTarget* target,
                              int left, int top, int width, int height,
                              GrPixelConfig,
                              void* buffer,
                              size_t rowBytes) = 0;

    // overridden by backend-specific derived class to perform the texture update
    virtual bool onWriteTexturePixels(GrTexture* texture,
                                      int left, int top, int width, int height,
                                      GrPixelConfig config, const void* buffer,
                                      size_t rowBytes) = 0;

    // overridden by backend-specific derived class to perform the resolve
    virtual void onResolveRenderTarget(GrRenderTarget* target) = 0;

    // width and height may be larger than rt (if underlying API allows it).
    // Should attach the SB to the RT. Returns false if compatible sb could
    // not be created.
    virtual bool createStencilAttachmentForRenderTarget(GrRenderTarget*, int width, int height) = 0;

    // attaches an existing SB to an existing RT.
    virtual bool attachStencilAttachmentToRenderTarget(GrStencilAttachment*, GrRenderTarget*) = 0;

    // clears target's entire stencil buffer to 0
    virtual void clearStencil(GrRenderTarget* target) = 0;

    virtual void didAddGpuTraceMarker() = 0;
    virtual void didRemoveGpuTraceMarker() = 0;

    void resetContext() {
        this->onResetContext(fResetBits);
        fResetBits = 0;
        ++fResetTimestamp;
    }

    void handleDirtyContext() {
        if (fResetBits) {
            this->resetContext();
        }
    }

    ResetTimestamp                                                      fResetTimestamp;
    uint32_t                                                            fResetBits;
    // To keep track that we always have at least as many debug marker adds as removes
    int                                                                 fGpuTraceMarkerCount;
    GrTraceMarkerSet                                                    fActiveTraceMarkers;
    GrTraceMarkerSet                                                    fStoredTraceMarkers;
    // The context owns us, not vice-versa, so this ptr is not ref'ed by Gpu.
    GrContext*                                                          fContext;

    typedef SkRefCnt INHERITED;
};

#endif