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

529 lines
19 KiB

/*
* Copyright 2016 Google Inc.
*
* Use of this source code is governed by a BSD-style license that can be
* found in the LICENSE file.
*/
#ifndef GrSurfaceProxy_DEFINED
#define GrSurfaceProxy_DEFINED
#include "../private/SkNoncopyable.h"
#include "GrBackendSurface.h"
#include "GrGpuResource.h"
#include "GrSurface.h"
#include "SkRect.h"
class GrCaps;
class GrOpList;
class GrProxyProvider;
class GrRenderTargetOpList;
class GrRenderTargetProxy;
class GrResourceProvider;
class GrSurfaceContext;
class GrSurfaceProxyPriv;
class GrTextureOpList;
class GrTextureProxy;
// This class replicates the functionality GrIORef<GrSurface> but tracks the
// utilitization for later resource allocation (for the deferred case) and
// forwards on the utilization in the wrapped case
class GrIORefProxy : public SkNoncopyable {
public:
void ref() const {
this->validate();
++fRefCnt;
if (fTarget) {
fTarget->ref();
}
}
void unref() const {
this->validate();
if (fTarget) {
fTarget->unref();
}
--fRefCnt;
this->didRemoveRefOrPendingIO();
}
#ifdef SK_DEBUG
bool isUnique_debugOnly() const { // For asserts.
SkASSERT(fRefCnt >= 0 && fPendingWrites >= 0 && fPendingReads >= 0);
return 1 == fRefCnt + fPendingWrites + fPendingReads;
}
#endif
void release() {
// The proxy itself may still have multiple refs. It can be owned by an SkImage and multiple
// SkDeferredDisplayLists at the same time if we are using DDLs.
SkASSERT(0 == fPendingReads);
SkASSERT(0 == fPendingWrites);
// In the current hybrid world, the proxy and backing surface are ref/unreffed in
// synchrony. Each ref we've added or removed to the proxy was mirrored to the backing
// surface. Though, that backing surface could be owned by other proxies as well. Remove
// a ref from the backing surface for each ref the proxy has since we are about to remove
// our pointer to the surface. If this proxy is reinstantiated then all the proxy's refs
// get transferred to the (possibly new) backing surface.
for (int refs = fRefCnt; refs; --refs) {
fTarget->unref();
}
fTarget = nullptr;
}
void validate() const {
#ifdef SK_DEBUG
SkASSERT(fRefCnt >= 0);
SkASSERT(fPendingReads >= 0);
SkASSERT(fPendingWrites >= 0);
SkASSERT(fRefCnt + fPendingReads + fPendingWrites >= 1);
if (fTarget) {
// The backing GrSurface can have more refs than the proxy if the proxy
// started off wrapping an external resource (that came in with refs).
// The GrSurface should never have fewer refs than the proxy however.
SkASSERT(fTarget->fRefCnt >= fRefCnt);
SkASSERT(fTarget->fPendingReads >= fPendingReads);
SkASSERT(fTarget->fPendingWrites >= fPendingWrites);
}
#endif
}
int32_t getBackingRefCnt_TestOnly() const;
int32_t getPendingReadCnt_TestOnly() const;
int32_t getPendingWriteCnt_TestOnly() const;
void addPendingRead() const {
this->validate();
++fPendingReads;
if (fTarget) {
fTarget->addPendingRead();
}
}
void completedRead() const {
this->validate();
if (fTarget) {
fTarget->completedRead();
}
--fPendingReads;
this->didRemoveRefOrPendingIO();
}
void addPendingWrite() const {
this->validate();
++fPendingWrites;
if (fTarget) {
fTarget->addPendingWrite();
}
}
void completedWrite() const {
this->validate();
if (fTarget) {
fTarget->completedWrite();
}
--fPendingWrites;
this->didRemoveRefOrPendingIO();
}
protected:
GrIORefProxy() : fTarget(nullptr), fRefCnt(1), fPendingReads(0), fPendingWrites(0) {}
GrIORefProxy(sk_sp<GrSurface> surface) : fRefCnt(1), fPendingReads(0), fPendingWrites(0) {
// Since we're manually forwarding on refs & unrefs we don't want sk_sp doing
// anything extra.
fTarget = surface.release();
}
virtual ~GrIORefProxy() {
// We don't unref 'fTarget' here since the 'unref' method will already
// have forwarded on the unref call that got us here.
