gecko-dev/image/Decoder.h
Andrew Osmond 4e0ba74469 Bug 1465619 - Part 7. Add support for recycling to image::Decoder. r=tnikkel
The owner for the decoder may implement IDecoderFrameRecycler to allow
the decoder to request a recycled frame instead of allocating a new one.
If none are available, it will fallback to allocating a new frame.

Not only may the IDecoderFrameRecycler not have any frames available for
recycling, the recycled frame itself may still be in use by other
entities outside of imagelib. Additionally it may still be required by
BlendAnimationFilter to restore the previous frame's data. It may even
be the same frame as to restore. In the worst case, we will simply
choose to allocate an entirely new frame, just like before.

When we allocate a new frame, that means the old frame we tried to
recycle will be taken out of circulation and not reused again,
regardless of why it failed.

Differential Revision: https://phabricator.services.mozilla.com/D7512
2018-10-22 13:40:35 -04:00

684 lines
22 KiB
C++

/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
#ifndef mozilla_image_Decoder_h
#define mozilla_image_Decoder_h
#include "FrameAnimator.h"
#include "RasterImage.h"
#include "mozilla/Maybe.h"
#include "mozilla/NotNull.h"
#include "mozilla/RefPtr.h"
#include "AnimationParams.h"
#include "DecoderFlags.h"
#include "Downscaler.h"
#include "ImageMetadata.h"
#include "Orientation.h"
#include "SourceBuffer.h"
#include "StreamingLexer.h"
#include "SurfaceFlags.h"
namespace mozilla {
namespace Telemetry {
enum HistogramID : uint32_t;
} // namespace Telemetry
namespace image {
class imgFrame;
struct DecoderFinalStatus final
{
DecoderFinalStatus(bool aWasMetadataDecode,
bool aFinished,
bool aHadError,
bool aShouldReportError)
: mWasMetadataDecode(aWasMetadataDecode)
, mFinished(aFinished)
, mHadError(aHadError)
, mShouldReportError(aShouldReportError)
{ }
/// True if this was a metadata decode.
const bool mWasMetadataDecode : 1;
/// True if this decoder finished, whether successfully or due to failure.
const bool mFinished : 1;
/// True if this decoder encountered an error.
const bool mHadError : 1;
/// True if this decoder encountered the kind of error that should be reported
/// to the console.
const bool mShouldReportError : 1;
};
struct DecoderTelemetry final
{
DecoderTelemetry(const Maybe<Telemetry::HistogramID>& aSpeedHistogram,
size_t aBytesDecoded,
uint32_t aChunkCount,
TimeDuration aDecodeTime)
: mSpeedHistogram(aSpeedHistogram)
, mBytesDecoded(aBytesDecoded)
, mChunkCount(aChunkCount)
, mDecodeTime(aDecodeTime)
{ }
/// @return our decoder's speed, in KBps.
int32_t Speed() const
{
return mBytesDecoded / (1024 * mDecodeTime.ToSeconds());
}
/// @return our decoder's decode time, in microseconds.
int32_t DecodeTimeMicros() { return mDecodeTime.ToMicroseconds(); }
/// The per-image-format telemetry ID for recording our decoder's speed, or
/// Nothing() if we don't record speed telemetry for this kind of decoder.
const Maybe<Telemetry::HistogramID> mSpeedHistogram;
/// The number of bytes of input our decoder processed.
const size_t mBytesDecoded;
/// The number of chunks our decoder's input was divided into.
const uint32_t mChunkCount;
/// The amount of time our decoder spent inside DoDecode().
const TimeDuration mDecodeTime;
};
/**
* Interface which owners of an animated Decoder object must implement in order
* to use recycling. It allows the decoder to get a handle to the recycled
* frames.
*/
class IDecoderFrameRecycler
{
public:
/**
* Request the next available recycled imgFrame from the recycler.
*
* @param aRecycleRect If a frame is returned, this must be set to the
* accumulated dirty rect between the frame being
* recycled, and the frame being generated.
*
* @returns The recycled frame, if any is available.
*/
virtual RawAccessFrameRef RecycleFrame(gfx::IntRect& aRecycleRect) = 0;
};
class Decoder
{
public:
NS_INLINE_DECL_THREADSAFE_REFCOUNTING(Decoder)
explicit Decoder(RasterImage* aImage);
/**
* Initialize an image decoder. Decoders may not be re-initialized.
*
* @return NS_OK if the decoder could be initialized successfully.
*/
nsresult Init();
/**
* Decodes, reading all data currently available in the SourceBuffer.
*
* If more data is needed and @aOnResume is non-null, Decode() will schedule
* @aOnResume to be called when more data is available.
*
* @return a LexerResult which may indicate:
* - the image has been successfully decoded (TerminalState::SUCCESS), or
* - the image has failed to decode (TerminalState::FAILURE), or
* - the decoder is yielding until it gets more data (Yield::NEED_MORE_DATA), or
* - the decoder is yielding to allow the caller to access intermediate
* output (Yield::OUTPUT_AVAILABLE).
*/
LexerResult Decode(IResumable* aOnResume = nullptr);
/**
* Terminate this decoder in a failure state, just as if the decoder
* implementation had returned TerminalState::FAILURE from DoDecode().
*
* XXX(seth): This method should be removed ASAP; it exists only because
* RasterImage::FinalizeDecoder() requires an actual Decoder object as an
* argument, so we can't simply tell RasterImage a decode failed except via an
* intervening decoder. We'll fix this in bug 1291071.
*/
LexerResult TerminateFailure();
/**
* Given a maximum number of bytes we're willing to decode, @aByteLimit,
* returns true if we should attempt to run this decoder synchronously.
*/
bool ShouldSyncDecode(size_t aByteLimit);
/**
* Gets the invalidation region accumulated by the decoder so far, and clears
* the decoder's invalidation region. This means that each call to
* TakeInvalidRect() returns only the invalidation region accumulated since
* the last call to TakeInvalidRect().
*/
nsIntRect TakeInvalidRect()
{
nsIntRect invalidRect = mInvalidRect;
mInvalidRect.SetEmpty();
return invalidRect;
}
/**
* Gets the progress changes accumulated by the decoder so far, and clears
* them. This means that each call to TakeProgress() returns only the changes
* accumulated since the last call to TakeProgress().
*/
Progress TakeProgress()
{
Progress progress = mProgress;
mProgress = NoProgress;
return progress;
}
/**
* Returns true if there's any progress to report.
*/
bool HasProgress() const
{
return mProgress != NoProgress || !mInvalidRect.IsEmpty() || mFinishedNewFrame;
}
/*
* State.
*/
/**
* If we're doing a metadata decode, we only decode the image's headers, which
* is enough to determine the image's intrinsic size. A metadata decode is
* enabled by calling SetMetadataDecode() *before* calling Init().
*/
void SetMetadataDecode(bool aMetadataDecode)
{
MOZ_ASSERT(!mInitialized, "Shouldn't be initialized yet");
mMetadataDecode = aMetadataDecode;
}
bool IsMetadataDecode() const { return mMetadataDecode; }
/**
* Sets the output size of this decoder. If this is smaller than the intrinsic
* size of the image, we'll downscale it while decoding. For memory usage
* reasons, upscaling is forbidden and will trigger assertions in debug
* builds.
*
* Not calling SetOutputSize() means that we should just decode at the
* intrinsic size, whatever it is.
*
* If SetOutputSize() was called, ExplicitOutputSize() can be used to
* determine the value that was passed to it.
*
* This must be called before Init() is called.
*/
void SetOutputSize(const gfx::IntSize& aSize);
/**
* @return the output size of this decoder. If this is smaller than the
* intrinsic size, then the image will be downscaled during the decoding
* process.
*
* Illegal to call if HasSize() returns false.
*/
gfx::IntSize OutputSize() const { MOZ_ASSERT(HasSize()); return *mOutputSize; }
/**
* @return either the size passed to SetOutputSize() or Nothing(), indicating
* that SetOutputSize() was not explicitly called.
*/
Maybe<gfx::IntSize> ExplicitOutputSize() const;
/**
* Sets the expected image size of this decoder. Decoding will fail if this
* does not match.
*/
void SetExpectedSize(const gfx::IntSize& aSize)
{
mExpectedSize.emplace(aSize);
}
/**
* Is the image size what was expected, if specified?
*/
bool IsExpectedSize() const
{
return mExpectedSize.isNothing() || *mExpectedSize == Size();
}
/**
* Set an iterator to the SourceBuffer which will feed data to this decoder.
* This must always be called before calling Init(). (And only before Init().)
*
* XXX(seth): We should eliminate this method and pass a SourceBufferIterator
* to the various decoder constructors instead.