}
// This GrIORefProxy was deferred before but has just been instantiated. To
// make all the reffing & unreffing work out we now need to transfer any deferred
// refs & unrefs to the new GrSurface
void transferRefs() {
SkASSERT(fTarget);
SkASSERT(fTarget->fRefCnt > 0);
fTarget->fRefCnt += (fRefCnt-1); // don't xfer the proxy's creation ref
fTarget->fPendingReads += fPendingReads;
fTarget->fPendingWrites += fPendingWrites;
}
int32_t internalGetProxyRefCnt() const {
return fRefCnt;
}
bool internalHasPendingIO() const {
if (fTarget) {
return fTarget->internalHasPendingIO();
}
return SkToBool(fPendingWrites | fPendingReads);
}
bool internalHasPendingWrite() const {
if (fTarget) {
return fTarget->internalHasPendingWrite();
}
return SkToBool(fPendingWrites);
}
// For deferred proxies this will be null. For wrapped proxies it will point to the
// wrapped resource.
GrSurface* fTarget;
private:
// This class is used to manage conversion of refs to pending reads/writes.
template <typename> friend class GrProxyRef;
void didRemoveRefOrPendingIO() const {
if (0 == fPendingReads && 0 == fPendingWrites && 0 == fRefCnt) {
delete this;
}
}
mutable int32_t fRefCnt;
mutable int32_t fPendingReads;
mutable int32_t fPendingWrites;
};
class GrSurfaceProxy : public GrIORefProxy {
public:
enum class LazyInstantiationType {
kSingleUse, // Instantiation callback is allowed to be called only once.
kMultipleUse, // Instantiation callback can be called multiple times.
kDeinstantiate, // Instantiation callback can be called multiple times,
// but we will deinstantiate the proxy after every flush.
};
enum class LazyState {
kNot, // The proxy is instantiated or does not have a lazy callback
kPartially, // The proxy has a lazy callback but knows basic information about itself.
kFully, // The proxy has a lazy callback and also doesn't know its width, height, etc.
};
LazyState lazyInstantiationState() const {
if (fTarget || !SkToBool(fLazyInstantiateCallback)) {
return LazyState::kNot;
} else {
if (fWidth <= 0) {
SkASSERT(fHeight <= 0);
return LazyState::kFully;
} else {
SkASSERT(fHeight > 0);
return LazyState::kPartially;
}
}
}
GrPixelConfig config() const { return fConfig; }
int width() const {
SkASSERT(LazyState::kFully != this->lazyInstantiationState());
return fWidth;
}
int height() const {
SkASSERT(LazyState::kFully != this->lazyInstantiationState());
return fHeight;
}
SkISize isize() const { return {fWidth, fHeight}; }
int worstCaseWidth() const;
int worstCaseHeight() const;
/**
* Helper that gets the width and height of the surface as a bounding rectangle.
*/
SkRect getBoundsRect() const {
SkASSERT(LazyState::kFully != this->lazyInstantiationState());
return SkRect::MakeIWH(this->width(), this->height());
}
/**
* Helper that gets the worst case width and height of the surface as a bounding rectangle.
*/
SkRect getWorstCaseBoundsRect() const {
SkASSERT(LazyState::kFully != this->lazyInstantiationState());
return SkRect::MakeIWH(this->worstCaseWidth(), this->worstCaseHeight());
}
GrSurfaceOrigin origin() const {
SkASSERT(kTopLeft_GrSurfaceOrigin == fOrigin || kBottomLeft_GrSurfaceOrigin == fOrigin);
return fOrigin;
}
const GrBackendFormat& backendFormat() const { return fFormat; }
class UniqueID {
public:
static UniqueID InvalidID() {
return UniqueID(uint32_t(SK_InvalidUniqueID));
}
// wrapped
explicit UniqueID(const GrGpuResource::UniqueID& id) : fID(id.asUInt()) { }
// deferred and lazy-callback
UniqueID() : fID(GrGpuResource::CreateUniqueID()) { }
uint32_t asUInt() const { return fID; }
bool operator==(const UniqueID& other) const {
return fID == other.fID;
}
bool operator!=(const UniqueID& other) const {
return !(*this == other);
}
void makeInvalid() { fID = SK_InvalidUniqueID; }
bool isInvalid() const { return SK_InvalidUniqueID == fID; }
private:
explicit UniqueID(uint32_t id) : fID(id) {}
uint32_t fID;
};
/*
* The contract for the uniqueID is:
* for wrapped resources:
* the uniqueID will match that of the wrapped resource
*
* for deferred resources:
* the uniqueID will be different from the real resource, when it is allocated
* the proxy's uniqueID will not change across the instantiate call
*
* the uniqueIDs of the proxies and the resources draw from the same pool
*
* What this boils down to is that the uniqueID of a proxy can be used to consistently
* track/identify a proxy but should never be used to distinguish between
* resources and proxies - beware!