*/
void SetIterator(SourceBufferIterator&& aIterator)
{
MOZ_ASSERT(!mInitialized, "Shouldn't be initialized yet");
mIterator.emplace(std::move(aIterator));
}
SourceBuffer* GetSourceBuffer() const
{
return mIterator->Owner();
}
/**
* Should this decoder send partial invalidations?
*/
bool ShouldSendPartialInvalidations() const
{
return !(mDecoderFlags & DecoderFlags::IS_REDECODE);
}
/**
* Should we stop decoding after the first frame?
*/
bool IsFirstFrameDecode() const
{
return bool(mDecoderFlags & DecoderFlags::FIRST_FRAME_ONLY);
}
/**
* Should blend the current frame with the previous frames to produce a
* complete frame instead of a partial frame for animated images.
*/
bool ShouldBlendAnimation() const
{
return bool(mDecoderFlags & DecoderFlags::BLEND_ANIMATION);
}
/**
* @return the number of complete animation frames which have been decoded so
* far, if it has changed since the last call to TakeCompleteFrameCount();
* otherwise, returns Nothing().
*/
Maybe<uint32_t> TakeCompleteFrameCount();
// The number of frames we have, including anything in-progress. Thus, this
// is only 0 if we haven't begun any frames.
uint32_t GetFrameCount() { return mFrameCount; }
// Did we discover that the image we're decoding is animated?
bool HasAnimation() const { return mImageMetadata.HasAnimation(); }
// Error tracking
bool HasError() const { return mError; }
bool ShouldReportError() const { return mShouldReportError; }
// Finalize frames
void SetFinalizeFrames(bool aFinalize) { mFinalizeFrames = aFinalize; }
bool GetFinalizeFrames() const { return mFinalizeFrames; }
/// Did we finish decoding enough that calling Decode() again would be useless?
bool GetDecodeDone() const
{
return mReachedTerminalState || mDecodeDone ||
(mMetadataDecode && HasSize()) || HasError();
}
/// Are we in the middle of a frame right now? Used for assertions only.
bool InFrame() const { return mInFrame; }
/// Is the image valid if embedded inside an ICO.
virtual bool IsValidICOResource() const
{
return false;
}
/// Type of decoder.
virtual DecoderType GetType() const
{
return DecoderType::UNKNOWN;
}
enum DecodeStyle {
PROGRESSIVE, // produce intermediate frames representing the partial
// state of the image
SEQUENTIAL // decode to final image immediately
};
/**
* Get or set the DecoderFlags that influence the behavior of this decoder.
*/
void SetDecoderFlags(DecoderFlags aDecoderFlags)
{
MOZ_ASSERT(!mInitialized);
mDecoderFlags = aDecoderFlags;
}
DecoderFlags GetDecoderFlags() const { return mDecoderFlags; }
/**
* Get or set the SurfaceFlags that select the kind of output this decoder
* will produce.
*/
void SetSurfaceFlags(SurfaceFlags aSurfaceFlags)
{
MOZ_ASSERT(!mInitialized);
mSurfaceFlags = aSurfaceFlags;
}
SurfaceFlags GetSurfaceFlags() const { return mSurfaceFlags; }
/// @return true if we know the intrinsic size of the image we're decoding.
bool HasSize() const { return mImageMetadata.HasSize(); }
/**
* @return the intrinsic size of the image we're decoding.
*
* Illegal to call if HasSize() returns false.
*/
gfx::IntSize Size() const
{
MOZ_ASSERT(HasSize());
return mImageMetadata.GetSize();
}
/**
* @return an IntRect which covers the entire area of this image at its
* intrinsic size, appropriate for use as a frame rect when the image itself
* does not specify one.
*
* Illegal to call if HasSize() returns false.
*/
gfx::IntRect FullFrame() const
{
return gfx::IntRect(gfx::IntPoint(), Size());
}
/**
* @return an IntRect which covers the entire area of this image at its size
* after scaling - that is, at its output size.
*
* XXX(seth): This is only used for decoders which are using the old
* Downscaler code instead of SurfacePipe, since the old AllocateFrame() and
* Downscaler APIs required that the frame rect be specified in output space.
* We should remove this once all decoders use SurfacePipe.
*
* Illegal to call if HasSize() returns false.
*/
gfx::IntRect FullOutputFrame() const
{
return gfx::IntRect(gfx::IntPoint(), OutputSize());
}
/// @return final status information about this decoder. Should be called
/// after we decide we're not going to run the decoder anymore.
DecoderFinalStatus FinalStatus() const;
/// @return the metadata we collected about this image while decoding.
const ImageMetadata& GetImageMetadata() { return mImageMetadata; }
/// @return performance telemetry we collected while decoding.
DecoderTelemetry Telemetry() const;
/**
* @return a weak pointer to the image associated with this decoder. Illegal
* to call if this decoder is not associated with an image.
*/
NotNull<RasterImage*> GetImage() const { return WrapNotNull(mImage.get()); }
/**
* @return a possibly-null weak pointer to the image associated with this
* decoder. May be called even if this decoder is not associated with an
* image.
*/
RasterImage* GetImageMaybeNull() const { return mImage.get(); }
RawAccessFrameRef GetCurrentFrameRef()
{
return mCurrentFrame ? mCurrentFrame->RawAccessRef()
: RawAccessFrameRef();
}
/**
* For use during decoding only. Allows the BlendAnimationFilter to get the
* current frame we are producing for its animation parameters.
*/
imgFrame* GetCurrentFrame()
{
return mCurrentFrame.get();
}
/**
* For use during decoding only. Allows the BlendAnimationFilter to get the
* frame it should be pulling the previous frame data from.
*/
const RawAccessFrameRef& GetRestoreFrameRef() const
{
MOZ_ASSERT(ShouldBlendAnimation());
return mRestoreFrame;
}
const gfx::IntRect& GetRestoreDirtyRect() const
{
MOZ_ASSERT(ShouldBlendAnimation());
return mRestoreDirtyRect;
}
const gfx::IntRect& GetRecycleRect() const
{
MOZ_ASSERT(ShouldBlendAnimation());
return mRecycleRect;
}
const gfx::IntRect& GetFirstFrameRefreshArea() const
{
return mFirstFrameRefreshArea;
}
bool HasFrameToTake() const { return mHasFrameToTake; }
void ClearHasFrameToTake() {
MOZ_ASSERT(mHasFrameToTake);
mHasFrameToTake = false;
}
IDecoderFrameRecycler* GetFrameRecycler() const { return mFrameRecycler; }
void SetFrameRecycler(IDecoderFrameRecycler* aFrameRecycler)
{
mFrameRecycler = aFrameRecycler;
}
protected:
friend class AutoRecordDecoderTelemetry;
friend class DecoderTestHelper;
friend class nsICODecoder;
friend class PalettedSurfaceSink;
friend class SurfaceSink;
virtual ~Decoder();
/*
* Internal hooks. Decoder implementations may override these and
* only these methods.
*
* BeforeFinishInternal() can be used to detect if decoding is in an
* incomplete state, e.g. due to file truncation, in which case it should
* return a failing nsresult.
*/
virtual nsresult InitInternal();
virtual LexerResult DoDecode(SourceBufferIterator& aIterator,
IResumable* aOnResume) = 0;
virtual nsresult BeforeFinishInternal();
virtual nsresult FinishInternal();
virtual nsresult FinishWithErrorInternal();
/**
* @return the per-image-format telemetry ID for recording this decoder's
* speed, or Nothing() if we don't record speed telemetry for this kind of
* decoder.
*/
virtual Maybe<Telemetry::HistogramID> SpeedHistogram() const { return Nothing(); }
/*
* Progress notifications.
*/
// Called by decoders when they determine the size of the image. Informs
// the image of its size and sends notifications.
void PostSize(int32_t aWidth,
int32_t aHeight,
Orientation aOrientation = Orientation());
// Called by decoders if they determine that the image has transparency.
//
// This should be fired as early as possible to allow observers to do things
// that affect content, so it's necessarily pessimistic - if there's a
// possibility that the image has transparency, for example because its header
// specifies that it has an alpha channel, we fire PostHasTransparency
// immediately. PostFrameStop's aFrameOpacity argument, on the other hand, is
// only used internally to ImageLib. Because PostFrameStop isn't delivered
// until the entire frame has been decoded, decoders may take into account the
// actual contents of the frame and give a more accurate result.
void PostHasTransparency();
// Called by decoders if they determine that the image is animated.
//
// @param aTimeout The time for which the first frame should be shown before
// we advance to the next frame.
void PostIsAnimated(FrameTimeout aFirstFrameTimeout);
// Called by decoders when they end a frame. Informs the image, sends
// notifications, and does internal book-keeping.
// Specify whether this frame is opaque as an optimization.
// For animated images, specify the disposal, blend method and timeout for
// this frame.
void PostFrameStop(Opacity aFrameOpacity = Opacity::SOME_TRANSPARENCY);
/**
* Called by the decoders when they have a region to invalidate. We may not
* actually pass these invalidations on right away.