*/
UniqueID uniqueID() const { return fUniqueID; }
UniqueID underlyingUniqueID() const {
if (fTarget) {
return UniqueID(fTarget->uniqueID());
}
return fUniqueID;
}
virtual bool instantiate(GrResourceProvider* resourceProvider) = 0;
void deinstantiate();
/**
* Proxies that are already instantiated and whose backing surface cannot be recycled to
* instantiate other proxies do not need to be considered by GrResourceAllocator.
*/
bool canSkipResourceAllocator() const;
/**
* @return the texture proxy associated with the surface proxy, may be NULL.
*/
virtual GrTextureProxy* asTextureProxy() { return nullptr; }
virtual const GrTextureProxy* asTextureProxy() const { return nullptr; }
/**
* @return the render target proxy associated with the surface proxy, may be NULL.
*/
virtual GrRenderTargetProxy* asRenderTargetProxy() { return nullptr; }
virtual const GrRenderTargetProxy* asRenderTargetProxy() const { return nullptr; }
bool isInstantiated() const { return SkToBool(fTarget); }
// If the proxy is already instantiated, return its backing GrTexture; if not, return null.
GrSurface* peekSurface() const { return fTarget; }
// If this is a texture proxy and the proxy is already instantiated, return its backing
// GrTexture; if not, return null.
GrTexture* peekTexture() const { return fTarget ? fTarget->asTexture() : nullptr; }
// If this is a render target proxy and the proxy is already instantiated, return its backing
// GrRenderTarget; if not, return null.
GrRenderTarget* peekRenderTarget() const {
return fTarget ? fTarget->asRenderTarget() : nullptr;
}
/**
* Does the resource count against the resource budget?
*/
SkBudgeted isBudgeted() const { return fBudgeted; }
/**
* The pixel values of this proxy's surface cannot be modified (e.g. doesn't support write
* pixels or MIP map level regen). Read-only proxies also bypass interval tracking and
* assignment in GrResourceAllocator.
*/
bool readOnly() const { return fSurfaceFlags & GrInternalSurfaceFlags::kReadOnly; }
void setLastOpList(GrOpList* opList);
GrOpList* getLastOpList() { return fLastOpList; }
GrRenderTargetOpList* getLastRenderTargetOpList();
GrTextureOpList* getLastTextureOpList();
/**
* Retrieves the amount of GPU memory that will be or currently is used by this resource
* in bytes. It is approximate since we aren't aware of additional padding or copies made
* by the driver.
*
* @return the amount of GPU memory used in bytes
*/
size_t gpuMemorySize() const {
SkASSERT(LazyState::kFully != this->lazyInstantiationState());
if (fTarget) {
return fTarget->gpuMemorySize();
}
if (kInvalidGpuMemorySize == fGpuMemorySize) {
fGpuMemorySize = this->onUninstantiatedGpuMemorySize();
SkASSERT(kInvalidGpuMemorySize != fGpuMemorySize);
}
return fGpuMemorySize;
}
// Helper function that creates a temporary SurfaceContext to perform the copy
// The copy is is not a render target and not multisampled.
static sk_sp<GrTextureProxy> Copy(GrContext*, GrSurfaceProxy* src, GrMipMapped, SkIRect srcRect,
SkBackingFit, SkBudgeted);
// Copy the entire 'src'
static sk_sp<GrTextureProxy> Copy(GrContext*, GrSurfaceProxy* src, GrMipMapped, SkBackingFit,
SkBudgeted budgeted);
// Test-only entry point - should decrease in use as proxies propagate
static sk_sp<GrSurfaceContext> TestCopy(GrContext* context, const GrSurfaceDesc& dstDesc,
GrSurfaceOrigin, GrSurfaceProxy* srcProxy);
bool isWrapped_ForTesting() const;
SkDEBUGCODE(void validate(GrContext*) const;)
// Provides access to functions that aren't part of the public API.
inline GrSurfaceProxyPriv priv();
inline const GrSurfaceProxyPriv priv() const;
GrInternalSurfaceFlags testingOnly_getFlags() const;
protected:
// Deferred version
GrSurfaceProxy(const GrBackendFormat& format, const GrSurfaceDesc& desc,
GrSurfaceOrigin origin, SkBackingFit fit,
SkBudgeted budgeted, GrInternalSurfaceFlags surfaceFlags)
: GrSurfaceProxy(nullptr, LazyInstantiationType::kSingleUse, format, desc, origin, fit,
budgeted, surfaceFlags) {
// Note: this ctor pulls a new uniqueID from the same pool at the GrGpuResources
}
using LazyInstantiateCallback = std::function<sk_sp<GrSurface>(GrResourceProvider*)>;
// Lazy-callback version
GrSurfaceProxy(LazyInstantiateCallback&&, LazyInstantiationType,
const GrBackendFormat& format, const GrSurfaceDesc&, GrSurfaceOrigin,
SkBackingFit, SkBudgeted, GrInternalSurfaceFlags);
// Wrapped version.
GrSurfaceProxy(sk_sp<GrSurface>, GrSurfaceOrigin, SkBackingFit);
virtual ~GrSurfaceProxy();
friend class GrSurfaceProxyPriv;
// Methods made available via GrSurfaceProxyPriv
int32_t getProxyRefCnt() const {
return this->internalGetProxyRefCnt();
}
bool hasPendingIO() const {
return this->internalHasPendingIO();
}
bool hasPendingWrite() const {
return this->internalHasPendingWrite();
}
void computeScratchKey(GrScratchKey*) const;
virtual sk_sp<GrSurface> createSurface(GrResourceProvider*) const = 0;
void assign(sk_sp<GrSurface> surface);
sk_sp<GrSurface> createSurfaceImpl(GrResourceProvider*, int sampleCnt, bool needsStencil,
GrSurfaceDescFlags, GrMipMapped) const;
bool instantiateImpl(GrResourceProvider* resourceProvider, int sampleCnt, bool needsStencil,
GrSurfaceDescFlags descFlags, GrMipMapped, const GrUniqueKey*);
// In many cases these flags aren't actually known until the proxy has been instantiated.
// However, Ganesh frequently needs to change its behavior based on these settings. For
// internally create proxies we will know these properties ahead of time. For wrapped
// proxies we will copy the properties off of the GrSurface. For lazy proxies we force the
// call sites to provide the required information ahead of time. At instantiation time
// we verify that the assumed properties match the actual properties.
GrInternalSurfaceFlags fSurfaceFlags;
private:
// For wrapped resources, 'fFormat', 'fConfig', 'fWidth', 'fHeight', and 'fOrigin; will always
// be filled in from the wrapped resource.
GrBackendFormat fFormat;
GrPixelConfig fConfig;
int fWidth;
int fHeight;
GrSurfaceOrigin fOrigin;
SkBackingFit fFit; // always kApprox for lazy-callback resources
// always kExact for wrapped resources
mutable SkBudgeted fBudgeted; // always kYes for lazy-callback resources
// set from the backing resource for wrapped resources
// mutable bc of SkSurface/SkImage wishy-washiness
const UniqueID fUniqueID; // set from the backing resource for wrapped resources
LazyInstantiateCallback fLazyInstantiateCallback;
// If this is set to kSingleuse, then after one call to fLazyInstantiateCallback we will cleanup
// the lazy callback and then delete it. This will allow for any refs and resources being held
// by the standard function to be released. This is specifically useful in non-dll cases where
// we make lazy proxies and instantiate them immediately.
// Note: This is ignored if fLazyInstantiateCallback is null.
LazyInstantiationType fLazyInstantiationType;
SkDEBUGCODE(void validateSurface(const GrSurface*);)
SkDEBUGCODE(virtual void onValidateSurface(const GrSurface*) = 0;)
static const size_t kInvalidGpuMemorySize = ~static_cast<size_t>(0);
SkDEBUGCODE(size_t getRawGpuMemorySize_debugOnly() const { return fGpuMemorySize; })
virtual size_t onUninstantiatedGpuMemorySize() const = 0;
bool fNeedsClear;
// This entry is lazily evaluated so, when the proxy wraps a resource, the resource
// will be called but, when the proxy is deferred, it will compute the answer itself.
// If the proxy computes its own answer that answer is checked (in debug mode) in
// the instantiation method.
mutable size_t fGpuMemorySize;
// The last opList that wrote to or is currently going to write to this surface
// The opList can be closed (e.g., no surface context is currently bound
// to this proxy).
// This back-pointer is required so that we can add a dependancy between
// the opList used to create the current contents of this surface
// and the opList of a destination surface to which this one is being drawn or copied.
// This pointer is unreffed. OpLists own a ref on their surface proxies.
GrOpList* fLastOpList;
typedef GrIORefProxy INHERITED;
};
#endif