*
* @param aRect The invalidation rect in the coordinate system of the unscaled
* image (that is, the image at its intrinsic size).
* @param aRectAtOutputSize If not Nothing(), the invalidation rect in the
* coordinate system of the scaled image (that is,
* the image at our output size). This must
* be supplied if we're downscaling during decode.
*/
void PostInvalidation(const gfx::IntRect& aRect,
const Maybe<gfx::IntRect>& aRectAtOutputSize = Nothing());
// Called by the decoders when they have successfully decoded the image. This
// may occur as the result of the decoder getting to the appropriate point in
// the stream, or by us calling FinishInternal().
//
// May not be called mid-frame.
//
// For animated images, specify the loop count. -1 means loop forever, 0
// means a single iteration, stopping on the last frame.
void PostDecodeDone(int32_t aLoopCount = 0);
/**
* Allocates a new frame, making it our current frame if successful.
*
* If a non-paletted frame is desired, pass 0 for aPaletteDepth.
*/
nsresult AllocateFrame(const gfx::IntSize& aOutputSize,
const gfx::IntRect& aFrameRect,
gfx::SurfaceFormat aFormat,
uint8_t aPaletteDepth = 0,
const Maybe<AnimationParams>& aAnimParams = Nothing());
private:
/// Report that an error was encountered while decoding.
void PostError();
/**
* CompleteDecode() finishes up the decoding process after Decode() determines
* that we're finished. It records final progress and does all the cleanup
* that's possible off-main-thread.
*/
void CompleteDecode();
/// @return the number of complete frames we have. Does not include the
/// current frame if it's unfinished.
uint32_t GetCompleteFrameCount()
{
if (mFrameCount == 0) {
return 0;
}
return mInFrame ? mFrameCount - 1 : mFrameCount;
}
RawAccessFrameRef AllocateFrameInternal(const gfx::IntSize& aOutputSize,
const gfx::IntRect& aFrameRect,
gfx::SurfaceFormat aFormat,
uint8_t aPaletteDepth,
const Maybe<AnimationParams>& aAnimParams,
RawAccessFrameRef&& aPreviousFrame);
protected:
Maybe<Downscaler> mDownscaler;
uint8_t* mImageData; // Pointer to image data in either Cairo or 8bit format
uint32_t mImageDataLength;
uint32_t* mColormap; // Current colormap to be used in Cairo format
uint32_t mColormapSize;
private:
RefPtr<RasterImage> mImage;
Maybe<SourceBufferIterator> mIterator;
IDecoderFrameRecycler* mFrameRecycler;
// The current frame the decoder is producing.
RawAccessFrameRef mCurrentFrame;
// The complete frame to combine with the current partial frame to produce
// a complete current frame.
RawAccessFrameRef mRestoreFrame;
ImageMetadata mImageMetadata;
gfx::IntRect mInvalidRect; // Tracks new rows as the current frame is decoded.
gfx::IntRect mRestoreDirtyRect; // Tracks an invalidation region between the
// restore frame and the previous frame.
gfx::IntRect mRecycleRect; // Tracks an invalidation region between the recycled
// frame and the current frame.
Maybe<gfx::IntSize> mOutputSize; // The size of our output surface.
Maybe<gfx::IntSize> mExpectedSize; // The expected size of the image.
Progress mProgress;
uint32_t mFrameCount; // Number of frames, including anything in-progress
FrameTimeout mLoopLength; // Length of a single loop of this image.
gfx::IntRect mFirstFrameRefreshArea; // The area of the image that needs to
// be invalidated when the animation loops.
// Telemetry data for this decoder.
TimeDuration mDecodeTime;
DecoderFlags mDecoderFlags;
SurfaceFlags mSurfaceFlags;
bool mInitialized : 1;
bool mMetadataDecode : 1;
bool mHaveExplicitOutputSize : 1;
bool mInFrame : 1;
bool mFinishedNewFrame : 1; // True if PostFrameStop() has been called since
// the last call to TakeCompleteFrameCount().
// Has a new frame that AnimationSurfaceProvider can take. Unfortunately this
// has to be separate from mFinishedNewFrame because the png decoder yields a
// new frame before calling PostFrameStop().
bool mHasFrameToTake : 1;
bool mReachedTerminalState : 1;
bool mDecodeDone : 1;
bool mError : 1;
bool mShouldReportError : 1;
bool mFinalizeFrames : 1;
};
} // namespace image
} // namespace mozilla
#endif // mozilla_image_Decoder_h